US20220125524A1 - Feedforward continuous positioning control of end-effectors - Google Patents

Abstract

A positioning controller (50) including a forward predictive model (60) and/or inverse control predictive model (70) for positioning control of an interventional device (30) including a portion (40) of an interventional device. In operation, the controller (50) may apply the forward predictive model (60) to a commanded positioning motion of the interventional device (30) to render a predicted navigated pose of the end-effector (40), and generate positioning data informative of a positioning by the interventional device (30) of said portion of interventional device (40) to a target pose based on the predicted navigated pose of said portion (40). Alternatively, antecedently or subsequently, the controller (50) may apply the control predictive model (70) to the target pose of the portion of interventional device (40) to render a predicted positioning motion of the interventional device (30), and generate positioning commands controlling a positioning by the interventional device (30) of said device portion (40) to the target pose based on the predicted positioning motion of the interventional device (30).

Description

FIELD OF THE INVENTION
• The present disclosure generally relates to a positioning control of portions of interventional devices (e.g. end-effectors of interventional devices) utilized in interventional procedures (e.g., minimally-invasive surgery, video-assisted thoracic surgery, vascular procedures, endoluminal procedures, orthopedic procedures). The present disclosure can specifically relate to incorporation of predictive models in the positioning control of such portions of interventional devices utilized in interventional procedures.
BACKGROUND OF THE INVENTION
• Continuous (or non-continuous) control—positioning of said device portion (e.g. end-effector) within a certain workspace—is one of the most commonly attempted forms of control in conventional rigid link robots. By taking the advantage of a discrete rigid link structure of the robot, a precise positioning of said portion of interventional device (e.g. end-effector) can be achieved as desired in structured applications such as manufacturing. However, usage of rigid link robots in clinical settings is less desired due to deformable, delicate nature of the soft-tissue human organs as well as patient safety.
• More particularly, robots that are biologically inspired may produce motion similar to snakes, elephants, and octopuses, which could be a very effective in manipulation of soft anatomical objects. Nonetheless, effective control, and especially effective continuous control, of robotic structures in clinical settings has proven extremely difficult to achieve in view of a complexity in continuum (or quasi-continuum) structures for a desired degrees of freedom being hard to either model mathematically or provide sufficient actuator inputs to enable consistent control.
• For example, one approach for continuous positioning control of an end-effector supported by a continuum robot involves a modelling and controlling of a configuration of the continuum robot whereby a static model, which is formulated as a set of nonlinear differential equations, attempts to account for lance deformation of the continuum robot due to bending, torsion, and elongation. However, the accuracy of mathematical modelling is susceptible to changes in the environmental conditions of the robot (e.g., temperature and humidity), which will influence mechanical properties of the robot components, and susceptible to a presence of any manufacturing inaccuracies or various working loads.
• By further example, another approach for positioning control of an end-effector supported by a robot a manipulation of the robot by projecting a set of allowable motions and a set of allowable forces into the joint space corresponding to a manipulator control. For instance, after insertion of the robot into a nasal cavity, the controller adjusts the position of each segment of the robot to increase the difference between a measured generalized force and an expected generalized force on each end disk. However, the increased degree-of-freedom of the robot to become more maneuverable has the adverse effect of complicating the kinematics for the robot. This may be particularly problematic in case of continuous control of a continuum robot.
• Moreover, even if effective positioning control of an end-effector supported by a robot is achieved, an incorrect calibration or usage of the robotic system or general wear and tear of the mechanical components may however negatively influence the prediction accuracy of the kinematics of the robotic structure. Again, this may be particularly problematic in case of continuous control of a continuum robot (or continuum robotic structure).
SUMMARY OF THE INVENTION
• Known techniques devised for positioning controls of portions of interventional devices (e.g. end-effectors) have provided limited benefits. Thus, there remains a need for improved techniques for providing an effective positioning controls of these portions of interventional devices. To this end, the present disclosure teaches a feed-forward positioning control, a feedback positioning control and a data collection. This control is preferably executed continuously.
• To this purpose, the invention proposes as a first embodiment a positioning controller for an interventional device including an imaging device as recited in any ofclaims1 to12.
• As a second and third embodiments, the invention proposes a non-transitory (or not) machine readable storage medium encoded with instructions according to claim13 and a method of positioning, executable by a positioning controller for an interventional device, according to claim14.
• Feed-Forward (preferably Continuous) Positioning Control. The present disclosure further teaches a predictive model approach for a feed-forward (preferably continuous) positioning control of a manual navigated positioning or an automated navigated positioning of a device portion (e.g. an end-effector) supported by an interventional device based on a predictive model configured with kinematics of the interventional device, optionally trained on these kinematics.
• One other embodiment of the present disclosure for a feed-forward (preferably continuous) positioning control of a device portion (e.g. an end-effector) is a (continuous) positioning controller including a forward predictive model configured with (optionally forward) kinematics of the interventional device (optionally trained on these forward kinematics of the interventional device) predictive of a navigated pose of the end-effector and/or an control (optionally inverse) predictive model configured with kinematics of the interventional device, optionally trained on inverse kinematics of the interventional device predictive of a positioning motion of the interventional device, predictive of a positioning motion of the interventional device.
• For purposes of the description and purposes of the present disclosure, the term “navigated pose” broadly encompasses a pose of said portion of the interventional device (e.g. an end-effector of the interventional device) upon being navigated via the interventional device to a spatial position during an interventional procedure, and the term “positioning motion” broadly encompasses any movement of the interventional device to navigate this device portion to the spatial position during the interventional procedure.
• In operation, the continuous positioning controller applies the forward predictive model to a commanded positioning motion of the interventional device to render a predicted navigated pose of the end-effector, and generates continuous positioning data informative of a positioning by the interventional device of said device portion to a target pose based on the predicted navigated pose of said device portion.
• Alternatively, antecedently or concurrently, the (preferably continuous) positioning controller applies the inverse predictive model to the target pose of said portion of the interventional device (e.g. end-effector) to render a predicted positioning motion of the interventional device, and generate (continuous) positioning commands controlling a positioning by the interventional device of said device portion to the target pose based on the predicted positioning motion of the interventional device.
• Feedback (preferably Continuous) Positioning Control. The present disclosure further teaches a predictive model approach for a feedback (preferably continuous) positioning control of a manual navigated positioning or an (semi-)automated navigated positioning of an imaging device associated with or attached to the portion of the interventional device (e.g. the end-effector of the interventional device) based on an imaging predictive model configured with kinematics of the interventional device (optionally correlated with image data) to receive imaging data from said imaging device as feedback to the manual navigated positioning or the automated navigated positioning of the imaging device (or of the portion of the interventional device linked to the imaging device—e.g. the end-effector) to a target pose. Said predictive model is or has been optionally trained on images generated by the device portion (e.g. end-effector).
• One embodiment of the present disclosure for the feedback (preferably continuous) control positioning of the imaging device (or of the portion of the interventional device linked to the imaging device—e.g. the end-effector) is a continuous positioning controller including an imaging predictive model trained on a correlation of a relative imaging by the end-effector and forward kinematics of the interventional device predictive of a navigated pose of the end-effector. The (continuous) positioning controller further may include an control predictive model configured with kinematics of the interventional device predictive of a corrective positioning motion of the interventional device. Optionally this control predictive model has been or is trained on inverse kinematics of the interventional device to output said predictive of a corrective positioning motion.
• For purposes of the description and purposes of the present disclosure, the term “relative imaging” broadly encompasses a generation of an image of the interventional procedure by the end-effector at a given pose relative to a reference image of the interventional procedure.
• In operation, subsequent to a navigation of the device portion (e.g. end-effector) to a target pose, the (preferably) continuous positioning controller applies the imaging predictive model to imaging data generated by the end-effector to render a predicted navigated pose of the device portion (e.g. end-effector), applies the control (or in particular inverse) predictive model to error positioning data derived from a differential aspect between the target pose of the end-effector and the predicted navigated pose of the end-effector to render a predicted corrective positioning motion of the interventional device, and generates continuous positioning commands controlling a corrective positioning by the interventional device of the imaging device (40) or of the portion of the interventional device associated with the imaging device (e.g. end-effector) to the target pose based on the predicted corrective positioning motion of the interventional device.
• Training Data Collection. Additionally, to facilitate the (optional continuous) positioning control of the manual navigated positioning or the automated navigated positioning of a portion of the interventional device (e.g. end-effector) via the predictive models being invariant to environmental differences (e.g., anatomical differences between patients such as patient size, heart location, etc.), the present disclosure further teaches (optionally training) data collection techniques premised on a navigated positioning of said interventional device portion via a pre-defined data point pattern and a recording of a spatial positioning of the interventional device and a pose of said interventional device portion at each data acquisition point to thereby collect (training) data for the predictive models to infer forward kinematics or inverse kinematics of the interventional device.
• One embodiment of the present disclosure for collection of (optionally training) data for predictive models is a (optionally training) data collection system for an interventional device including said portion of interventional device (e.g. an end-effector) and sensors adapted to provide position and/or orientation and/or shape information, at least a part of the sensors (332) being affixed to said interventional device portion (optionally in a fixed shape). Such sensors may comprise marking visible from an imaging system (e.g. X-Ray, MRI system), electromagnetic tracking sensors, transducer sensors and/or optical shape sensing provided by/in an optical fiber.
• One particular embodiment of this embodiment is a (optionally training) data collection system for interventional device including a portion of an interventional device (e.g. an end-effector) and an optical shape sensor with a segment of the optical shape sensor being affixed to the end-effector (optionally in a fixed shape).
• The (training) data collection system may employ a robot controller, a data acquisition controller, a positioning determination module (or shape sensing controller in the above-mentioned particular embodiment) and a data storage controller.
• In operation, the data acquisition controller may command the robot controller to control motion variable(s) of the interventional device in accordance with a pre-defined data point pattern, and the positioning determination module (or shape sensing controller) is configured to determine position information based on position and/or orientation and/or shape information received from sensors so as to output an estimation of a pose of said portion of the interventional device portion and/or an estimation of positioning motion of the interventional device at each data point of the pre-defined data point pattern. Said determined position information is thus retrieved or derived or extracted or received to the purpose of said estimations, optionally based on kinematics or interventional device behavior configuring the positioning determination module. In a particular embodiment, the positioning determination module may derive or received a derived shape data from said position and/or orientation and/or shape information to determine said estimations. In case this positioning determination module is a shape sensing controller (as in the particular embodiment above-mentioned) then it controls a shape sensing of the optical shape sensor including an estimation of a pose of the end-effector and an estimation of positioning motion of the interventional device at each data point of the pre-defined data point pattern.
• Said “derivation of shape data from said position and/or orientation and/or shape information” may be implemented according to known technique of deriving shape from the data provided by the “sensors”. As an example, the positions tracked by the sensors may give a good indication of the general shape of the segment of the interventional device bearing these sensors, and an algorithm (more or less developed according to eh distance between the sensors and of possible shape of this interventional device along this segment) may be developed to derive or reconstruct this shape. A dynamic tracking of this positioning may also give indication about the orientation of the deformations. Sensors may also provide strain information (e.g. Rayleigh or Bragg grating sensors) which can indicate local positioning and orientation of the interventional device from which shape can be derived and reconstructed (known techniques too).
• The estimation of the pose of the interventional device portion (e.g. an end-effector) is derived from the at least a part of the sensors affixed to the end-effector (optionally in a fixed shape).
• Throughout the robot controller may control a motion variable(s) of the interventional device in accordance with pre-defined data point pattern, the data storage controller may receive a communication from the shape sensing controller of an estimated pose of the end-effector for each data point, may receive a communication from the positioning determination module (or shape sensing controller) of an estimated positioning motion of the interventional device for each data point and may receive a communication from the robot controller of the at least one motion variable of the interventional device for each data point.
• In response to the communications, the data storage controller may store a temporal data sequence for the interventional device derived from the estimated pose of the end-effector at each data point, the estimated spatial positioning of the interventional device at each data point and the motion variable(s) of the interventional device at each data point. The temporal data sequence may serve as training data for machine learning models of any type, particularly for the predictive models of the present disclosure.
