Underwater multi-robot obstacle avoidance device and method under communication connectivity maintenance constraintTechnical Field
The invention relates to the field of underwater multi-robot cluster system control, in particular to an underwater multi-robot obstacle avoidance device and method under the constraint of communication connectivity.
Background
The 21 st century is the century of the ocean, and along with the development of economy and the progress of scientific technology, the intellectualization of ocean exploration equipment is a future development trend. The underwater multi-robot cluster system has group cooperation advantages in aspects of large-scale sea exploration, sea search and rescue, tracking and trapping, and the like, but communication problems and control problems of the underwater multi-robot cluster system are caused while the efficiency is improved, and the underwater environment is different from the ground, so that the underwater multi-robot cluster system has the specificity. First, compared with the ground, the underwater communication means are limited and are easily affected by water temperature, light, noise and the like, and the unstable communication connection can cause the cruising task of the underwater multi-robot cluster system to fail. Secondly, the underwater environment is complex and changeable and is full of various uncertain factors, such as submarine reefs, corals and fish shoals, so that the underwater multi-robot cluster system has a challenge of collision-free movement. These factors present significant difficulties in the cruising task of the multiple-robot cluster system under water.
In the prior art, the Chinese patent with the publication number of CN115185287A discloses an intelligent multi-underwater robot dynamic obstacle avoidance and capture control system, which is used for sending the capture target position, moving direction and speed to other robots in real time through a communication device after a certain underwater robot finds a target, and capturing the target through an enclosure range calculated in advance; according to the scheme, the dynamic trapping of the target can be realized through the cooperation of the multiple underwater robots, but communication connectivity between the scheme and other underwater robots is not considered when target information is broadcast to other robots, the marine environment is complex and changeable, communication connection interruption is caused when communication connectivity constraint is not considered, and further the task failure of the crowd collecting system of the multiple underwater machines is caused.
Furthermore, the Chinese patent with publication number CN103529844A discloses an underwater robot obstacle avoidance method based on forward looking sonar, which introduces forward looking sonar image data into a robot obstacle avoidance strategy, so that the collision avoidance blind area of the robot is reduced. The method can provide long-distance and high-resolution underwater images for the underwater robot, but when the underwater robot is in a near-distance complex obstacle scene, the sonar resolution is low due to the influence of underwater noise and multipath effects, so that the obstacle avoidance performance of the robot is reduced.
Aiming at the defects, how to design a communication connectivity maintaining constraint to improve the communication quality of the multi-underwater robot and design a multi-underwater robot obstacle avoidance method and device under the constraint is particularly important.
Disclosure of Invention
The invention aims to solve the technical problem of providing the underwater multi-robot obstacle avoidance device and the method under the communication connectivity maintenance constraint, which can realize stable robot cluster system obstacle avoidance control under the communication connectivity maintenance constraint and improve the system communication and motion stability.
In order to solve the technical problems, the invention adopts the following technical scheme:
an underwater multi-robot obstacle avoidance device under communication connectivity maintenance constraint, wherein each underwater robot comprises an underwater robot main body, an underwater sound wireless communication device and a binocular vision obstacle avoidance device;
The underwater robot main body comprises a robot supporting body forming a robot main body frame, a power system comprising six propeller modules, 4 buoyancy materials symmetrically fixed on the front side and the rear side of the robot main body frame in pairs, a control cabin fixedly arranged in the center of the robot main body frame, a battery cabin internally provided with a power supply system and 2 underwater searchlights fixedly arranged at the bottom of the front end of the robot main body frame;
The underwater sound wireless communication device comprises an underwater sound transducer, a modem, a communication module, a singlechip microprocessor and a lithium battery power supply system, wherein the underwater sound transducer is used for sending and receiving communication high-frequency ultrasonic waves, the modem is used for converting analog signals into digital signals, the communication module is used for carrying out data exchange between the singlechip microprocessor and the modem, the singlechip microprocessor is used for processing data information sent by the modem, and the lithium battery power supply system is used for supplying power to the underwater sound wireless communication device;
The binocular vision obstacle avoidance device is composed of a first monocular camera, a second monocular camera, a third monocular camera and a fourth monocular camera, wherein the first monocular camera, the second monocular camera, the third monocular camera and the fourth monocular camera are respectively and fixedly installed right and left, right and above and right below the front part of the robot main body frame and used for capturing optical images of an underwater environment in real time.
