CN104809935A - Simulation training method for special situation fault of unmanned aerial vehicle and system thereof - Google Patents

Abstract

The invention provides a simulation training method for a special situation fault of an unmanned aerial vehicle and a system thereof, and belongs to the field of data processing of the unmanned aerial vehicle. A fault tree theory-based unmanned combat aerial vehicle system fault traversing analysis method is provided; a fault which affects the safety of flight is used as a top event of a fault tree to decompose the fault step by step to obtain all fault modes which affect the safety of the flight; meanwhile, a simulation snapshot technology-based unmanned combat aerial vehicle training process decoupling design technology is provided, current training state data is extracted in real time to establish a snapshot database, snapshot data of corresponding stages is only required to be called aiming at different training stages, and simulation initialization is performed, so the problem of extraction of initial training parameters of different task stages is solved; the method and the system have a random selection function at the specific training stage of the special situation fault of the unmanned aerial vehicle and a random fault selection function.

Description

The special feelings fault simulation training method of a kind of unmanned plane and system thereof

Technical field

The present invention relates to Unmanned Aerial Vehicle Data process field, particularly relate to the special feelings fault simulation training method of a kind of unmanned plane and system thereof.

Background technology

Special feelings refer to the special circumstances run in flight course.Frequent entail dangers to flight safety during appearance special circumstances; for improving the fighting efficiency of aircraft system; carrying out the training of special feelings in advance can intense strain effectively faced by ease pilot during special feelings; for unmanned plane, by height special feelings simulated training true to nature, the training of unmanned aerial vehicle operating personnel is become gradually to the Main Means of training unmanned plane manipulation hand.

In prior art, flight simulator Tactical Unmanned Air Vehicle being carried out to simulated training accurately can not determine that pilot needs fault type to be processed and special circumstances, some conventional Universal Faults can only be set usually, on the other hand, Whole Process Simulation training can only be carried out, when for particular flight stage-training (as landed), need to spend a lot of time wait, as shown in Figure 3, strong coupling is there is between training stage, after the special feelings training in a stage, unmanned plane may have left normal task envelope curve, the manual correction of long period must be carried out by trainer and just can get back to normal mission phase, carry out the training of next stage, due to the singularity that special feelings are trained, the controllability being in the unmanned plane of special situation state reduces greatly, once the training failure to train in certain stage causes the crash of unmanned plane, whole task envelope curve need be repeated simultaneously, simultaneously, the exploitation of special feelings simulated training is lagged behind to the development of aircraft, development progress is slow.

Summary of the invention

In order to solve the problem, the invention provides the special feelings fault simulation training method of a kind of unmanned plane and system thereof, on the basis of the special feelings fault tree analysis of UAS and training stage snapshot decoupling zero, have employed the thought of modular design, achieve the integrated of different unmanned plane emulation module, solve the special feelings fault simulation training of unmanned plane of different phase.

The present invention first aspect proposes a kind of unmanned plane special feelings fault simulation training method, comprising:

Using special for unmanned plane feelings fault as top event, the bottom event that all and described top event directly or indirectly associates is listed by traversal, described traversal is, find out likely directly cause the intermediate event of described top event, to each intermediate event, find out likely directly cause the secondary one-level event of described intermediate event, find out all bottom events until follow the trail of, described bottom event is the fundamental cause causing the special feelings fault of described unmanned plane;

The whole process of special for described unmanned plane feelings fault simulation training is divided into multiple training stage, and in each training stage, the flight status parameter of extract real-time current training stage is preserved, and sets up snapshot database;

When unmanned plane arbitrary training stage carries out the training of special feelings fault simulation, read the snapshot data in the snapshot database corresponding with this training stage, original state is it can be used as to be loaded in digital aircraft emulation system, choose bottom event to join this training stage as special feelings fault and carry out simulated training, described digital aircraft emulation system is the sequencing simulation of true aircraft simultaneously;

Simulated training process graphicalization is shown.

Preferably, the described intermediate event of top event that directly causes comprises sensor fault, flight control computer fault, actuator fault, Mechatronic Systems fault, power system fault and Data-Link fault.