• Additionally, the data acquisition controller may further command the robot controller to control the motion variable(s) of the interventional device in accordance with additional pre-defined data point pattern(s) whereby the data storage controller generates additional temporal data sequences for machine learning models of any type, particularly for the predictive models of the present disclosure.
• Also, for purposes of the description and claims of the present disclosure:
• (1) terms of the art including, but not limited to, “end-effector”, “kinematics”, “position”, “positioning”, “pose”, “posing”, “motion” and “navigation” are to be interpreted as known in the art of the present disclosure and as exemplary described in the present disclosure;
• (2) examples of an end-effector include, but are not limited to, an intraoperative imaging devices, interventional tools/surgical instruments and surgical sutures, as known in the art of the present disclosure and hereinafter conceived;
• (3) the term “intraoperative imaging device” broadly encompasses all imaging devices, as known in the art of the present disclosure and hereinafter conceived, for illustrating anatomical objects/regions. Examples of intraoperative imaging devices include, but are not limited to, a transesophageal echocardiography transducer (e.g., X7-2t transducer, Philips), laparoscopic ultrasound transducers (e.g., L10-4lap transducer, Philips), optical cameras and sensing devices (e.g., tissue spectral sensing sensors, ECG electrodes, probes for electrophysiological mapping);
• (4) examples of interventional tools/surgical instruments include, but are not limited to, scalpels, cauterizers, ablation devices, needles, forceps, k-wires and associated drivers, endoscopes, awls, screwdrivers, osteotomes, chisels, mallets, curettes, clamps, forceps, periosteomes and j-needles, as known in the art of the present disclosure and hereinafter conceived;
• (5) the term “interventional device” broadly encompasses all devices, as known in the art of the present disclosure and hereinafter conceived, for supporting a positioning of end-effector during an application. Examples of an interventional device include, but are not limited to, continuum flexible robots, flexible interventional scopes and guidewires;
• (6) examples of flexible robots include, but are not limited to, hyper-redundant robots (e.g., multiple discrete links, serpentine links or concentric tubes), continuous backbone section robot (e.g., cable actuated), tendon-drive robots, gel-like soft robot and fluid filled tube actuated robot, as known in the art of the present disclosure and hereinafter conceived;
• (7) examples of flexible interventional scopes include, but are not limited to, an endoscope of a transoesophageal echocardiography (TEE) probe, an endoscope of a intracardiac echocardiography (ICE) probe, a laparoscope and a bronchoscope, as known in the art of the present disclosure and hereinafter conceived;
• (8) the term “predictive model” broadly encompasses all types of models, as known in the art of the present disclosure and hereinafter conceived, predictive of navigation variables related to a (optionally) continuous positioning control of portions of interventional device (e.g. end-effectors) in accordance with the present disclosure as exemplary described in the present disclosure. Optionally these predictive models may be configured with kinematics to output this prediction based on positioning dataset. Optionally, the predictive models have been trainable or are trained on kinematics of an interventional device. Examples of predictive models include, but are not limited to, artificial neural networks (e.g., feed-forward convolutional neural networks, recurrent neural networks, long short-term memory networks, autoencoder networks, generative adversarial networks, and many others deep learning neural networks);
• (9) the term “controller” broadly encompasses all structural configurations, as understood in the art of the present disclosure and as exemplary described in the present disclosure, of main circuit board or integrated circuit for controlling an application of various inventive principles of the present disclosure as exemplary described in the present disclosure. The structural configuration of the controller may include, but is not limited to, processor(s), computer-usable/computer readable storage medium(s), an operating system, application module(s), peripheral device controller(s), slot(s) and port(s), instructions to control. A controller may be housed within or communicatively linked to a workstation;
• (10) the term “application module” broadly encompasses an application incorporated within or accessible by a controller consisting of an electronic circuit (e.g., electronic components and/or hardware) and/or an executable program (e.g., executable software stored on non-transitory (or not) computer readable medium(s) and/or firmware) for executing a specific application; and
• (11) the terms “signal”, “data” and “command” broadly encompasses all forms of a detectable physical quantity or impulse (e.g., voltage, current, or magnetic field strength) as understood in the art of the present disclosure and as exemplary described in the present disclosure for transmitting information and/or instructions in support of applying various inventive principles of the present disclosure as subsequently described in the present disclosure. Signal/data/command communication various components of the present disclosure may involve any communication method as known in the art of the present disclosure including, but not limited to, signal/data/command transmission/reception over any type of wired or wireless datalink and a reading of signal/data/commands uploaded to a computer-usable/computer readable storage medium.
• The foregoing embodiments and other embodiments of the present disclosure as well as various structures and advantages of the present disclosure will become further apparent from the following detailed description of various embodiments of the present disclosure read in conjunction with the accompanying drawings. The detailed description and drawings are merely illustrative of the present disclosure rather than limiting, the scope of the present disclosure being defined by the appended claims and equivalents thereof.
BRIEF DESCRIPTION OF THE DRAWINGS
• FIG. 1 illustrates an exemplary embodiment of an interventional device including an end-effector as known in the art of the present disclosure.
• FIG. 2A illustrates an exemplary embodiment of a continuous position controller in accordance with various aspects of the present disclosure.
• FIG. 2B illustrates an exemplary embodiment of a continuous positioning state machine in accordance with various aspects of the present disclosure.
• FIG. 3A illustrates an exemplary embodiment of a transesophageal echocardiography (TEE) probe as known in the art of the present disclosure.
• FIG. 3B illustrates an exemplary embodiment of a handle of a transesophageal echocardiography (TEE) probe as known in the art of the present disclosure.
• FIGS. 3C-3F illustrate exemplary motions of the TEE probe ofFIG. 2A as known in the art of the present disclosure
• FIG. 3G illustrates an exemplary embodiment of robot manipulated transthoracic echocardiography (TTE) probe as known in the art of the present disclosure.
• FIG. 3H illustrates an exemplary embodiment of an optical shape sensed continuum robot as known in the art of the present disclosure.
• FIGS. 4A-4E illustrate first exemplary embodiments of a forward predictive model and a forward predictive method of the present disclosure.
• FIGS. 5A-5E illustrate first exemplary embodiments of an inverse predictive model and an inverse predictive method of the present disclosure.
• FIGS. 6A-6E illustrate second exemplary embodiments of a forward predictive model and a forward predictive method of the present disclosure.
• FIGS. 7A-7E illustrate second exemplary embodiments of an inverse predictive model and an inverse predictive method of the present disclosure.
• FIGS. 8A-8E illustrate third exemplary embodiments of a forward predictive model and a forward predictive method of the present disclosure.
• FIGS. 9A-9F illustrate fourth exemplary embodiments of a forward predictive model and a forward predictive method of the present disclosure.
• FIGS. 11A-11E illustrate first exemplary embodiments of an image predictive model and an image predictive method of the present disclosure.
• FIGS. 12A-12E illustrate second exemplary embodiments of a forward predictive model and forward predictive method of the present disclosure.
• FIGS. 13A-13D illustrate a closed loop pose control of the present disclosure.
• FIGS. 14A-14D illustrate a closed loop vector velocity control of the present disclosure.
• FIG. 15-FIG. 18 illustrate a training data collection system and method of the present disclosure.
• FIG. 19 illustrates an exemplary embodiment of a continuous positioning controller of the present disclosure.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
• The present disclosure is applicable to numerous and various applications that require continuous position control of an end-effector. Examples of such applications include, but are not limited to, minimally-invasive procedures (e.g., endoscopic hepatectomy, necrosectomy, prostatectomy, etc.), video-assisted thoracic surgery (e.g., lobetectomy, etc.), minimally-vascular procedures (e.g., via catheters, sheaths, deployment systems, etc.), minimally medical diagnostic procedures (e.g., endoluminal procedures via endoscopes or bronchoscopes), orthopedic procedures (e.g., via k-wires, screwdrivers, etc.) and non-medical applications.
• The present disclosure improves upon continuous position control of an end-effector during such applications by providing a prediction of poses of the end-effector and/or positioning motions of the interventional device that may be utilized to control and/or corroborate a manual navigated positioning or an automated navigated positioning of the end-effector.
• To facilitate an understanding of the present disclosure, the following description ofFIG. 1 teaches exemplary embodiments of an interventional device including an end-effector as known in the art of the present disclosure. From the description ofFIG. 1, those having ordinary skill in the art of the present disclosure will appreciate how to apply the present disclosure to make and use additional embodiments of interventional device including an end-effector as known in the art of the present disclosure and hereinafter conceived.
• Referring toFIG. 1, in practice, aninterventional device30 supports a manual navigated positioning or an automated navigated positioning of an end-effector40 during an application as symbolized by the arrows extending from end-effector40. Examples ofinterventional device30 include, but are not limited to, continuum flexible robots, flexible interventional scopes and guidewires, and examples of end-effector40 include, but are not limited to, intraoperative imaging devices, interventional tools/surgical instruments and surgical sutures.
• In operation, navigation command(s)31, actuation signal(s)32 and/or navigation force(s)33 are communicated to/imposed ontointerventional device30 wherebyinterventional device30 is translated, rotated and/or pivoted in accordance with the navigation command(s)31, actuation signal(s)32 and/or navigation force(s)33 to thereby navigate end-effector40 to a target pose (i.e., a location and an orientation in the application space).
• For example,FIGS. 3A and 3B shows a transesophageal echocardiography (TEE)probe130 as an embodiment ofinterventional device30 having an imaging end-effector140 that is insertable through a mouth of a patient P into an esophagus to capture images of a heart of the patient P, and a practitioner (not shown) or a robot controller100 (FIG. 2B) operates ahandle132 ofTEE probe130 to repositionTEE probe130 within patient P to thereby navigate imaging end-effector140 to a target pose.
• More particularly,TEE probe130 includes flexibleelongate member131, handle132 and imaging end-effector140. The flexibleelongate member131 is sized and/or shaped, structurally arranged, and/or otherwise configured to be positioned within a body lumen of a patient, such as an esophagus. The imaging end-effector140 is mounted at a distal end of themember131 and includes one or more ultrasound transducer elements whereby imaging end-effector140 is configured to emit ultrasonic energy towards an anatomy (e.g., the heart) of the patient P. The ultrasonic energy is reflected by the patient's vasculatures and/or tissue structures whereby the ultrasound transducer elements in the imaging end-effector140 receive the reflected ultrasound echo signals. In some embodiments, theTEE probe130 may include an internal or integrated processing component that can process the ultrasound echo signals locally to generate image signals representative of the patient P's anatomy under imaging. In practice, the ultrasound transducer element(s) may be arranged to provide two-dimensional (2D) images or three-dimensional (3D) images of the patient P's anatomy. The images acquired by theTEE probe130 are dependent on the depth of insertion, the rotation, and/or the tilt of the imaging end-effector140, as described in greater detail herein.
• Thehandle132 is coupled a proximal end of themember131. Thehandle132 includes control elements for navigating the imaging end-effector140 to the target pose. As shown, thehandle132 includesknobs133 and134, and aswitch135. Theknob133 flexes themember131 and the imaging end-effector140 along an anterior-posterior plane of the patient P (e.g., heart). Theknob134 flexes themember131 and the imaging end-effector140 along a left-right plane of the patient P. Theswitch135 controls beamforming at the imaging end-effector140 (e.g., adjusting an angle of an imaging plane).
• In a manual navigated embodiment, the practitioner manually dials theknobs133 and134 and/or manually turns theswitch135 on and/or off as needed to navigate imaging end-effector140 to the target pose. The practitioner may receive a display of the images generated by imaging end-effector140 to thereby impose navigation forces33 (FIG. 1) to control theknobs133 and134 and/or theswitch135 on thehandle132.
• In an automated navigated embodiment, a robotic system (not shown) may include electrical and/or mechanical components (e.g., motors, rollers, and gears) configured to dial theknobs133 and134 and/or turn theswitch135 on and/or off wherebyrobot controller100 may receive motion control commands31 (FIG. 1) from a navigation controller (not shown) or an input device (not shown) to thereby control theknobs133 and134 and/or theswitch135 on thehandle132. Alternatively,robot controller100 may to configured to directly manipulate theTEE probe130 via actuation signals32 (FIG. 1) based on a guidance method implemented byrobot controller100.
• TheTEE probe130 is maneuverable in various degrees of freedom.FIGS. 3C-3F illustrate various mechanisms for maneuvering theTEE probe130.
• FIG. 3C is a schematic diagram illustratingTEE probe130 may be manually advanced into a patient's esophagus as shown by thearrow131bor withdrawn from the patient's esophagus as shown by thearrow131c. TheTEE probe130 can be manually or robotically rotated left (e.g., counter-clockwise) or right (e.g., clockwise) with respect to alongitudinal axis131aof theTEE probe130 as shown by thearrows139aand139b, respectively. The rotations of themember131 can be described by a parameter, denoted as γ.
• FIG. 3D is a schematic diagram illustratingTEE probe130 being electronically rotated from 0 degree to 180 degrees (e.g., for beamforming) as shown by thearrows136aand136b, for example, by a manual or robotic control ofswitch135 on thehandle132. The rotations of the imaging planes can be described by a parameter, denoted as ω.
• FIG. 3E is a schematic diagram illustratingTEE probe130 being flexed along an anterior-posterior plane, for example, with respect to a patient's heart, as shown by thearrows137aand137b, for example, by a manual or robotic dialing theknob134 on thehandle132. The flexing along the anterior-posterior plane can be described by a parameter, denoted as α.
• FIG. 3F is a schematic diagram illustratingTEE probe130 being flexed along a left-right plane, for example, with respect to a patient's heart, as shown by thearrows138aand138b, for example, by a manual or robotic dialing theknob133 on thehandle132. The flexing along the left-right plane can be described by a parameter, denoted as β.