The technical scheme of the invention is further improved in that the robot bearing body comprises a first bearing body, a second bearing body and a third bearing body, wherein the first bearing body and the second bearing body form an outer side frame of a robot main body frame, and the third bearing body forms the bottom of the robot main body frame;
Six propeller modules contained in the power system specifically refer to two ascending/descending propellers fixedly arranged on the left side and the right side of the control cabin body and four advancing/retreating propellers which are fixed on a first supporting body and a second supporting body below a buoyancy material at an included angle of 45 degrees with the horizontal direction, wherein the buoyancy material is positioned on the front side and the rear side of the ascending/descending propellers;
The control cabin body comprises a motor driving module, a singlechip microprocessor unit and a microcomputer image processing unit, and is fixedly arranged at the middle part of the third supporting body.
An underwater multi-robot obstacle avoidance method under communication connectivity maintenance constraint comprises the following steps:
Step 1, acquiring an environment image of an underwater cruising task area and an image of an underwater robot, preprocessing an area of the acquired image, which needs obstacle avoidance, and generating a corresponding data set, performing offline training on the data set through a deep convolutional neural network, and deploying a trained model to a microcomputer image processing unit in each underwater robot control cabin;
Step 2, each underwater robot is deployed to a cruising task area, the underwater robot captures the information of the environmental image in real time through a binocular camera carried by the underwater robot, a microcomputer image processing unit judges whether the image has an obstacle avoidance area through a deployed neural network model, if the image has the obstacle avoidance area, the step 3 is carried out, and if the image does not have the obstacle avoidance area, the step 4 is carried out;
Step 3, acquiring the coordinates of a pixel point at the center of an obstacle avoidance area on the image, and calculating the parallax angle value of the obstacle avoidance area through the coordinates of the pixel point;
Step 4, respectively solving a neighbor set of each underwater robot based on the communication radius of the underwater acoustic wireless communication device, and then calculating the communication gravitational field value between each underwater robot and other underwater robots in the neighbor set;
Step 5, constructing a reward function of the underwater robot through the parallax angle information obtained in the step 3 and the communication gravitational field obtained in the step 4, constructing a value function based on the reward function, and fitting the value function through a deep reinforcement learning neural network;
And 6, repeating the steps 2 to 5 until the optimal value function is obtained, wherein the deep reinforcement learning neural network is converged, and the deep reinforcement learning neural network is deployed on each underwater robot so as to obtain the optimal control strategy.
In the step 3, if an obstacle avoidance area exists in an image, marking a pixel seat of a central point as (X, Y), wherein X and Y are respectively the horizontal coordinate and the vertical coordinate of the pixel of the central point of the obstacle avoidance area, and calculating the horizontal parallax angle and the vertical parallax angle of the obstacle avoidance area by acquiring the central pixel coordinate of the obstacle avoidance area of the image:
Wherein, θH and θV are respectively the horizontal parallax angle and the vertical parallax angle of the obstacle region, θT is the obstacle avoidance parallax angle threshold, and θA、θB、θC and θD are respectively the horizontal parallax angle and the vertical parallax angle values calculated by the first monocular camera, the second monocular camera, the third monocular camera and the fourth monocular camera according to the central pixel coordinates of the obstacle avoidance region of the image.
The technical scheme of the invention is further improved in that in the step 4, the communication function between the underwater robot Um and the underwater robot Un is as follows:
Wherein, M, n e {1,..M }, Xm＝[xm,ym,zm]T and Xn＝[xn,yn,zn]T respectively represent the positions of the underwater robots Um and Un under the world coordinate system, L (Xm,rv)＝{X∈R2：||X-Xm||≤rv) represents a circular region with a communication radius rv centered on the underwater robot Um, and the neighbor set of the underwater robot Um is:
Pm＝{Xn·fmn(Xm)} (4)
the communication gravitational field generated by the robot Um in the neighbor set is as follows:
Where dmn＝||Xm-Xn l represents the distance between the underwater robot Um and the underwater robot Un, and rs is the maximum stable communication distance.