In above-mentioned either a program preferably, the fundamental cause of the special feelings fault of described unmanned plane refers to the bottom fault in the special feelings fault of unmanned plane, this bottom fault cannot continue to be split as other fault, or lower one deck fault of this bottom fault does not belong to maintenance range when the special feelings fault simulation of unmanned plane is trained.

In above-mentioned either a program preferably, the training stage add bottom event carry out simulated training as fault time, comprising:

By Matlab, all events of the special feelings fault of unmanned plane are emulated and verified, build the realistic model of digital aircraft;

Above-mentioned realistic model is converted into the binary code of dynamic link library form, this dynamic link library makes each bottom event realistic model corresponding with it be associated;

In VC, call above-mentioned dynamic link library carry out simulated training.

In above-mentioned either a program preferably, described when event model is converted into dynamic link library, comprise and first event model is changed into ANSI-C code, again this ANSI-C code compilation is generated dynamic link library afterwards.

In above-mentioned either a program preferably, in VC, call dynamic link library carry out simulated training, comprising:

Call initialize routine to digital aircraft emulation system initialization;

Call periodic function and the control command arranging instruction and land station of instructor station is passed to digital aircraft emulation system;

Numeral aircraft emulation system is resolved above-mentioned instruction, and calculation result is passed to instructor station;

Instructor station drives vision simulation program to carry out vision simulation;

Numeral aircraft emulation system repeats above-mentioned periodic function and subsequent step, until receive stopping or the reset instruction of instructor station.

The present invention provides a kind of unmanned plane special feelings fault simulation training system on the other hand, comprising:

Numeral aircraft module, for emulating the special feelings fault of unmanned plane and verify, and sets up its dynamic link library;

Instructor station's module, for in each training stage, the flight status parameter of extract real-time current training stage is preserved, set up snapshot database, and for by the middle of snapshot database and special feelings direct fault location to the training stage of digital aircraft emulation system, described by special feelings direct fault location to the digital aircraft emulation system training stage time, described instructor station's module calls the dynamic link library in the middle of described digital aircraft emulation system module, thus calls the special feelings fault emulated;

What comes into a driver's module, it has built model bank and terrain lib, for showing simulated training process graphicalization.

Key point of the present invention and advantage are:

The present invention, proposes the system failure traversal analytical approach based on fault tree theory, decomposes step by step, draw whole fault mode, ensure that the training of special feelings is pointed; Propose the training flow process Decoupling design technology based on training snapshot, set up snapshot database, realize any selection of training stage, save training time and cost; Exploitation integration mode is versatile and flexible, can develop, achieve the intervention as early as possible of pilot training with model is parallel; And by replacing configuration to key model, realizing transplanting between different model using, significantly improving training effectiveness.

Accompanying drawing explanation

Fig. 1 is the implementation process flow diagram of Calling MATLAB in the VC according to a preferred embodiment of unmanned plane of the present invention special feelings fault simulation training method.

Fig. 2 is the present invention's Construction of Fault Tree schematic diagram embodiment illustrated in fig. 1.

Fig. 3 is traditional unmanned plane training method schematic diagram.

Fig. 4 is the present invention's unmanned plane based on analogue system snapping technique embodiment illustrated in fig. 1 special feelings training process flow diagram.

Fig. 5 is the present invention's digital flight simulation process flow diagram embodiment illustrated in fig. 1.

Embodiment

Below in conjunction with accompanying drawing, the present invention is described in further details.

The present invention first proposed a kind of unmanned plane special feelings fault simulation training method, comprising:

Using special for unmanned plane feelings fault as top event, the bottom event that all and described top event directly or indirectly associates is listed by traversal, described traversal is, find out likely directly cause the intermediate event of described top event, to each intermediate event, find out likely directly cause the secondary one-level event of described intermediate event, find out all bottom events until follow the trail of, described bottom event is the fundamental cause causing the special feelings fault of described unmanned plane;

The whole process of special for described unmanned plane feelings fault simulation training is divided into multiple training stage, and in each training stage, the flight status parameter of extract real-time current training stage is preserved, and sets up snapshot database;

When unmanned plane arbitrary training stage carries out the training of special feelings fault simulation, read the snapshot data in the snapshot database corresponding with this training stage, original state is it can be used as to be loaded in digital aircraft emulation system, choose bottom event to join this training stage as special feelings fault and carry out simulated training, described digital aircraft emulation system is the sequencing simulation of true aircraft simultaneously;

Simulated training process graphicalization is shown.