• By further example of an exemplary embodiment of interventional device30 (FIG. 1),FIG. 3G shows transthoracic echocardiography (TTE)probe240 configured to capture ultrasound images of an anatomy of patient P from the outside of the body of patient P and arobot230 as shown for manually or robotically handlingTTE probe240 to reposition theTTE probe240 to a target pose (i.e., a location and/or an orientation ofTTE probe240 relative to patient P). More particularly,robot230 includes a plurality oflinks231 coupled to a plurality ofjoints232 configured to hold theTTE probe240 and maneuver theTTE probe240 on an external surface of a patient P (e.g., around the chest area for imaging the heart).
• In a manual navigated embodiment, the practitioner manually imposes navigation forces33 (FIG. 1) onto links(s)231 to thereby translates, rotates and/or pivots link(s)231 ofrobot230 to navigateimaging TTE probe240 to the target pose. The practitioner may receive a display of the images generated byTTE probe240 to utilize as a basis for control oflinks231 ofrobot230.
• In an automated navigated embodiment, a robotic system (not shown) include electrical and/or mechanical components (e.g., motors, rollers, and gears) configured to steer thelinks231 ofrobot230 wherebyrobot controller101 receives motion control commands32 (FIG. 1) from a navigation controller (not shown) or an input device (not shown) in the form of Cartesian velocity parameters or joint velocity parameters to thereby steer thelinks231 ofrobot230. Alternatively,robot controller101 may to configured to directly manipulate theTTE probe240 via actuation signals32 (FIG. 1) based on a guidance method implemented byrobot controller101.
• By further example of an exemplary embodiment of interventional device30 (FIG. 1A),FIG. 3H shows a shape-sensedguidewire332 embedded or attached to acontinuum robot331 having an end-effector340. The shape-sensedguidewire232 incorporates optical shape sensing (OSS) technology as known in the art of the present disclosure. More particularly, ashape sensing controller103 uses light along a multicoreoptical fiber333 ofguidewire332 for device localization and navigation during surgical intervention. The principle involved makes use of distributed strain measurements in the optical fiber using characteristic Rayleigh backscatter or controlled grating patterns (e.g. Fiber Bragg Gratings). In practice,shape sensing controller103 acquires registered shapes of thecontinuum robot331 via shape-sensedguidewire332 asrobot controller102 or a practitioner (not shown) navigatescontinuum robot331 within patient P to position end-effector340 at a target pose.
• To further facilitate an understanding of the present disclosure, the following description ofFIGS. 2A and 2B teaches exemplary embodiments of a continuous position controller of the present disclosure and a continuous positioning state machine of the present disclosure, respectively. From the description ofFIGS. 2A and 2B, those having ordinary skill in the art of the present disclosure will appreciate how to apply the present disclosure to make and use additional embodiments of a continuous position controller of the present disclosure and a continuous positioning state machine of the present disclosure.
• Additionally, TEE probe130 (FIG. 3A),robot230/TTE probe240 (FIG. 3G) and continuum robot331 (FIG. 3H) are utilized herein as non-limiting examples of aninterventional device30 including end-effector40 (FIG. 1) to support the description of various embodiments of a continuous position controller of the present disclosure. Nonetheless, those having ordinary skill in the art of the present disclosure will appreciate how to apply the present disclosure to various and numerous additional embodiments ofinterventional device30 including end-effector40.
• Referring toFIGS. 1, 2A and 2B,interventional device30 including end-effector40 and acontinuous positioning controller50 of the present disclosure represent a continuouspositioning state machine90.
• Specifically, as shown inFIG. 2B, a state S92 of continuouspositioning state machine90 encompasses a navigation ofinterventional device30 including end-effector40 in accordance with a particular application (e.g., a minimally-invasive procedure, a video-assisted thoracic surgery, a minimally-vascular procedure, a minimally medical diagnostic procedure or an orthopedic procedure). In practice, the application may involve a manual navigation positioning or an automated navigation positioning of the end-effector40 to a target pose whereby the application incorporates imaging guidance (e.g., image segmentation, image registration, path planning, etc.), interventional device tracking (e.g., electromagnetic, optical or shape sensing), and/or any other navigation technique for positioning end-effector40 to the target pose.
• An execution of state S92 of continuouspositioning state machine90 results in a generation ofnavigation data34 and an optional generation ofauxiliary data35. Generally, in practice,navigation data34 will be in the form of navigation command(s)31, actuation signal(s)32 and/or navigation force(s)33 communicated to/imposed ontointerventional device30, andauxiliary data35 will be in the form of image(s) ofinterventional device30 and/or end-effector40, operational characteristics of interventional device30 (e.g., shape, strain, twist, temperature, etc.) and operational characteristics of end-effector40 (e.g., pose, forces, etc.).
• In response thereto, a state S94 of continuouspositioning state machine90 encompasses continuous positioning control bycontinuous position controller50 of the navigation ofinterventional tool30 and end-effector40 in accordance with the state S92. To this end,continuous position controller50 employs a forwardpredictive model60 of the present disclosure, an inversepredictive model70 of the present disclosure and/or an imagingpredictive model80 of the present disclosure.
• In practice, as will be further explained in the present disclosure, forwardpredictive model60 may be any type of machine learning model or equivalent for a regression of a positioning motion of theinterventional device30 to a navigated pose of end-effector40 (e.g., a neural network) that is suitable for the particular type ofinterventional device30 being utilized in the particular type of application being implemented, whereby forwardpredictive model60 is trained on the forward kinematics ofinterventional device30 that is predictive of a navigated pose of end-effector40.
• In operation, forwardpredictive model60 inputs navigation data34 (andauxiliary data35 if communicated) associated with a manual navigation or an automated navigation ofinterventional device30 to thereby predict a navigated pose of end-effector40 corresponding to the navigation ofinterventional device30, and outputscontinuous positioning data51 informative of a positioning byinterventional device30 of the end-effector40 to the target pose based on the predicted navigated pose of end-effector40.Continuous positioning data51 may be utilized in state S92 as an control to determine an accuracy and/or execute a recalibration of the manual navigation or the automated navigation ofinterventional device30 to position end-effector40 to the target pose.
• In practice, as will be further explained in the present disclosure, inversepredictive model70 may be any type any type of machine learning model or equivalent for a regression a target pose of end-effector40 to a positioning motion of the interventional device30 (e.g., a neural network) that is suitable for the particular type ofinterventional device30 being utilized in the particular type of application being implemented, whereby inversepredictive model60 is trained on the inverse kinematics ofinterventional device30 that is predictive of a positioning motion of theinterventional device30.
• In operation, inversepredictive model70 inputs navigation data34 (andauxiliary data35 if communicated) associated with a target pose of end-effector40 to thereby predict a positioning motion ofinterventional device30 for positioning end-effector40 to the target pose, and outputs continuous positioning commands52 for controlling a positioning byinterventional device30 of end-effector40 to the target pose based on the predicted positioning motion ofinterventional device30. Continuous positioning commands52 may be utilized in state S92 as a control to execute the manual navigation or the automated navigation ofinterventional device30 to position end-effector40 to the target pose.
• In practice, as will be further explained in the present disclosure, imagingpredictive model80 may be any type of machine learning model or equivalent for a regression of relative imaging by the end-effector40 to a navigated pose of end-effector40 (e.g., a neural network or a scale invariant feature transform network) that is suitable for the particular type of end-effector40 being utilized in the particular type of application being implemented, whereby inversepredictive model60 is trained on a correlation of relative imaging by end-effector40 and forward kinematics of theinterventional device30 that is predictive of a navigated pose of end-effector40.
• In operation, imagingpredictive model60 inputsauxiliary data35 in the form of images generated by the end-effector40 at one or more poses to thereby predict the navigated pose of end-effector40 as feedback data informative of a corrective positioning byinterventional device30 of the end-effector40 to a target pose. The feedback data is utilized in a closed loop of state S94 to generate a differential between a target pose of end-effector40 and the predicted navigated pose of the end-effector40 whereby inverse predictmodel70 may input the differential to predict a corrective positioning motion ofinterventional device30 to reposition end-effector30 to the target pose.
• In practice, an embodiment ofcontinuous positioning controller50 may employ forwardpredictive model60, inversepredictive model70 and/or imagingpredictive model80.
• For example, an embodiment ofcontinuous positioning controller50 may employ only forwardpredictive model60 to facilitate a display of an accuracy of a manual navigation or the automated navigation ofinterventional device30 to position end-effector40 to the target pose.
• More particularly, a user interface is provided to display an image of an attempted navigation of end-effector40 to the target pose and an image of the predicted navigated pose of end-effector40 by forwardpredictive model60. A confidence ratio of the prediction is shown to the user. To evaluate prediction uncertainty, multiple feedforward iterations of forwardpredictive model60 are performed with dropout enabled stochastically as known in the art of the present disclosure.
• By further example, an embodiment ofcontinuous positioning controller50 may only employ inversepredictive model70 to command a manual navigation or an automated navigation ofinterventional device30 to position end-effector40 to the target pose.
• By further example, an embodiment ofcontinuous positioning controller50 may only employ imagingpredictive model80 to provide feedback data informative of a corrective positioning byinterventional device30 of the end-effector40 to a target pose.
• By further example, an embodiment ofcontinuous positioning controller50 may employ forwardpredictive model60 and inversepredictive model70 to thereby command a manual navigation or an automated navigation ofinterventional device30 to position end-effector40 to the target pose and to display of an accuracy of a manual navigation or the automated navigation ofinterventional device30 to position end-effector40 to the target pose.
• By further example, an embodiment ofcontinuous positioning controller50 may employ inversepredictive model70 and imagingpredictive model80 to thereby command a manual navigation or an automated navigation ofinterventional device30 to position end-effector40 to the target pose and to provide feedback data informative of a corrective positioning byinterventional device30 of the end-effector40 to the target pose.
• To further facilitate an understanding of the present disclosure, the following description ofFIGS. 4-14 teaches exemplary embodiments of forward predictive models, inverse predictive models and imaging predictive models of the present disclosure. From the description ofFIGS. 4-14, those having ordinary skill in the art of the present disclosure will appreciate how to apply the present disclosure to make and use additional embodiments of forward predictive models, inverse predictive models and imaging predictive models of the present disclosure.
• FIGS. 4A-4E illustrate a training and an application of a forwardpredictive model60aof the present disclosure trained to forward kinematics of an interventional device30 (FIG. 1) predictive of a navigated pose of an end-effector40 (FIG. 1) to thereby, during an interventional procedure, facilitate an application of forwardpredictive model60ato a commanded positioning motion ofinterventional device30 to render a predicted navigated pose of end-effector40, whereby continuous positioning controller50 (FIG. 1) may generatecontinuous positioning data51 informative of a positioning byinterventional device30 of end-effector40 to a target pose based on the predicted navigated pose of end-effector40.
• More particularly, referring toFIG. 4A, a training phase of forwardpredictive model60ainvolves a training controller (not shown) configured to interpret ground-truth training dataset D as exemplary taught in a subsequent description ofFIGS. 16-18 of the present disclosure. The dataset D consists of n sequences W that contains i data points represented by a 2-tuple: d_i=(T_i, Q_i). The 2-tuple consists of an end-effector pose (T∈SE (3))62aand a sequence of j consecutive joint variables (Q∈(q_t, q_t+1. . . q_t+j))61aacquired at sequential time points starting from t to t+j. Entry q_tstands for all joint variables that are controlled by the robot controller (not shown).
• In practice, training dataset D is a collection of expert data with reasonable coverage of different navigations ofinterventional device30. To this end, the diverse dataset training dataset D for learning should incorporate manufacturing differences between various types of robots, performance characteristics, wear and tear of the hardware components, and other system dependent and independent factors, such as temperature or humidity of the environment in which robot operates.
• Referring toFIG. 4B, an application phase of forwardpredictive model60ainvolvescontinuous positioning controller50 executing a deep learning type algorithm using feed-forwardpredictive model60afor regression of the commanded position motion ofinterventional device30 to a navigated pose of end-effector40. Forwardpredictive model60ais configured to infer the pose ({circumflex over (T)}∈SE(3))62bof end-effector40 given a sequence Q of j consecutivejoint variables61bof interventional device30 (e.g., parameters α, β as shown inFIG. 3B).
• In one embodiment as shown inFIG. 4E, forwardpredictive model60aemploys aneural network base160aincluding an input layer, hidden layers and an output layer derived from a combination of one or more fully connected layers (FCL)163a, one or more convolutional layers (CNL)164a, one or more recurrent layers (RCL)165a, and one or more long term short memory (LSTM) layers166a.
• In practice, the combination of layers is configured to implement a regression of joint variables Q to pose {circumflex over (T)}.
• In one embodiment for implementing a regression of joint variables Q to pose {circumflex over (T)},neural network base160aincludes a set of N number of fully connectedlayers163a.
• In a second embodiment for implementing a regression of joint variables Q to pose {circumflex over (T)},neural network base160aincludes a set of N convolutional layers164afollowed by either a set of M fully connectedlayers163aor a set of Wrecurrent layers165aor a set of W long term short memory layers166a.
• In a third embodiment for implementing a regression of joint variables Q to pose {circumflex over (T)},neural network base160aincludes a set of N convolutional layers164afollowed combination of a set of M fully connectedlayers163aand a set of Wrecurrent layers165aor a set W of long term short memory layers166a.
• In practice, a fully connectedlayer163amay include K neurons, where N, M, W, K may be any positive integer, and values may vary depending on the embodiments. For example, N may be about 8, M may be about 2, W may be about 2, and K can be about 1000. Alson, aconvolutional layer164amay implement a non-linear transformation, which may be a composite function of operations (e.g., batch normalization, rectified linear units (ReLU), pooling, dropout and/or convolution), and aconvolutional layer164amay also include a non-linearity function (e.g. including rectified non-linear ReLU operations) configured to extract rectified feature maps.