In step 5, through the parallax angle information of the obstacle and the constraint of the communication gravitational field obtained in the step of the technical scheme, a single-step rewarding function can be constructed to calculate the rewarding value of the strategy at the moment, the larger the rewarding is, the better the control strategy is, and the rewarding function is as follows:
Wherein, Rm(Xm,τm) is a single step prize of the underwater robot Um, τm is a control input of the underwater robot Um at this time, and K1 and K2 are weight coefficients.
In step 6, update the value function based on the single step rewarding function in step 5, the definition of the value function is as follows:
Q(Xmk,τmk)＝Rm(Xmk,τmk)+γ×maxQ(Xmk+1,τmk+1) (7)
Wherein Xmk and τmk are the position and control input of the underwater robot Um in time step k, 0< gamma is less than or equal to 1 and is a discount factor, fitting iterative updating is carried out on a value function through a deep reinforcement learning neural network, steps 2 to 5 are repeated until the convergence requirement of the neural network is met, and at the moment, the optimal control strategy is obtained through the neural network
By adopting the technical scheme, the invention has the following technical progress:
1. The binocular parallax obstacle avoidance method based on binocular vision can realize that the underwater multi-robot cluster system can avoid collision between the underwater robot and the underwater obstacle and other underwater robots when working in a complex obstacle water area, gets rid of the influence of low short-distance resolution and acoustic multipath effect of the traditional acoustic obstacle avoidance sensor, and improves the control stability of the underwater robot.
2. The communication connectivity maintenance constraint scheme provided by the invention improves the communication stability of the underwater multi-robot collaborative work, solves the problem of the disconnection of the underwater robot crowd system, combines with the deep reinforcement learning neural network, and improves the communication connectivity and the control stability of the underwater multi-robot crowd system.
Drawings
For a clearer description of embodiments of the invention or of the solutions of the prior art, the drawings that are used in the description of the embodiments or of the prior art will be briefly described, it being obvious that the drawings in the description below are some embodiments of the invention, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art;
fig. 1 is a schematic perspective view of an underwater robot carrying a binocular vision obstacle avoidance apparatus according to an embodiment of the present invention;
FIG. 2 is a schematic side view of a submerged robot structure according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of an underwater acoustic wireless communications device in accordance with an embodiment of the present invention;
fig. 4 is a diagram illustrating binocular vision parallax detection in an embodiment of the present invention;
FIG. 5 is a schematic diagram of communication area division of a underwater robot in an embodiment of the present invention;
FIG. 6 is a flow chart of a method for multi-robot obstacle avoidance under communication connectivity retention constraints in an embodiment of the invention;
Wherein, 1, buoyancy material, 2, underwater searchlight, 3-1, first carrier, 3-2, second carrier, 4, rising/submerging propeller, 5-1, first monocular camera, 5-2, second monocular camera, 5-3, third monocular camera, 5-4, fourth monocular camera, 6, third carrier, 7, control cabin, 8, battery cabin, 9, advancing/retreating propeller, 10, threading bolt.
Detailed Description
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
The invention is described in further detail below with reference to the attached drawings and examples:
Referring to fig. 1 and 2, and fig. 3 and 4, there is shown an underwater robot obstacle avoidance device under communication connectivity maintenance constraint in an embodiment of the present invention, each underwater robot including an underwater robot body, an underwater acoustic wireless communication device, and a binocular vision obstacle avoidance device;
the underwater robot main body comprises a robot carrier, a power system, a buoyancy material 1, a control cabin 7, a battery cabin 8 and an underwater searchlight 2;
The robot bearing body is a robot main body frame formed by a first bearing body 3-1, a second bearing body 3-2 and a third bearing body 6, wherein the first bearing body 3-1 and the second bearing body 3-2 are vertically arranged in parallel to form an outer side frame of the underwater robot main body frame, and the third bearing body 6 forms the bottom of the underwater robot main body frame, is in vertical relation with the first bearing body 3-1 and the second bearing body 3-2 and is fixedly connected with the first bearing body 3-1 and the second bearing body 3-2;
The power system comprises six propeller modules, specifically, two ascending/descending propellers 4 are fixed on the left side and the right side of a control cabin 7, and four advancing/retreating propellers 9 are fixed below the buoyancy material 1 and on the first supporting body 3-1 and the second supporting body 3-2 at an included angle of 45 degrees with the horizontal direction.