Concrete, adopt the Unmanned combat aircraft system failure traversal analytical approach based on fault tree theory, to the fault of flight safety be affected as top event, fault is decomposed step by step, draw the whole fault modes affecting flight safety, the described intermediate event of top event that directly causes comprises sensor fault, flight control computer fault, actuator fault, Mechatronic Systems fault, power system fault and Data-Link fault.

Fault tree analysis, as a kind of technology of reliability and safety analysis, is a kind of symbolic logic analytical approach be based upon on operational research and system reliability basis, has clear thinking, feature that logicality is strong.Usual fault tree graph can be converted into reliability block diagram easily, thus comprises the distribution changed because of the time with fiduciary level formula, and therefore it both can do qualitative analysis, also can carry out quantitative test.In described Unmanned combat aircraft system failure traversal analytical approach, first the fault tree by decomposing step by step, find out the institute's likely immediate cause causing top event to occur, i.e. intermediate event, follow the trail of again and find out all possible causes causing each intermediate event to occur, finally track each fundamental cause, i.e. bottom event; In described Unmanned combat aircraft system failure traversal analytical approach, fault tree graph is converted into the Boolean algebra equation of equivalence, and then tries to achieve the probability of happening of top event.Be specially:

As shown in Figure 2, set up fault tree, in tree, each symbol implication is as follows: the special feelings of A unmanned plane; B1 sensor fault; C1 air data computer fault; D1 air height fault; D2 air Mach number fault; D3 air stagnation pressure fault; D4 air static pressure fault; D5 air indicated airspeed fault; D6 air elevation rate fault; D7 angle of attack sensor fault; D8 sideslip sensor fault; C2GPS fault; D9GPS height fault; D10GPS elevation rate fault; D11GPS positional information lost efficacy; D12GPS velocity information lost efficacy; C3 inertial navigation fault; D13 inertial navigation axial acceleration fault; D14 inertial navigation lateral overload fault; D15 inertial navigation normal g-load fault; D16 inertial navigation longitudinal velocity fault; D17 inertial navigation side velocity fault; D18 inertial navigation normal velocity fault; Dress angular speed fault is rolled in D19 inertial navigation; D20 inertial navigation pitch rate fault; D21 inertial navigation yawrate fault; D22 inertial navigation ground velocity fault; B2 flight control computer fault; E1 flight control computer primary fault; E2 flight control computer secondary failure; E3 flight control computer lost efficacy; B3 actuator fault; E4 hydraulic system 1 fault; E5 hydraulic system 2 fault; ; E6 left inside side rudder face fault; E7 Right Inboard rudder face fault; E8 left side rudder face fault; E9 right side rudder face fault; Rudder face fault on E10 left-external side; Rudder face fault under E11 left-external side; Rudder face fault on the right outside side of E12; Rudder face fault under the right outside side of E13; E14 front-wheel carrying fault; E15 right main wheel carrying fault; E16 left main wheel carrying fault; E17 main brake fault; E18 backup brake fault; B4 Mechatronic Systems fault; The left generator failure of F1; The right generator failure of F2; F3EPU fault; B5 power system fault; F4 fuel feeding house steward not supercharging; F5 excess oil 400; The right oil feed pump fault of F6 left oil feed pump fault F7; F8 throttle fault; F9 engine backup fault; F10 engine surge fault; F11 Engine Anti-Ice fault; F12 engine anti-asthma fault; F13 engine overrun fault; F14 engine overtemperature fault; F15 motor oil fault; F16 engine monitoring fault; Trouble of shutdown in F17 engine air; F18 computer in the engine 1 fault; F19 computer in the engine 2 fault; B6 Data-Link fault; G1 Data-Link interrupts; G2 Data-Link postpones.