• Further in practice, one of thelayers163aor164aserve as an input layer for inputting asequence161aof joint variables Q, whereby a size of the sequence of joint variables Q may be ≥1, and one of thelayers163a,165aand166amay serve as an output layer for outputting apose162aof end-effector40 in Cartesian space (e.g., a translation and a rotation of the end-effector40 in Cartesian space). The outputted pose of end-effector40 in Cartesian space may be represented as vectorial parametrization and/or non-vectorial parametrization of a rigid-body position and orientation. More particularly, the parametrizations may be in the form of Euler angles, quaternions, matrix, exponential map, and/or angle-axis representing rotations and/or translations (e.g., including a direction and a magnitude for the translations).
• Also in practice, the output layer may be a non-linear fully connectedlayer163athat gradually shrinks a high-dimensional output of the lastconvolutional layer164aofneural network base160ato produce a set of output variables.
• In training, training weights of forwardpredictive model60aare constantly updated by comparing the outputs inferred forward predictive model ({circumflex over (T)})—given input sequence Q—with ground-truth end-effector pose T_ifrom a batch of training datasets D, which may be systematically or randomly selected from data memory (not shown). More particularly, the coefficients for the filters may be initialized with predefined or arbitrary values. The coefficients for the filters are applied to the batch of training datasets D via a forward propagation, and are adjusted via a backward propagation to minimize any output error.
• In application, forwardpredictive model60ainfers the pose ({circumflex over (T)})62aof end-effector40 given a sequence Q of j consecutivejoint variables62a.
• Still referring toFIG. 4E, the neural architecture of the exemplary embodiment of forwardpredictive model60aas illustrates has a unique number of layers, depending on the complexity of the task, which is specified in the output layer169adefined to predict the pose ({circumflex over (T)})62aof the end-effector40 given the sequence Q of j consecutivejoint variables61a. Loss function for training of this neural architecture may be defined as a sum of Euclidean distances between translational and rotational components, such as
• $\mathrm{loss}-{\hat{x}-x}_2+{\hat{r}-\frac{r}{r}}_2.$
• Referring toFIG. 4C andFIG. 4D, a stage S92aof an exemplaryinterventional procedure90aemploying TEE probe130 (FIG. 3A) and robot controller100 (FIG. 3B) and encompassesrobot controller100 receiving a position ofTEE probe130 as joint variables and sending movement signals to handle132 (FIG. 3A) ofTEE probe130. Using motorized knobs, handle132 pull/loosen the tendons ofTEE probe130, which will result in the motion of the end-effector140 (FIGS. 3C-3F) to a target pose. In practice, the position ofTEE probe130 as joint variables may be indicated by a user or an external tracking device or a guidance system.
• A stage S94aofprocedure90ainvolvesrobot controller100, forwardpredictive model50aand adisplay controller104.Robot controller100 stores and communicates the sequence of consecutive joint variables (Q)61ato forwardpredictive model60athat predicts the navigated pose {circumflex over (T)} of theend effector140.Continuous positioning controller50agenerates a confidence ratio of the prediction derived from uncertainty multiple feedforward iterations of forwardpredictive model60aperformed with dropout enabled stochastically as known in the art of the present disclosure. Forwardpredictive model50acommunicatescontinuous positioning data51aincluding the predicted navigated pose {circumflex over (T)} of theend effector140 and the confidence ratio to adisplay controller104, which in turn controls a display of animage105aof the navigated pose of end-effector140, animage106aof the navigated pose of end-effector140 and the confidence ratio for guidance purposes to end-effector140 to the target pose.
• FIGS. 5A-5E illustrate a training and an application of an inversepredictive model70aof the present disclosure trained to on inverse kinematics of an interventional device30 (FIG. 1) predictive of a positioning motion ofinterventional device30 to thereby, during an interventional procedure, facilitate an application of inversepredictive model70ato target pose (T)71aof end-effector40 to render a joint variable motion (q)72_tofinterventional device30, whereby continuous positioning controller50 (FIG. 1) may generate continuous positioning commands controlling a repositioning byinterventional device30 of the end-effector40 to the target pose based on the predicted positioning motion of theinterventional device40.
• More particularly, referring toFIG. 5A, a training phase of inversepredictive model70ainvolves a training controller (not shown) configured to interpret ground-truth training dataset D as exemplary taught in the description ofFIGS. 17-18. The dataset D consists of n sequences W that contains i data points represented by a 2-tuple: d_i=(T_i, Q_i). The 2-tuple consists of an end-effector pose (T∈SE(3))71aand a sequence of j consecutive joint variables (Q∈(q_t, q_t+1. . . q_t+j))72aacquired at sequential time points starting from t to t+j. Entry q_tstands for all joint variables that are controlled by the robot controller (not shown). Variable j may also equal to 1, meaning predictive model will be trained to infer single set of joint variables.
• In practice, training dataset D is a collection of expert data with reasonable coverage of different navigations ofinterventional device30. To this end, the diverse dataset training dataset D for learning should incorporate mechanical differences between various types of robots, wear and tear of the hardware components, and other system dependent factors.
• Referring toFIG. 5B, an application phase of inversepredictive model70ainvolvescontinuous positioning controller50 executing a deep learning type algorithm using inversepredictive model70afor motion regression. Inversepredictive model70ais configured to infer sequence {circumflex over (Q)} off consecutive joint variables72_t(e.g., parameters α, β as shown inFIG. 3B) to reach the pose (T∈SE(3))71tof end-effector40.
• In one embodiment as shown inFIG. 5E, inversepredictive model70aemploys aneural network base170aincluding an input layer, hidden layers and an output layer derived from a combination of one or more fully connected layers (FCL)173a, one or more convolutional layers (CNL)174a, one or more recurrent layers (RCL)175a, and one or more long term short memory (LSTM) layers176a.
• In practice, the combination of layers is configured to implement a regression of pose T to joint variables {circumflex over (Q)}.
• In one embodiment for implementing a regression of pose T to joint variables {circumflex over (Q)},neural network base170aincludes a set of N number of fully connectedlayers173a.
• In a second embodiment for implementing a regression of pose T to joint variables {circumflex over (Q)},neural network base170aincludes a set of N convolutional layers174afollowed by either a set of M fully connectedlayers173aor a set of Wrecurrent layers175aor a set of W long term short memory layers176a.
• In a third embodiment for implementing a regression of pose T to joint variables {circumflex over (Q)},neural network base170aincludes a set of N convolutional layers174afollowed combination of a set of M fully connectedlayers173aand a set of Wrecurrent layers175aor a set W of long term short memory layers176a.
• In practice, a fully connectedlayer173amay include K neurons, where N, M, W, K may be any positive integer, and values may vary depending on the embodiments. For example, N may be about 8, M may be about 2, W may be about 2, and K can be about 1000. Also, aconvolutional layer174amay implement a non-linear transformation, which may be a composite function of operations (e.g., batch normalization, rectified linear units (ReLU), pooling, dropout and/or convolution), and aconvolutional layer174amay also include a non-linearity function (e.g. including rectified non-linear ReLU operations) configured to extract rectified feature maps.
• Further in practice, one of thelayers173aor174aserve as an input layer for inputtingpose171aof end-effector40 in Cartesian space (e.g., a translation and a rotation of the end-effector40 in Cartesian space), and one of thelayers173a,175aand176amay serve as an output layer for outputting asequence172aof joint variables Q, whereby a size of the sequence of joint variables Q may be ≥1. The inputted pose of end-effector40 in Cartesian space may be represented as vectorial parametrization and/or non-vectorial parametrization of a rigid-body position and orientation. More particularly, the parametrizations may be in the form of Euler angles, quaternions, matrix, exponential map, and/or angle-axis representing rotations and/or translations (e.g., including a direction and a magnitude for the translations).
• Also in practice, the output layer may be a non-linear fully connectedlayer173athat gradually shrinks a high-dimensional output of the lastconvolutional layer174aofneural network base170ato produce a set of output variables.
• In training, training weights of inversepredictive model70aare constantly updated by comparing the output from the inverse predicted model {circumflex over (Q)}—given as input a ground-truth end effector pose T—with a ground-truth sequence Q_ifrom a batch of training datasets D, which may be systematically or randomly selected from data memory (not shown). More particularly, the coefficients for the filters may be initialized with predefined or arbitrary values. The coefficients for the filters are applied to the batch of training datasets D via a forward propagation, and are adjusted via a backward propagation to minimize any output error.
• In application, inversepredictive model70ainfers sequence {circumflex over (Q)} off consecutivejoint variables72b(e.g., parameters α, β as shown inFIG. 3B) to reach the provided pose (T∈SE(3))71bof end-effector40.
• Still referring toFIG. 5E, the neural architecture of the exemplary embodiment of inversepredictive model70aas illustrates has a unique number of layers, depending on the complexity of the task, which is specified in the output layer179adefined to predict sequence {circumflex over (Q)} of j consecutivejoint variables72b(e.g., parameters α, β as shown inFIG. 3B) to reach the pose (T∈SE(3))71bof end-effector40. A mean squared error (MSE) may be used as a loss function.
• Referring toFIG. 5C andFIG. 5D, a stage S92bof an exemplaryinterventional procedure90bemploying TEE probe130 (FIG. 3A) and anavigation controller103 and encompassesnavigation controller103 determining the target pose of end-effector140 and sending the target pose71ato inversepredictive model70a. In practice,navigation controller130 may implement any known guidance algorithm known in the art of the present disclosure.
• A stage S94A ofprocedure90binvolves inversepredictive model70aandrobot controller100. Inversepredictive model70ainfers sequence {circumflex over (Q)} of j consecutivejoint variables72a(e.g., parameters α, β as shown inFIG. 3B) to reach the pose (T∈SE(3))71aof end-effector40, andcontinuous positioning controller50bcommunicates acommand controller104 to control a positioning of end-effector50 via theinterventional device30 of end-effector40 to the target pose based on the predicted positioning motion of theinterventional device40.
• FIGS. 6A-6E illustrate a training and an application of a forwardpredictive model60bof the present disclosure trained on forward kinematics of an interventional device30 (FIG. 1) predictive of a navigated pose of an end-effector40 (FIG. 1) to thereby, during an interventional procedure, facilitate an application of forwardpredictive model60bto a commanded n-vector of joint velocities ofinterventional device30 to render a predicted linear velocity and/or angular velocity of end-effector40, whereby continuous positioning controller50 (FIG. 1) may generatecontinuous positioning data51 informative of a repositioning byinterventional device30 of end-effector40 to a target pose based on the predicted linear velocity and/or angular velocity of end-effector40.
• More particularly, referring toFIG. 6A, a training phase of forwardpredictive model60binvolves a training controller (not shown) configured to interpret ground-truth training dataset D as exemplary taught in the description ofFIGS. 16-18. The dataset D consists of n sequences W that contains i data points represented by a 2-tuple:
• $d_i=\left({\stackrel{.}{Q}}_{\iota },\left[\begin{array}{c}{v}_i\\ {\omega }_i\end{array}\right]\right).$
• The 2-tuple consists of commanded sequence of j consecutive joint velocities E ({dot over (Q)}∈({dot over (q)}_t,{dot over (q)}_t+1. . . {dot over (q)}_t+j))61band linear velocity and/orangular velocity62bof the end effector
• $\left[\begin{array}{c}v\\ \omega \end{array}\right].$
• Entry {dot over (q)}_istands for all joint velocities that are controlled by the robot controller (not shown). Sequence may also contain only one entry.
• In practice, training dataset D is a collection of expert data with reasonable coverage of different navigations ofinterventional device30. To this end, the diverse dataset training dataset D for learning should incorporate mechanical differences between various types of robots, wear and tear of the hardware components, and other system dependent factors.
• Referring toFIG. 6B, an application phase of forwardpredictive model60binvolvescontinuous positioning controller50 executing a deep learning type algorithm using feed-forwardpredictive model60bfor end effect motion regression. Forwardpredictive model60bis configured to infer the linear velocity and/orangular velocity62bof end-effector40 given n-vector ofjoint velocities61bofinterventional device30.
• In one embodiment as shown inFIG. 6E, forwardpredictive model60bemploys aneural network base160bincluding an input layer, hidden layers and an output layer derived from a combination of one or more fully connected layers (FCL)163b, one or more convolutional layers (CNL)164b, one or more recurrent layers (RCL)165b, and one or more long term short memory (LSTM) layers166b.
• In practice, the combination of layers is configured to implement a regression of joint velocities of theinterventional device30 to a linear velocity and/or an angular velocity of end-effector40.
• In one embodiment for implementing a regression of joint velocities of theinterventional device30 to a linear velocity and/or an angular velocity of end-effector40,neural network base160bincludes a set of N number of fully connectedlayers163b.
• In a second embodiment for implementing a regression of joint velocities of theinterventional device30 to a linear velocity and/or an angular velocity of end-effector40,neural network base160bincludes a set of N convolutional layers164bfollowed by either a set of M fully connectedlayers163bor a set of Wrecurrent layers165bor a set of W long termshort memory layers166b.
• In a third embodiment for implementing a regression of joint velocities of theinterventional device30 to a linear velocity and/or an angular velocity of end-effector40,neural network base160bincludes a set of N convolutional layers164bfollowed combination of a set of M fully connectedlayers163band a set of Wrecurrent layers165bor a set W of long termshort memory layers166b.
• In practice, a fully connectedlayer163bmay include K neurons, where N, M, W, K may be any positive integer, and values may vary depending on the embodiments. For example, N may be about 8, M may be about 2, W may be about 2, and K can be about 1000. Also, aconvolutional layer164bmay implement a non-linear transformation, which may be a composite function of operations (e.g., batch normalization, rectified linear units (ReLU), pooling, dropout and/or convolution), and aconvolutional layer164bmay also include a non-linearity function (e.g. including rectified non-linear ReLU operations) configured to extract rectified feature maps.
• Further in practice, one of thelayers163bor164bserve as an input layer for inputting a sequence of j consecutive joint velocities ({dot over (Q)}∈({dot over (q)}_t, {dot over (q)}_t+1. . . {dot over (q)}_t+j)), whereby a size of the a sequence of j consecutive joint velocities ({dot over (Q)}∈({dot over (q)}_t, {dot over (q)}_t+1. . . {dot over (q)}_t+j)) may be ≥1, and one of thelayers163b,165band166bmay serve as an output layer for outputting a linear and angular velocity of the end effector
• $\left[\begin{array}{c}\hat{v}\\ \hat{\omega }\end{array}\right]\hspace{1em}$
• as regressed from last fully-connected layer (e.g. 6 units, 3 units for linear and 3 units for angular velocity) with linear or non-linear activation function.
• In training, training weights of forwardpredictive model60bare constantly updated by comparing the predicted linear and angular velocity
• $\left[\begin{array}{c}\hat{v}\\ \hat{\omega }\end{array}\right]\hspace{1em}$
• via forward velocity predictive model the —given a sequence of joint velocities—with linear velocity and/orangular velocity62bfrom a batch of training datasets D, which may be systematically or randomly selected from data memory (not shown). More particularly, the coefficients for the filters may be initialized with predefined or arbitrary values. The coefficients for the filters are applied to the batch of training datasets D via a forward propagation, and are adjusted via a backward propagation to minimize any output error.