The buoyancy materials 1 are four in number and are symmetrically fixed on the front side and the rear side of the robot main body frame in pairs.
The control cabin body 7 is fixed in the center of the robot main body frame, the inside of the control cabin body comprises a motor driving module, a single-chip microcomputer microprocessor unit and a microcomputer image processing unit, the motor driving module is used for driving a power system to work, the single-chip microcomputer microprocessor is used for receiving signals sent by all sensors and sending control instructions to drive the robot to move, and the microcomputer image processing unit is used for processing environment image information acquired by the binocular vision system in real time.
The battery compartment 8 is fixed at the middle part of the upper side of the third supporting body 6 and is used for assembling a power supply system of the underwater robot.
The underwater searchlight 2 is divided into two parts, is fixed on the inner sides of the bottoms of the first supporting body 3-1 and the second supporting body 3-2, and is connected with the inside of the battery compartment 8 through the threading bolt 10, so that illumination can be provided for the underwater robot in a deep water area.
Referring to fig. 3, a schematic diagram of an underwater acoustic wireless communication device according to an embodiment of the present invention is shown, where the underwater acoustic wireless communication device includes an underwater acoustic transducer, a modem, a communication module, a single-chip microprocessor, and a lithium battery power supply system;
The underwater acoustic transducer is connected with the modem and is used for sending and receiving high-frequency ultrasonic signals;
The modem converts an analog signal transmitted by the underwater acoustic transducer into a digital signal;
The communication module is respectively connected with the modem and the singlechip microprocessor, and the modem transmits digital signals to the singlechip microprocessor through the communication module according to a communication protocol;
the singlechip microprocessor analyzes the data information sent by the communication module according to the communication protocol;
the lithium battery power supply system supplies power to the whole underwater sound wireless communication device.
As shown in fig. 4, which shows a binocular vision parallax detection schematic diagram in the embodiment of the present invention, the binocular vision obstacle avoidance device is formed by four monocular cameras, the first monocular camera 5-1 and the second monocular camera 5-2 are respectively fixed on the outer sides of the middle parts of the first carrier 3-1 and the second carrier 3-2, the third monocular camera 5-3 is fixed on the upper front side of the control cabin 7, the fourth monocular camera 5-4 is fixed on the lower front side of the third carrier 6, the horizontal symmetrical distribution of the first monocular camera 5-1 and the second monocular camera 5-2 forms a horizontal binocular camera, the vertical symmetrical distribution of the third monocular camera 5-3 and the fourth monocular camera 5-4 forms a vertical binocular camera, and the four monocular cameras are used for capturing optical images of the underwater environment in real time, and calculating a horizontal parallax angle and a vertical parallax angle for the obstacle avoidance area respectively.
As shown in fig. 6, an underwater multi-robot obstacle avoidance method under communication connectivity maintenance constraint specifically includes the following steps:
Step 1, combining characteristic information of an underwater obstacle and an underwater robot, cleaning an acquired underwater cruising task area environment image and an underwater robot image, processing a required obstacle avoidance area to generate a data set to be trained, performing off-line training on the data set by building a convolutional neural network model, finishing training when a neural network loss function converges to a set threshold value, and deploying the trained model to a microcomputer image processing unit in an underwater robot control cabin body 7;
And 2, respectively deploying all the underwater robots to a cruising task area, and enabling any robot to normally communicate with all the underwater robots except the robot before cruising tasks are started. The underwater robot captures the information of the image of the environment in real time through a binocular camera carried by the underwater robot, the image is transmitted to a microcomputer image processing unit in a control cabin 7 in real time, the prediction is carried out through the neural network model trained in the step 1, if the image has an area needing to avoid the obstacle, the step 3 is carried out, otherwise, the step 4 is carried out;
Step 3, acquiring coordinates (X, Y) of a central pixel point of the obstacle avoidance area on the image, wherein X and Y are respectively the abscissa and the ordinate of the pixel of the central point of the obstacle avoidance area, and further, calculating the horizontal parallax angle and the vertical parallax angle of the obstacle avoidance area through the coordinates of the central pixel point of the obstacle avoidance area:
Wherein, θH and θV are respectively the horizontal parallax angle and the vertical parallax angle of the obstacle region, θT is the obstacle avoidance parallax angle threshold, and θA、θB、θC and θD are respectively the horizontal and vertical parallax angle values calculated by the first monocular camera 5-1, the second monocular camera 5-2, the third monocular camera 5-3 and the fourth monocular camera 5-4 according to the central pixel coordinates of the image obstacle avoidance region.