Secondly fault tree is analyzed, comprises Boolean algebra equation fault tree graph being converted into equivalence as follows:

$A=\underset{i=1}{\overset{6}{\∪}}{B}_i=(C_1\∪C_2\∪C_3)\∪\left(\underset{k=1}{\overset{18}{\∪}}{E}_k\right)\∪\left(\underset{l=1}{\overset{19}{\∪}}{F}_l\right)\∪\left(\underset{m=1}{\overset{2}{\∪}}{G}_m\right)=\left(\underset{j=1}{\overset{22}{\∪}}{D}_j\right)\∪\left(\underset{k=1}{\overset{18}{\∪}}{E}_k\right)\∪\left(\underset{l=1}{\overset{19}{\∪}}{F}_l\right)\∪\left(\underset{m=1}{\overset{2}{\∪}}{G}_m\right)$

From above formula, fault tree is made up of 61 single-order minimal cut sets, and suppose that the probability of happening of single-order minimal cut set is, then the probability of happening of top event is

$p\left(A\right)\≈\underset{j=1}{\overset{22}{\mathrm{\Σ}}}p\left(D_j\right)+\underset{k=1}{\overset{18}{\mathrm{\Σ}}}p\left(E_k\right)+\underset{l=1}{\overset{19}{\mathrm{\Σ}}}p\left(F_l\right)+\underset{m=1}{\overset{2}{\mathrm{\Σ}}}p\left(G_m\right)$

Engineering practice shows, from reliability, safety perspective, cause the reason of the special feelings of UAS a lot, but most of unmanned plane during flying fault is analyzed and shows engine system fault (F4-F19), actuator fault (E4-E18), the probability that inertial navigation fault (D13-D22) and GPS fault (D9-D12) occur is larger, it is the main cause that special feelings appear in UAS, individual event training and combined training should be carried out emphatically in task envelope curve, in addition data link postpones (G12) as the distinctive problem of UAS, also number realization should be strengthened when consequent malfunction is simulated.

In the present embodiment, should be understood that, the fundamental cause of the special feelings fault of described unmanned plane refers to the bottom fault in the special feelings fault of unmanned plane, this bottom fault cannot continue to be split as other fault, or lower one deck fault of this bottom fault does not belong to maintenance range when the special feelings fault simulation of unmanned plane is trained, such as D17 is inertial navigation side velocity fault, this fault cannot continue to be split as other Low Fault, therefore D17 is bottom fault, F14 engine overtemperature fault for another example, this fault next stage is certain part overtemperature fault, this fault is not in the maintenance range of this simulated training system, therefore F14 is bottom fault.

Whole process by special for described unmanned plane feelings fault simulation training of the present invention is divided into multiple training stage and divides as the case may be, in the present embodiment, the flight course of unmanned plane is reduced to 11 stages, as shown in Figure 3, should be understood that, being divided into multiple training stage is not limited thereto 11 stages, Fig. 3 is the special feelings fault simulation training of classic method, strong coupling is there is as seen from the figure between training stage, after the special feelings training in a stage, unmanned plane may have left normal task envelope curve, the manual correction of long period must be carried out by trainer and just can get back to normal mission phase, carry out the training of next stage, meanwhile, due to the singularity that special feelings are trained, the controllability being in the unmanned plane of special situation state reduces greatly, once the training failure to train in certain stage causes the crash of unmanned plane, need repeat whole task envelope curve.Above 2 all will cause the cycle of training elongated, and efficiency is lower.

Be exactly in training system operational process based on training system snapping technique, in real time by the flight status parameter of moment (mainly comprise: mission phase, working method, position, highly, speed, attitude etc.) preserve, set up a snapshot database, different training stage snapshot datas is selected for only needing during the different training stages to ask, and assignment is carried out to variable corresponding in realistic model, it can be used as original state, realistic model brings into operation.So just achieve the decoupling zero between the training stage, as shown in Figure 4, each training use-case only need be loaded into the training system state snapshot of moment, and starts anew without the need to system, the special feelings training of complexity is converted into multiple independently individual event training, drastically increases the efficiency of training.