• In application, forwardpredictive model60binfer the linear velocity and/orangular velocity62bof end-effector40 given sequence ofjoint velocities61bofinterventional device30.
• Still referring toFIG. 6E, the neural architecture consisting of inputs, neural network base described previously, and outputs. Input is a sequence of joint velocities, and output is linear and angular velocity that could be regressed from fully-connected layer with 6 units. The loss function may be a MSE as listed:
• $\mathrm{MSE}=\frac{1}{n}{\sum }_{i=1}^n\left(\left[\begin{array}{c}v\\ \omega \end{array}\right]-\left[\begin{array}{c}\hat{v}\\ \hat{\omega }\end{array}\right]\right).$
• Referring toFIG. 6C andFIG. 6D, a stage S92cof an exemplaryinterventional procedure90cemploying robot230 (FIG. 3A), TTE probe240 (FIG. 3A), and robot controller101 (FIG. 3G) and encompassesrobot controller101 receiving a position ofTTE probe240 as n-vector ofjoint velocities61bofinterventional device30 and sending n-vector ofjoint velocities61bofinterventional device30 torobot230. In practice, the position ofTTE probe240 as n-vector ofjoint velocities61bofinterventional device30 may be indicated by a user or an external tracking device or a guidance system.
• A stage S94cofprocedure90cinvolvesrobot controller101, forwardpredictive model50aand adisplay controller104.Robot controller101 stores and communicates the n-vector ofjoint velocities61bofinterventional device30 to forwardpredictive model60bthat predicts the linear velocity and/orangular velocity62bofTTE probe240.Continuous positioning controller50cgenerates a confidence ratio of the prediction derived from uncertainty multiple feedforward iterations of forwardpredictive model60bperformed with dropout enabled stochastically as known in the art of the present disclosure. Forwardpredictive model60bcommunicatescontinuous positioning data51bincluding a predicted navigated pose i of theTTE probe240 derived from the predicted linear velocity and/orangular velocity62bofTTE probe240, and further including the confidence ratio to adisplay controller104, which in turn controls a display of animage105aof the navigated pose ofTTE probe240, animage106aof the navigated poseTTE probe240 and the confidence ratio for guidance purposes to end-effector240 to the target pose.
• FIGS. 7A-7E illustrate a training and an application of an inversepredictive model70bof the present disclosure trained to on inverse kinematics of an interventional device30 (FIG. 1) predictive of a positioning motion ofinterventional device30 to thereby, during an interventional procedure, facilitate an application of inversepredictive model70bto a target pose of end-effector40 to render a predicted positioning motion ofinterventional device30, whereby continuous positioning controller50 (FIG. 1) may generate continuous positioning commands controlling a repositioning byinterventional device30 of the end-effector40 to the target pose based on the predicted positioning motion of theinterventional device40.
• More particularly, referring toFIG. 7A, a training phase of inversepredictive model70binvolves a training controller (not shown) configured to interpret ground-truth training dataset D as exemplary taught in the description ofFIGS. 17-18. The dataset D consists of n sequences W that contains i data points represented by a 2-tuple:
• $d_i=\left(\left[\begin{array}{c}{v}_i\\ {\omega }_i\end{array}\right],{\stackrel{.}{Q}}_i\right).$
• The 2-tuple consists of a linear velocity and/or angular velocity of the end effector and sequence of consecutive joint velocities ({dot over (q)}_t) acquired at sequential time points starting from t to t+j. Entry {dot over (q)}_tstands for all joint velocities that are controlled by the robot controller (not shown).
• In practice, training dataset D is a collection of expert data with reasonable coverage of different navigations ofinterventional device30. To this end, the diverse dataset training dataset D for learning should incorporate mechanical differences between various types of robots, wear and tear of the hardware components, and other system dependent factors.
• Referring toFIG. 7B, an application phase of inversepredictive model70binvolvescontinuous positioning controller50 executing a deep learning type algorithm using inversepredictive model70bfor joint velocity regression. Inversepredictive model70bis configured to infer a n-vector ofjoint velocities61bofinterventional device30 given a linear velocity and/orangular velocity62bof end-effector40.
• In one embodiment as shown inFIG. 7E, inversepredictive model70bemploys aneural network base170bincluding an input layer, hidden layers and an output layer derived from a combination of one or more fully connected layers (FCL)173b, one or more convolutional layers (CNL)174b, one or more recurrent layers (RCL)175b, and one or more long term short memory (LSTM) layers176b.
• In practice, the combination of layers is configured to implement a regression of a linear velocity and/or an angular velocity of end-effector40 to joint velocities of theinterventional device30.
• In one embodiment for implementing a regression of a linear velocity and/or an angular velocity of end-effector40 to joint velocities of theinterventional device30,neural network base170bincludes a set of N number of fully connectedlayers173b.
• In a second embodiment for implementing a regression of a linear velocity and/or an angular velocity of end-effector40 to joint velocities of theinterventional device30,neural network base170bincludes a set of N convolutional layers174bfollowed by either a set of M fully connectedlayers173bor a set of Wrecurrent layers175bor a set of W long termshort memory layers176b.
• In a third embodiment for implementing a regression of a linear velocity and/or an angular velocity of end-effector40 to joint velocities of theinterventional device30,neural network base170bincludes a set of N convolutional layers174bfollowed combination of a set of M fully connectedlayers173band a set of Wrecurrent layers175bor a set W of long termshort memory layers176b.
• In practice, a fully connectedlayer173bmay include K neurons, where N, M, W, K may be any positive integer, and values may vary depending on the embodiments. For example, N may be about 8, M may be about 2, W may be about 2, and K can be about 1000. Also, aconvolutional layer174bmay implement a non-linear transformation, which may be a composite function of operations (e.g., batch normalization, rectified linear units (ReLU), pooling, dropout and/or convolution), and aconvolutional layer174bmay also include a non-linearity function (e.g. including rectified non-linear ReLU operations) configured to extract rectified feature maps.
• Further in practice, one of thelayers173bor174bserve as an input layer for inputting angular and linear velocity
• $\left[\begin{array}{c}v\\ \omega \end{array}\right],$
• and one of thelayers173b,175band176bmay serve as an output layer for outputting a a sequence of j consecutive joint velocities ({dot over (Q)}∈({dot over (q)}_t, {dot over (q)}_t+1. . . {dot over (q)}_t+j)) that provided as output fromLSTM layers176b. Alternatively, single joint velocities may be regressed from fully-connectedlayer173bconsisting of m units, each unit for every join in the robot controlled by the robot controller. Fully-connectedlayer173bmay have linear or non-linear activation functions.
• In training, training weights of inversepredictive model70bare constantly updated by comparing the predicted sequence of joint velocities {dot over ({circumflex over (Q)})}—given a linear and angular velocity at input
• $\left(\begin{array}{c}{v}_i\\ {\omega }_i\end{array}\right)\hspace{1em}$
• —with the ground-truth sequence {dot over (Q)}_iof joint velocities from a batch of training datasets D, which may be systematically or randomly selected from data memory (not shown). More particularly, the coefficients for the filters may be initialized with predefined or arbitrary values. The coefficients for the filters are applied to the batch of training datasets D via a forward propagation, and are adjusted via a backward propagation to minimize any output error.
• In application, inversepredictive model70binfers a n-vector ofjoint velocities61bofinterventional device30 given a linear velocity and/orangular velocity62bof end-effector40.
• Still referring toFIG. 7E, the neural architecture of the exemplary embodiment of inverse velocity model consisting of inputs, neural network base, and outputs. Input is angular and linear velocity
• $\left[\begin{array}{c}v\\ \omega \end{array}\right]\hspace{1em}$
• and output is a sequence {dot over ({circumflex over (Q)})} of joint velocities that provided as output from LSTM layers. Alternatively, single joint velocities may be regressed from fully-connected layer consisting of m units, each unit for every join in the robot controlled by the robot controller. Fully-connected layer may have linear or non-linear activation functions.
• Referring toFIG. 7C andFIG. 7D, a stage S92dof an exemplaryinterventional procedure90demploying TTE probe240 (FIG. 3G) and anavigation controller103 and encompassesnavigation controller103 determining a linear velocity and/orangular velocity62bofTTE probe240 to a target pose and sending the linear velocity and/orangular velocity62bto inversepredictive model70b. In practice,navigation controller130 may implement any known guidance algorithm known in the art of the present disclosure.
• A stage S94dofprocedure90dinvolves inversepredictive model70bandrobot controller101. Inversepredictive model70binfers a n-vector ofjoint velocities61bofinterventional device30 given linear velocity and/orangular velocity62b, andcontinuous positioning controller50ccommunicates acontinuous positioning command52btorobot controller101 to thereby control a positioning ofTTE probe240 via the robot230 (FIG. 3G) to the target pose based on the predicted positioning motion of theinterventional device40.
• In practice, forwardpredictive model60a(FIG. 4A), inversepredictive model70a(FIG. 5A), forwardpredictive model60b(FIG. 64A) and inversepredictive model70b(FIG. 7A) may use additional auxiliary information, such as, for example, an image of the anatomy (e.g., ultrasound, endoscopic, or X-ray image), a force measured at the end effector and a shape of the robot. Other inputs, including information from spectral tissue sensing device, ECG or EEG signal, tissue conductivity, or other physiological signals could also be included depending on the application. For example, if the continuum-like robot operates inside the human heart, features available in the ultrasound image as well as electrophysiological signal could improve the localization of the end effector in respect to the anatomy, hence improve the guidance accuracy.
• Referring toFIGS. 8A and 8B, a forwardpredictive model60cis shown as being trained on ashape35a, animage35band aforce35cof the interventional device in addition to forward kinematics of the interventional device in addition to sequence Q ofjoint variables61aand end-effector posesT62a. Thus, in application, forwardpredictive model60cwill be capable of predicting a navigated pose of the end-effector from sequence Q ofjoint variables61aas well as theshape35a, theimage35band theforce35cof the interventional device.
• Those having ordinary skill in the art will know how to applyshape35a, animage35band aforce35cof the interventional device as well as any other additional auxiliary information to inversepredictive model70a, forwardpredictive model60band inversepredictive model70b.
• FIGS. 9A-9E illustrate a training and an application of a forwardpredictive model60dof the present disclosure trained on the forward kinematics of interventional device30 (FIG. 1) predictive of a navigated pose of an end-effector40 (FIG. 1) as well as shape of the robot to thereby, during an interventional procedure, facilitate an application of forwardpredictive model60dto sequence of consecutive shapes of aninterventional device30 with embedded OSS technology to render a predicted navigated pose of end-effector40 as well as shape, whereby continuous positioning controller50 (FIG. 1) may generatecontinuous positioning data51cinformative of a repositioning byinterventional device30 of end-effector40 to a target pose based on the predicted navigated pose of end-effector40.
• More particularly, referring toFIG. 9A, a training phase a training controller (not shown) configured to interpret ground-truth training dataset D as exemplary taught in the description ofFIGS. 16-18. This dataset consists of n sequences W that contains i data points represented by a 2-tuple: d_i=(H_i, H_i+1). This 2-tuple consists of a sequence of kconsecutive shapes61d(H_i∈(h_t, h_t+1. . . h_t+k)), where h∈(p₁. . . p_m) is a set of m vectors p_mthat describe both the position of an OSS sensor embedded in interventional device30 (e.g., shape-sensed guidewire) in 3D Euclidean space and auxiliary shape parameters such as strain, curvature, and twist. This 2-tuple further consists a sequence of kconsecutive shapes62bcontaining a future time point h_t+k+1such as H_i+1∈(h_t+1, h_t+2. . . h_t+k+1), where h∈(p₁. . . p_m) is a set of m vectors p_mthat describe both the position of OSSinterventional device30 in 3D Euclidean space and auxiliary shape parameters such as strain, curvature, and twist.
• In practice, training dataset D is a collection of expert data with reasonable coverage of different navigations of OSSinterventional device30. To this end, the diverse dataset training dataset D for learning should incorporate anatomies with different curvatures, magnitude of motion, mechanical differences between various types of robots, wear and tear of the hardware components, and other system independent factors such as temperature and humidity of the environment.
• Referring toFIG. 9B, an application phase of forwardpredictive model60dinvolves acontinuous positioning controller50dexecuting a deep learning type algorithm using forward with recurrent layers trained on expert data with reasonable coverage of different cases. Diverse dataset for learning incorporates various working conditions (temperature, fiber bending, etc.), different operational motion of the device, and differences in hardware (fiber, interrogator, etc.).
• In one embodiment as shown inFIG. 9E, forwardpredictive model60demploys a neural architecture sequentially including aninput layer163a, a sequence-to-sequence model263a,output layer262aand anextraction layer264. The neural architecture is a deep convolutional neural network with recurrent layers that are configured to infer the future sequence Ĥ_t+1consisting of k shapes and therefore use the last shape ĥ_t+k+1in the predicted sequence to estimate the position of OSSinterventional device30 in the future time point.
• In training, training weights of forwardpredictive model60dare constantly updated by comparing the sequence of future shapes Ĥ_i+1— predicted by a the model given an input sequence H_i— with ground-truth future sequence H_i+1from the training dataset D, which may be systematically or randomly selected from data memory (not shown). More particularly, the coefficients for the filters may be initialized with predefined or arbitrary values. The coefficients for the filters are applied to the batch of training datasets D via a forward propagation, and are adjusted via a backward propagation to minimize any output error.
• In application, forwardpredictive model60dinfers the future sequence Ĥ_t+1consisting of k shapes and therefore use the last shape ĥ_t+k+1in the predicted sequence to estimate the position of OSSinterventional device30 in the future time point.
• In an alternative embodiment as shown inFIG. 9F, forwardpredictive model60demploys a many-to-onemodel263bin lieu of sequence-to-sequence model263awhereby the final layer will be the last shape ĥ_t+k+1.
• Referring toFIG. 9C andFIG. 9D, a stage S92eof an exemplaryinterventional procedure90eemploying OSS guidewire332 (FIG. 3F) and shape sensing controller103 (FIG. 3F) and encompassesshape sensing controller103 measure and storing a shape of OSS guidewire332 during a navigation of an end-effector340 to a target pose.
• A stage S94eofprocedure90einvolvesshape sensing controller103, forwardpredictive model50dand adisplay controller104.shape sensing controller103 communicates asequence61dof k consecutive shapes to forward predictive model60eto thereby inf infers the following sequence of shapes Ĥ_t+1, where the last shape ĥ_t+k+1is the predicted position ofOSS guidewire332.Display controller104 controls a display of a sensedposition image105aof OSS guidewire332 for guidance purposes to the end-effector340 to the target pose.