Step 4, taking the underwater robot Um as an example, the communication relationship between the underwater robot Un is as follows:
Wherein M, n e { 1..m }, M is the number of all underwater robots contained in the underwater multi-robot cluster system, Xm＝[xm,ym,zm ] T and Xn＝[xn,yn,zn]T represent the positions of the underwater robots Um and Un in the world coordinate system, respectively, L (Xm,rv)＝{X∈R2：||X-Xm||≤rv) represents a circular area with a communication radius rv centered on the position Xm of the underwater robot Um:
Pm＝{Xn·fmn(Xm)} (4)
Further, as shown in fig. 5, a schematic diagram of division of communication areas of the underwater robot in the embodiment of the present invention is shown, in which a circular area with a communication radius rv centered on the underwater robot Um is divided into a maximum stable communication area with a communication radius rs centered on the underwater robot Um, and according to this division rule, a communication gravitational field generated by the underwater robot Um in its neighbor set Pm is calculated:
Wherein dmn＝||Xm-Xn is the distance between the underwater robot Um and the underwater robot Un, the underwater robot Un and the underwater robot Um in the largest stable communication area of the underwater robot Um have very safe and reliable communication guarantee, no communication attractive force is generated at the moment, and when the underwater robot Un and the underwater robot Um are positioned outside the largest safe communication area, the communication capability between the underwater robot Un and the underwater robot Um is weaker, and mutual attractive communication attractive force is required to be generated to ensure the subsequent safe, stable and reliable communication capability;
and 5, constructing a single-step rewarding function of the underwater robot Um through the parallax angle information acquired in the step 3 and the communication gravitational field acquired in the step 4:
Wherein, Rm(Xm,τm) is a single-step reward of the underwater robot Um, the main body of the single-step reward is composed of an obstacle parallax angle and a communication constraint relation, when the underwater robot Um is closer to the obstacle, the larger the obtained obstacle horizontal parallax angle θH and the obtained obstacle vertical parallax angle θV are, the smaller the obtained reward value is, and conversely, the larger the reward is. Similarly, when the underwater robot Un is closer to the underwater robot Um, the more stable the communication is, the larger the obtained prize value is, and conversely, the smaller the prize is. τm is the control input of the underwater robot Um at the moment, K1 and K2 are weight coefficients, the magnitude of the single-step rewards reflects the degree of quality of the control strategy τm at the moment, and the larger the rewards, the better the control strategy, and the worse the control strategy.
Step 6, updating a value function through the single step rewards in step 5, wherein the value function is defined as follows:
Q(Xmk,τmk)＝Rm(Xmk,τmk)+γ×maxQ(Xmk+1,τmk+1) (7)
where Xmk and τmk are the position and control inputs of the underwater robot Um at time step k, 0< γ≤1 is a discount factor. And (3) fitting, iterating and updating the value function through the built deep reinforcement learning neural network of each underwater robot, and repeating the steps (2) to (5) until the neural network converges. The controllers are respectively deployed on the single-chip microcomputer microprocessor units in the corresponding underwater robot control cabin body 7, and at the moment, the underwater robot can acquire an optimal control strategy
It should be noted that the above embodiments are merely for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the above embodiments, it should be understood by those skilled in the art that the technical solution described in the above embodiments may be modified or some or all of the technical features may be equivalently replaced, and these modifications or substitutions do not make the essence of the corresponding technical solution deviate from the scope of the technical solution of the embodiments of the present invention.