• Referring toFIGS. 11A and 11B, an imagingpredictive model80aof the present disclosure is trained on expert data with reasonable coverage of different cases. Thisneural network80a, given anultrasound image81a, infers therelative position82aof thisimage81ain respect to the reference anatomy. In result, a motion between a former pose of an end-effector and a current pose of an end-effector in the anatomy (e.g., a heart) may be calculated as will be further described herein.
• FIG. 11A shows a batchwise training of the imagingpredictive model80a. During training, network constantly imagingpredictive model80aupdates its weights using 2-tuples from the ground truth dataset consisting of ultrasound image and relative position of this image in respect to the reference anatomy.
• FIG. 11B shows a real-time inference using imagingpredictive model80athat, given animage81a, predicts therelative pose82aof the end-effector n respect to the reference anatomy, (e.g., a reference ultrasound image).
• In a training phase, a data acquisition controller (not shown) is configured to receive and interpret information from both a robot and end-effector (e.g., an ultrasound device), and save data on a data storage media (not shown) by a data acquisition controller (not shown) in a format defined by the following specifications: training dataset D consisting of i data points represented by a 2-tuple: d_i=(U_i, T_i). This 2-tuple consists of anultrasound image U_i81athat was acquired at a certain position T∈SE(3)82ain respect to the reference positon.
• A training controller is configured to interpret the training dataset D saved on the data storage media. This dataset D consists of i data points represented by a 2-tuple: d_i=(U_i, T_i). This 2-tuple consists of:ultrasound image U_i81aof the anatomy andrelative motion T_i82abetween current pose of the end-effector at which ultrasound image U_iwas acquired and some arbitrarily chosen reference position.
• In one embodiment as shown inFIG. 11E, imagepredictive model80aemploys aneural network base180aincluding an input layer, hidden layers and an output layer derived from a combination of one or more fully connected layers (FCL)183a, one or more convolutional layers (CNL)184a, one or more recurrent layers (RCL)185a, and one or more long term short memory (LSTM) layers186a.
• In practice, the combination of layers is configured to implement a relative positioning of image U_Cto a reference image to thereby pose {circumflex over (T)}.
• In one embodiment for implementing a relative positioning of image U_Cto a reference image to thereby pose {circumflex over (T)},neural network base180aincludes a set of N number of fully connectedlayers183a.
• In a second embodiment for implementing a relative positioning of image U_Cto a reference image to thereby pose {circumflex over (T)},neural network base180aincludes a set of N convolutional layers184afollowed by either a set of M fully connectedlayers183aor a set of Wrecurrent layers185aor a set of W long term short memory layers186a.
• In a third embodiment for implementing a relative positioning of image U_Cto a reference image to thereby pose {circumflex over (T)},neural network base180aincludes a set of N convolutional layers184afollowed combination of a set of M fully connectedlayers183aand a set of Wrecurrent layers185aor a set W of long term short memory layers186a.
• In practice, a fully connectedlayer183amay include K neurons, where N, M, W, K may be any positive integer, and values may vary depending on the embodiments. For example, N may be about 8, M may be about 2, W may be about 2, and K can be about 1000. Also, aconvolutional layer184amay implement a non-linear transformation, which may be a composite function of operations (e.g., batch normalization, rectified linear units (ReLU), pooling, dropout and/or convolution), and aconvolutional layer184amay also include a non-linearity function (e.g. including rectified non-linear ReLU operations) configured to extract rectified feature maps.
• Further in practice, one of thelayers183aor184aserve as an input layer for inputting image U_C, and one of thelayers183a,185aand186amay serve as an output layer for outputting a pose182aof end-effector40 in Cartesian space (e.g., a translation and a rotation of the end-effector40 in Cartesian space). The outputted pose of end-effector40 in Cartesian space may be represented as vectorial parametrization and/or non-vectorial parametrization of a rigid-body position and orientation. More particularly, the parametrizations may be in the form of Euler angles, quaternions, matrix, exponential map, and/or angle-axis representing rotations and/or translations (e.g., including a direction and a magnitude for the translations).
• Also in practice, the output layer may be a non-linear fully connectedlayer183athat gradually shrinks a high-dimensional output of the lastconvolutional layer184aofneural network base180ato produce a set of output variables.
• In training, training weights of imagepredictive model80aare constantly updated by comparing a predicted relative motion of the end-effector ({circumflex over (T)}) in respect to some reference anatomy using image predictive model—given and ultrasound image161cas input—with ground-truth relative motion T from a batch of training datasets D, which may be systematically or randomly selected from data memory (not shown). More particularly, the coefficients for the filters may be initialized with predefined or arbitrary values. The coefficients for the filters are applied to the batch of training datasets D via a forward propagation, and are adjusted via a backward propagation to minimize any output error.
• Referring toFIGS. 12A and 12B, an imagingpredictive model80baof the present disclosure is trained on expert data with reasonable coverage of different cases. Thisneural network80ba, given anultrasound image81a, infers therelative position82aof thisimage81ain respect to the reference anatomy. In result, a motion between a former pose of an end-effector and a current pose of an end-effector in the anatomy (e.g., a heart) may be calculated as will be further described herein.
• FIG. 12A shows a batchwise training of the imagingpredictive model80ba. During training, network constantly imagingpredictive model80baupdates its weights using 2-tuples from the ground truth dataset consisting of ultrasound image and relative position of this image in respect to the reference anatomy.
• FIG. 12B shows a real-time inference using imagingpredictive model80bathat, given animage81a, predicts83alinear velocity and angular velocity of the end-effector with respect to the reference anatomy, (e.g., a reference ultrasound image).
• In a training phase, a data acquisition controller (not shown) is configured to receive and interpret information from both a robot and end-effector (e.g., an ultrasound device), and save data on a data storage media (not shown) by a data acquisition controller (not shown) in a format defined by the following specifications: training dataset D consisting of i data points represented by a 2-tuple: d_i=(U_i, V_i). This 2-tuple consists of anultrasound image U_i81athat was acquired at a certain position T∈SE(3)82ain respect to the reference position via the-vector83aof linear velocity and angular velocity of the end-effector.
• A training controller is configured to interpret the training dataset D saved on the data storage media. This dataset D consists of i data points represented by a 2-tuple: d_i=(U_i, V_i). This 2-tuple consists of:ultrasound image U_i81aof the anatomy andrelative n vector83aof linear velocity and angular velocity of the end-effector at which ultrasound image U_iwas acquired and some arbitrarily chosen reference position.
• In one embodiment as shown inFIG. 12E, imagepredictive model80bemploys aneural network base180bincluding an input layer, hidden layers and an output layer derived from a combination of one or more fully connected layers (FCL)183b, one or more convolutional layers (CNL)184b, one or more recurrent layers (RCL)185b, and one or more long term short memory (LSTM) layers186b.
• In practice, the combination of layers is configured to implement a relative positioning of image U_Cto a reference image to thereby derive a linear velocity and/or an angular velocity of end-effector40.
• In one embodiment for implementing relative positioning of image U_Cto a reference image to thereby derive a linear velocity and/or an angular velocity of end-effector40,neural network base180bincludes a set of N number of fully connectedlayers183b.
• In a second embodiment for implementing relative positioning of image U_Cto a reference image to thereby derive a linear velocity and/or an angular velocity of end-effector40,neural network base180bincludes a set of N convolutional layers184bfollowed by either a set of M fully connectedlayers183bor a set of Wrecurrent layers185bor a set of W long termshort memory layers186b.
• In a third embodiment for implementing relative positioning of image U_Cto a reference image to thereby derive a linear velocity and/or an angular velocity of end-effector40,neural network base180bincludes a set of N convolutional layers184bfollowed combination of set of M fully connectedlayers183band a set of Wrecurrent layers185bor a set W of long termshort memory layers186b.
• In practice, a fully connectedlayer183bmay include K neurons, where N, M, W, K may be any positive integer, and values may vary depending on the embodiments. For example, N may be about 8, M may be about 2, W may be about 2, and K can be about 1000. Also, aconvolutional layer184bmay implement a non-linear transformation, which may be a composite function of operations (e.g., batch normalization, rectified linear units (ReLU), pooling, dropout and/or convolution), and aconvolutional layer184bmay also include a non-linearity function (e.g. including rectified non-linear ReLU operations) configured to extract rectified feature maps.
• Further in practice, one of thelayers183bor184bserve as an input layer for inputting image U_C, and one of thelayers183b,185band186bmay serve as an output layer for outputting a linear and angular velocity of the end effector
• $\left[\begin{array}{c}\hat{v}\\ \hat{\omega }\end{array}\right]\hspace{1em}$
• as regressed from last fully-connected layer (e.g. 6 units, 3 units for linear and 3 units for angular velocity) with linear or non-linear activation function.
• In training, training weights of imagepredictive model80bare constantly updated by comparing a predicted linear and angular velocity of the end-effector in respect to some reference anatomy—given ultrasound image161cas input—with ground-truth end-effector linear and angular velocity describing motion to some reference anatomy from a batch of training datasets D, which may be systematically or randomly selected from data memory (not shown). More particularly, the coefficients for the filters may be initialized with predefined or arbitrary values. The coefficients for the filters are applied to the batch of training datasets D via a forward propagation, and are adjusted via a backward propagation to minimize any output error.
• Referring toFIGS. 13A and 13B, acontinuous positioning controller50femploys inversepredictive model70a, imagepredictive model80a, asubtractor53 and acontrol law54ato execute a closed loop continuous position control method of the present disclosure as represented byflowchart190a.
• In one TEE probe embodiment, stage S192aofprocedure190aencompasses TEE probe handle132 (FIG. 3A) being inserted into the robotic controller100 (FIG. 3B) as known in the art of the present disclosure to controls the dials ofhandle132 as well as a rotation ofTEE probe130. A TEE transducer140 (FIG. 3A) is inserted into the body through the esophagus and positioned in the vicinity of the anatomical structure of interest, such as heart, for instance at a mid-esophageal position as shown inFIG. 13C. Ultrasound image parameters are defined at this target pose ofTEE transducer140.
• A stage S194aofprocedure190aencompasses imagepredictive model90 processing acurrent ultrasound image81a, to which we will refer as former ultrasound image U_f, to predict relative position of this image plane in respect to the reference anatomy^ref{circumflex over (T)}_f. Sonographer observes the anatomy using ultrasound image and decides about the desired movement of the transducer from its current position T_d. Alternatively, desired motion of the transducer could be provided from: external tracking devices, user interfaces, other imaging modalities that are registered to the ultrasound image, such as X-ray images that are registered to 3D TEE images using EchoNavigator (Philips).
• Based on T_d, inversepredictive model70apredicts joint variables {circumflex over (q)}_t72athat are required to move the robot to the desired position.Robot controller100 receives the joint variable72as and movesTEE probe130 accordingly.
• A stage S196aofprocedures190aencompasses the ultrasound transducer reaching another position at which ultrasound image U_cis acquired. Image predictive model90guses process the current ultrasound image U_cto predict relative position of both current image plane in respect to the reference anatomy^ref{circumflex over (T)}_c. In result a motion between former and current position in the anatomy, such as heart, can be calculated as follows:
• ${}_{}{}^c\hat{T}_f^{}={\left({}_{}{}^{\mathrm{ref}}\hat{T}_c^{}\right)}^{-1}{}_{}{}^{\mathrm{ref}}\hat{T}_f^{}$
• In a second TEE probe embodiment, as shown inFIG. 13D, a selection of a path on animage201 generated by an external imaging modality registered to the ultrasound image may be utilized to determine the desired position T_d(e.g., X-ray image, as well as cone-beam CT image can be registered to the ultrasound image using methods known in the art (Philips EchoNavigator)).
• Referring toFIGS. 14A and 14B, acontinuous positioning controller50gemploys inversepredictive model70b, imagepredictive model80b, asubtractor53 and acontrol law54bto execute a closed loop continuous position control method of the present disclosure as represented byflowchart190b.
• In one TEE probe embodiment, stage S192bofprocedure190bencompasses TEE probe handle142 (FIG. 3A) being inserted into the robotic controller100 (FIG.3B3) as known in the art of the present disclosure to controls the dials of handle142 as well as a rotation ofTEE probe140. A TEE transducer140 (FIG. 3A) is inserted into the body through the esophagus and positioned in the vicinity of the anatomical structure of interest, such as heart, for instance at a mid-esophageal position as shown inFIG. 14C. Ultrasound image parameters are defined at this target pose ofTEE transducer140
• A stage S194bofprocedure190bencompasses a user desiring to move the transducer in the image space for instance by selecting the path from point A to point B inultrasound image203 or a transformation between image plane A and B. In this embodiment, a first linear andangular velocity202 defined by the path on the image is transformed to the end effector coordinate system using^end-effectorJ_imageJacobian204, which is calculated by knowing the spatial relationship between end effector and image coordinate system using methods known in art of the present disclosure.
• Based on V_d, inversepredictive model70b(FIG.7B3) predicts joint velocities {dot over ({circumflex over (q)})}_t72bthat are required to move the robot to the desired position.Robot controller100 receives the Joint velocities {dot over ({circumflex over (q)})}_t72band movesTEE probe130 accordingly.
• A stage S196bofprocedures190aencompasses the ultrasound transducer reaching another position at which ultrasound image U_cis acquired. Imagepredictive model80bprocess the currentultrasound image U_c81ato predict avelocity vector83aof the end-effector between points A and B. The imagepredictive model80bestimating the function between Cartesian velocity of the end effector and the velocities in the joint space, i.e. neural network models a manipulator Jacobian —given a 6-vector consisting of the linear and angular velocities of the end effector71cto predict n-vector72cof joint velocities.
• As understood by those skilled in the art of the present disclosure, neural networks that model spatial relationships between images of the anatomy, such as heart, require large training dataset, which are specific to given organ.
• In an alternative embodiment, features are directly extracted from the images in order to validate the position of the transducer. In this embodiment as shown inFIG. 14C, a user selects the object on theimage205, e.g. apical wall, around which system will extract certain salient features via afeature extraction206. These features, which may include edges, shape, and size of the object, are first detected using methods know in the art of the present disclosure (e.g. Canny edge detector, morphological operation, etc.), and second tracked using scale invariant feature transform (SIFT) that is known in the art of the present disclosure. Finally system defines a path between selected object and the center of the field of view, which indicates the required motion on the image. By tracking the salient features using SIFT,continuous position controller50g(FIG. 14A) can correct the predictions from the network within the close-loop control.
• More particular, a velocity-based control system of the continuum-like robot. Desired motion on the image is identified as soon as the user selects certain object on the image, e.g. apical wall (see red dot on the ultrasound image). Motion is defined by a path between center of the field of view and the selected object, which can be transformed to linear and angular velocities in the end effectorspace using Jacobian204. This Cartesian velocity is then sent to the neural network which will infer joint velocities. Achieved position will be iteratively validated against the path defined by a constantly tracked object, and center of the field of view.
• In practice, the closed control loop ofFIGS. 13 and 14 may closed using other modalities, such as optical shape sensing (OSS), electromagnetic tracking, or X-ray images registered to the ultrasound images using for instance Philips EchoNavigator.
• Also in practice, prediction accuracy of the neural network can be affected by the configuration of the flexible endoscope. Thus, a position of a transducer in respect to heart is first defined, using for instance neural network g, or Philips HeartModel, which will implicitly define one of the possible configurations. Second, a certain set of network weights will be loaded into the models according to the detected configuration, thus improving the prediction accuracy.
• Similar approach can be used to guide the user to the location at which the most optimal images and guidance can be provided.
• Furthermore, one of the hardest issues in machine/deep learning is the accessibility of large data in the right format for training of the predictive models. More particularly, collecting and constructing the training and validation sets is very time consuming and expensive because it requires domain-specific knowledge. For instance, to train a predictive model to accurately differentiate a benign breast tumor and a malignant breast tumor, such training needs several thousands of ultrasound images annotated by expert radiologists and transformed into a numerical representation that a training algorithm can understand. Additionally, the image datasets might be inaccurate, corrupted or labelled with noise, all leading to detection inaccuracies, and an acquisition of large medical datasets may trigger ethical and privacy concerns, and many others concerns.
• Referring toFIG. 15, a training data collection system of the present disclosure employs ashape sensing controller103 providing3D position vectors233 for each point on the opticalshape sensing fiber332 that is mounted to bothendoscope131 and tendon-drivenmanipulator231 of the robot. A distal part of theoptical fiber332 is embedded in the plastic case which rigidly connects optical fiber to the end-effector and introduce a certain curvature within the shape.
• For example,FIGS. 16A and 16B show adistal end332dof an opticalshape sensing fiber332 embedded in theplastic casing350 that is rigidly attached to anultrasound transducer232 ofmanipulator231 as shown inFIGS. 16C and 16D. Theplastic casing350 reinforces a certain curvature in the shape thus enabling estimation of the end-effector pose using template matching algorithm as known in the art of the present disclosure (e.g., adata acquisition sequence370 as shown inFIG. 17 with an alpha and beta on each axis corresponds to knobs positions on the TEE handle132 shown inFIG. 15).
• Referring back toFIG. 16, generally, by detecting this pattern one can estimate the pose T∈SE(3) of the end-effector using methods known in art. Arobot controller100 sends motion commands to the robot, which controls actuated knobs responsible for pulling or loosening the tendons. By altering the state of the tendons, the position of the end-effector is changed as shown on inFIGS. 3D and 3E. Adata storage controller190 receives shape h∈(p₁. . . p_n) of optical fiber, pose T of the end-effector as well as motion commands, i.e. joint positions q_t, from shape sensing and robot controller respectively. Data is stored as a 3-tuple on the storage device and used later on for training of the deep convolutional neural network of the present disclosure
• More particular, a shape-sensedguidewire332 is embedded or attached to the continuum robot, which uses an optical shape sensing (OSS) technology known in the art. OSS uses light along a multicore optical fiber for device localization and navigation during surgical intervention. The principle involved makes use of distributed strain measurements in the optical fiber using characteristic Rayleigh backscatter or controlled grating patterns.
• Shape sensing controller103 is configured to acquire the shape of shape-sensed guidewire322, estimate the pose T∈SE(3) of the end-effector that is rigidly attached to aplastic casing350, which enforces a certain curvature in the guidewire as previously describe herein. The methods for estimating pose T based on a well-defined curvature as well as template matching algorithm as known in the art of the present disclosure.
• Data acquisition controller191 is configured to generate a sequence of motor commands according to pre-defined acquisition pattern (e.g., a spiral, radial, or square motion, etc.) and send movement commands torobot controller100.
• Robot controller100 is configured to receive robot position and send movement signals to the robot. Using motorized knobs robot will pull/loosen the tendons, which will result in the motion of the end-effector. Robot controller is further configured to receive and interpret information from thedata acquisition controller191, and change the robot position based on the information from thedata acquisition controller191.
• Data storage controller190 is configured to receive and interpret information from bothrobot controller100 andshape sensing controller103, and save data on the data storage media (not shown) in a format defined by the following specifications: First specification is to acquire training dataset D for all configurations pre-defined by thedata acquisition controller191. A dataset D consists of a set of n sequences W: D={W₁, W₂, . . . , W_n}, each sequence W_nconsists of i data points d_i: W_n={d₁, d₂, . . . , d_i}; and each data point d_ifrom the sequence is defined by a 3-tuple: d_i=(T_i, H_i, Q_i)
• The 3-tuple consists of end-effector pose T∈SE(3), sequence of k consecutive shapes such as H∈(h_t, h_t+1. . . h_t+k), where h∈(p₁. . . p_m) is a set of m vectors p_mthat describe both the position of the shape-sensed guidewire in 3D Euclidean space and auxiliary shape parameters such as strain, curvature, and twist, and a sequence of j consecutive joint variables Q∈(q_t, q_t+1. . . q_t+j) acquired at time points starting from t to t+j. For instance, entry q_tcould be an angle on the control knobs α, β acquired at a time point t.
• Referring toFIG. 19, a trainingdata collection method360 of the present disclosure is executed by the training data collection system ofFIG. 15.
• Referring to bothFIGS. 15 and 19, a stage S362 ofmethod360 encompasses the robot being moved to a home position by arobot controller100 using for instance limit switches or proximity sensors. A distal part of the shape-sensedguidewire232 is inserted in theindent353 provided in theplastic casing350. Thisindent353 will enforce a certain curvature in the shape.
• Casing350 is rigidly attached to the end-effector of the continuum-like robot.
• By using template matching algorithm as known in the art of the present disclosure during a stage S364 ofmethod360,shape sensing controller103 can now estimate the pose T∈SE(3) of the end-effector. Preferably, the coordinate system of the end-effector is defined by the template, however additional calibration matrixes can be used. When robotic system is still in the home position, pose of the end-effector is acquired in the OSS coordinate system. Every following pose that is acquired during the experiment is estimated relative to this initial position.
• Data acquisition controller191 generates a motion sequence, i.e. set of joint variables, according to pre-defined acquisition pattern (e.g.,pattern370 ofFIG. 18). Motion sequence is sent iteratively bydata acquisition controller191 to therobot controller100 that moves the robot according to the generated joint variables.
• Stage S366 of method300 encompasses, at each time point, an acquisition and storage of a data tuple d_i=(T_i, H_i, Q_i) thedata storage controller190. Of importance, because H_iand Q_iare sequential, all former time points are kept in the memory by thedata storage controller190.
• To facilitate a further understanding of the various inventions of the present disclosure, the following description ofFIG. 19 teaches an exemplary embodiment of a continuous positioning controller of the present disclosure. From this description, those having ordinary skill in the art will appreciate how to apply various aspects of the present disclosure for making and using additional embodiments of continuous positioning controller of the present disclosure.
• Referring toFIG. 19, acontinuous positioning controller400 includes one or more processor(s)401,memory402, a user interface403, anetwork interface404, and a storage405 interconnected via one or more system buses406.
• Eachprocessor401 may be any hardware device, as known in the art of the present disclosure or hereinafter conceived, capable of executing instructions stored inmemory402 or storage or otherwise processing data. In a non-limiting example, the processor(s)401 may include a microprocessor, field programmable gate array (FPGA), application-specific integrated circuit (ASIC), or other similar devices.
• Thememory402 may include various memories, e.g. a non-transitory and/or static memory, as known in the art of the present disclosure or hereinafter conceived, including, but not limited to, L1, L2, or L3 cache or system memory. In a non-limiting example, thememory402 may include static random access memory (SRAM), dynamic RAM (DRAM), flash memory, read only memory (ROM), or other similar memory devices.
• The user interface403 may include one or more devices, as known in the art of the present disclosure or hereinafter conceived, for enabling communication with a user such as an administrator. In a non-limiting example, the user interface may include a command line interface or graphical user interface that may be presented to a remote terminal via thenetwork interface404.
• Thenetwork interface404 may include one or more devices, as known in the art of the present disclosure or hereinafter conceived, for enabling communication with other hardware devices. In a non-limiting example, thenetwork interface404 may include a network interface card (NIC) configured to communicate according to the Ethernet protocol. Additionally, thenetwork interface404 may implement a TCP/IP stack for communication according to the TCP/IP protocols. Various alternative or additional hardware or configurations for thenetwork interface404 will be apparent.
• The storage405 may include one or more machine-readable storage media, as known in the art of the present disclosure or hereinafter conceived, including, but not limited to, read-only memory (ROM), random-access memory (RAM), magnetic disk storage media, optical storage media, flash-memory devices, or similar storage media. In various non-limiting embodiments, the storage405 may store instructions for execution by the processor(s)401 or data upon with the processor(s)401 may operate. For example, the storage405 may store a base operating system for controlling various basic operations of the hardware. The storage405 also stores application modules in the form of executable software/firmware for implementing the various functions of the controller400aas previously described in the present disclosure including, but not limited to, forward predictive model(s)60, inverse predictive model(s)70 and imaging predictive model(s)80 as previously described in the present disclosure.
• In practice,controller400 may be installed within anX-ray imaging system500, an intervention system501 (e.g., an intervention robot system), or a stand-alone workstation502 in communication withX-ray imaging500 system and/or intervention system501 (e.g., a client workstation or a mobile device like a tablet). Alternatively, components ofcontroller400 may be distributed amongX-ray imaging system500,intervention system501 and/or stand-alone workstation502.
• Also in practice, additional controllers of the present disclosure including a shape sensing controller, a data storage controller and a data acquisition controller may also each include one or more processor(s), memory, a user interface, a network interface, and a storage interconnected via one or more system buses as arranged inFIG. 19, whereby the storage contains applicable application modules of that controller as previously described herein. Alternatively, two or more controllers of the present disclosure may be integrated as a single controller, where the storage contains applicable application modules of the two or more controllers as previously described herein.
• Referring toFIGS. 1-19, those having ordinary skill in the art of the present disclosure will appreciate numerous benefits of the present disclosure.
• Further, as one having ordinary skill in the art will appreciate in view of the teachings provided herein, structures, elements, components, etc. described in the present disclosure/specification and/or depicted in the Figures may be implemented in various combinations of hardware and software, and provide functions which may be combined in a single element or multiple elements. For example, the functions of the various structures, elements, components, etc. shown/illustrated/depicted in the Figures can be provided through the use of dedicated hardware as well as hardware capable of executing software in association with appropriate software for added functionality. When provided by a processor, the functions can be provided by a single dedicated processor, by a single shared processor, or by a plurality of individual processors, some of which can be shared and/or multiplexed. Moreover, explicit use of the term “processor” or “controller” should not be construed to refer exclusively to hardware capable of executing software, and can implicitly include, without limitation, digital signal processor (“DSP”) hardware, memory (e.g., read only memory (“ROM”) for storing software, random access memory (“RAM”), non-volatile storage, etc.) and virtually any means and/or machine (including hardware, software, firmware, combinations thereof, etc.) which is capable of (and/or configurable) to perform and/or control a process.
• Moreover, all statements herein reciting principles, aspects, and embodiments of the invention, as well as specific examples thereof, are intended to encompass both structural and functional equivalents thereof. Additionally, it is intended that such equivalents include both currently known equivalents as well as equivalents developed in the future (e.g., any elements developed that can perform the same or substantially similar function, regardless of structure). Thus, for example, it will be appreciated by one having ordinary skill in the art in view of the teachings provided herein that any block diagrams presented herein can represent conceptual views of illustrative system components and/or circuitry embodying the principles of the invention. Similarly, one having ordinary skill in the art should appreciate in view of the teachings provided herein that any flow charts, flow diagrams and the like can represent various processes which can be substantially represented in computer readable storage media and so executed by a computer, processor or other device with processing capabilities, whether or not such computer or processor is explicitly shown.
• Having described preferred and exemplary embodiments of the various and numerous inventions of the present disclosure (which embodiments are intended to be illustrative and not limiting), it is noted that modifications and variations can be made by persons skilled in the art in light of the teachings provided herein, including the Figures. It is therefore to be understood that changes can be made in/to the preferred and exemplary embodiments of the present disclosure which are within the scope of the embodiments disclosed herein.
• Moreover, it is contemplated that corresponding and/or related systems incorporating and/or implementing the device/system or such as may be used/implemented in/with a device in accordance with the present disclosure are also contemplated and considered to be within the scope of the present disclosure. Further, corresponding and/or related method for manufacturing and/or using a device and/or system in accordance with the present disclosure are also contemplated and considered to be within the scope of the present disclosure.

Claims (14)

1. A positioning controller for an interventional device including a device portion, the positioning controller comprising:

a memory including at least one of:

a forward predictive model configured with embedded kinematics of interventional device to receive commanded positioning motion of the interventional device and to output data related to a prediction of a navigated pose of the device portion, and

a control predictive model configured with kinematics of interventional device to receive a target pose data of the interventional device and to output data related to a prediction of a positioning motion of the interventional device; and

at least one processor in communication with the memory, wherein the at least one processor is configured to at least one of:

(i) apply the forward predictive model to a commanded positioning motion of the interventional device to render a predicted navigated pose of the device portion, and generate positioning data informative of a positioning by the interventional device of the device portion to a target pose based on the predicted navigated pose of the device portion; and

(ii) apply the control predictive model to the target pose of the device portion to render a predicted positioning motion of the interventional device, and generate positioning commands controlling a positioning of the interventional device of the device portion to the target pose based on the predicted positioning motion of the interventional device.

2. The positioning controller ofclaim 1, wherein the imaging predictive model has been or is trained on forward kinematics of the interventional device, and/or the control predictive model is an inverse predictive model which has been or is trained on inverse kinematics of the interventional device.

3. The positioning controller ofclaim 1, wherein the device portion is an end-effector of the interventional device.

4. The positioning controller ofclaim 1, further configured to generate said positioning data continuously and/or said positioning commands continuously such that the positioning controller is considered as a continuous positioning controller.

5. The positioning controller ofclaim 1, wherein the forward predictive model includes:

a neural network base having an input layer configured to input joint variables of the interventional robot representative of the commanded positioning motion of the interventional device, and an output layer configured to output at least one of a translation, a rotation and a pivoting of the device portion derived from a regression of the joint variables of the interventional robot,

wherein the at least one of the translation, the rotation and the pivoting of the device portion infers the predicted navigated pose of the device portion.

6. The positioning controller ofclaim 1, wherein the control predictive model includes:

a neural network base having an input layer configured to input at least one of a translation, a rotation and a pivoting of the device portion, and an output layer configured to output joint variables of the interventional robot derived from a regression at least one of the translation, the rotation and the pivoting of the device portion,

wherein the joint variables of the interventional robot infer the predicted positioning motion of the interventional device.

7. The positioning controller ofclaim 1, wherein the forward predictive model includes:

a neural network base having an input layer configured to input joint velocities of the interventional device representative of the commanded positioning motion of the interventional device, and an output layer configured to output at least one of a linear velocity and an angular velocity of the device portion from a regression of the joint velocities of the interventional device,

wherein at least one of a linear velocity and an angular velocity of the device portion infers the predicted navigated pose of the interventional device.

8. The positioning controller ofclaim 1, wherein the control predictive model includes:

a neural network base having an input layer configured to input at least one of a linear velocity and an angular velocity of the device portion to the target pose and an output layer configured to output joint velocities of the interventional device from a regression at least one of a linear velocity and an angular velocity of the device portion to the target pose,

wherein the joint velocities of the interventional device infer the predicted positioning motion of the interventional device.

9. The positioning controller ofclaim 1, wherein the forward predictive model includes:

a neural network base having an input layer configured to input a preceding sequence of shapes of the interventional device representative of the commanded positioning motion of the interventional device, and an output layer configured to output a succeeding sequence of shapes of the interventional device derived from a time series prediction of the preceding sequence of shapes of the interventional device,

wherein the succeeding sequence of shapes of the interventional device infer the predicted navigated pose of the device portion.

10. The positioning controller ofclaim 1, wherein the forward predictive model includes:

a neural network base having an input layer configured to input a preceding sequence of shapes of the interventional device representative of the commanded positioning motion of the interventional device, and an output layer configured to output a succeeding shape of the interventional device derived from a time series prediction of the preceding sequence of shapes of the interventional device,

wherein the succeeding shape of the interventional device infers the predicted navigated pose of the device portion.

11. The positioning controller ofclaim 2,

wherein at least one of:

the forward predictive model is further trained on at least one navigation parameter of the interventional device auxiliary to the forward kinematics of the interventional device predictive of the pose of the device portion, and

the control predictive model is further trained on the at least one navigation parameter of the interventional device auxiliary to the inverse kinematics of the interventional device predictive of the positioning motion of the interventional device; and

wherein the at least one processor is configured to at least one of:

(i′) apply the forward predictive model to both the commanded positioning motion of the interventional device and the at least one navigation parameter auxiliary to the forward kinematics of the interventional device to render the predicted navigated pose of the device portion; and

(ii′) apply the control predictive model to both the target pose of the device portion and the at least one navigation parameter auxiliary to the inverse kinematics of the interventional device to render the predicted positioning motion of the interventional device.

12. The positioning controller ofclaim 1,

wherein at least one of:

the forward predictive model is configured to further receive at least one auxiliary navigation parameter of the interventional device, and further process it to output the prediction of navigated pose of the device portion, and

the control predictive model is configured to further receive at least one auxiliary navigation parameter of the interventional device, and further process it to output the prediction of positioning motion of the interventional device; and

wherein the at least one processor is configured to at least one of:

(i′) apply the forward predictive model to both the commanded positioning motion of the interventional device and at least one auxiliary navigation parameter to render the predicted navigated pose of the device portion; and

(ii′) apply the control predictive model to both the target pose of the device portion and at least one auxiliary navigation parameter to render the predicted positioning motion of the interventional device.

13. A machine-readable storage medium, encoded with instructions for execution by at least one processor, to instruct an international device including a device portion, the machine-readable storage medium storing:

at least one of:

a forward predictive model configured with kinematics of interventional device, to receive commanded positioning motion of the interventional device and to output data related to a prediction of a navigated pose of the device portion, and

an control predictive model configured with kinematics of interventional device to receive target pose data of the interventional device and to output data related to a prediction of a positioning motion of the interventional device; and

instruction to at least one of:

(i) apply the forward predictive model to a commanded positioning motion of the interventional device to render a predicted navigated pose of the device portion, and generate positioning data informative of a positioning by the interventional device of the device portion to a target pose based on the predicted navigated pose of the device portion; and

(ii) apply the control predictive model to the target pose of the device portion to render a predicted positioning motion of the interventional device, and generate positioning commands controlling a positioning of the interventional device of the device portion to the target pose based on the predicted positioning motion of the interventional device.

14. A positioning method executable by a positioning controller for an interventional device including a device portion,

the positioning controller storing at least one of:

a forward predictive model configured with embedded kinematics of interventional device to receive commanded positioning motion of the interventional device and to output data related to a prediction of a navigated pose of the device portion, and

an control predictive model configured with kinematics of interventional device to receive a target pose data of the interventional device and to output data related to a prediction of a positioning motion of the interventional device,

wherein the positioning method comprising the position controller executing at least one of:

(i) applying the forward predictive model to a commanded positioning motion of the interventional device to render a predicted navigated pose of the device portion, and generating positioning data informative of a positioning by the interventional device of the device portion to a target pose based on the predicted navigated pose of the device portion; and

(ii) applying the control predictive model to the target pose of the device portion (40) to render a predicted positioning motion of the interventional device, and generating positioning commands controlling a positioning of the interventional device of the device portion (40) to the target pose based on the predicted positioning motion of the interventional device.

Images