The present invention the training stage add bottom event carry out simulated training as fault time, comprising:

By Matlab, all events of the special feelings fault of unmanned plane are emulated and verified, build the realistic model of digital aircraft;

Above-mentioned realistic model is converted into the binary code of dynamic link library form, this dynamic link library makes each bottom event realistic model corresponding with it be associated;

In VC, call above-mentioned dynamic link library carry out simulated training.

Specifically as shown in Figure 1, first build special feelings training pattern under Matlab environment, aircraft body dynamics, sensor, actuator, flight control, navigation, flight management, mechanical electronic hydraulic pressing system are emulated and verified; Secondly, by Matlab RTW, the special feelings training pattern of building is converted into the binary code of dynamic link library form; Finally, the dynamic link library of generation is called when realizing digital flight simulation software in VC to realize the real-time simulation of different digital aircraft module.Described when event model is converted into dynamic link library, comprise and first event model is changed into ANSI-C code, again this ANSI-C code compilation is generated dynamic link library afterwards.Mdl file described in Fig. 1 is a certain event of failure through emulation and checking, as a fault in sensor, actuator, flight control, navigation.

In VC, call dynamic link library carry out simulated training, comprising:

Call initialize routine to digital aircraft emulation system initialization, call int FCRT_Init (void*) function during simulation initialisation and utilize the disposal of the Initialize installation of instructor station or unmanned plane solidification to arrange to realize the Initialize installation of digital aircraft;

Call periodic function and the control command arranging instruction and land station of instructor station is passed to digital aircraft emulation system, as shown in Figure 5, emulation cycle is 10ms, and each cycle calls int FCRT_Run (void*, void*) once;

Numeral aircraft emulation system is resolved above-mentioned instruction, and calculation result is passed to instructor station, carries out real-time storage and monitoring by instructor station's software;

Instructor station drives vision simulation program to carry out vision simulation;

Numeral aircraft emulation system repeats above-mentioned periodic function and subsequent step, until receive stopping or the reset instruction of instructor station, calls the emulation of int FCRT_Stop () end number aircraft.

The present invention provides a kind of unmanned plane special feelings fault simulation training system on the other hand, comprising:

Numeral aircraft module, for emulating the special feelings fault of unmanned plane and verify, and sets up its dynamic link library;

Instructor station's module, for in each training stage, the flight status parameter of extract real-time current training stage is preserved, set up snapshot database, and for by the middle of snapshot database and special feelings direct fault location to the training stage of digital aircraft emulation system, described by special feelings direct fault location to the digital aircraft emulation system training stage time, described instructor station's module calls the dynamic link library in the middle of described digital aircraft emulation system module, thus calls the special feelings fault emulated;

What comes into a driver's module, it has built model bank and terrain lib, for showing simulated training process graphicalization.

Concrete, digital aircraft module: this module real time workshop technology that have employed based on Matlab RTW (Real-TimeWorkshop) realizes unmanned plane body dynamics simulation, flies guard system emulation, aircraft on-board systems emulation.

The file destination constructive process of RTW can be divided into two stages, and first stage is code generation process; Second stage be RTW according to template binding file generated makefile file, and with this file, compiling be finally linked as executable program is carried out to generated code.Can be applied in our special feelings training system to make the code automatically generated, we need carry out interface amendment, recompility and link to generated real-time code, finally obtain the dll as dynamic link library file under windows platform, call this dll by the every 10ms of Visual Studio and realize unmanned plane real-time simulation.Compared with the method for traditional hand-coding simulation software code, this method shortens the lead time of system, and ensure that the reliability of software to a certain extent.

Instructor station's module, this module is connected with database by ADO, controls, can carry out spy while management and supervision neatly and be with one's heart at the foundation of system snapshot and loading, spy and care for into, flight data recording, playback and analysis realizing simulation flow.

What comes into a driver's module, this module adopts Vega Prime5 and Visual Studio2005 to realize, adopt modular design method, reserved standard interface has also built model bank and terrain lib, the visual simulation of different unmanned plane can be realized easily, and continuous print landform, airport, runway, buildings, road and significant topography and geomorphology can be constructed, the visual effects such as simulate day, dusk, night, rain, mist, snow, for operating personnel provide the simulation visual field at the 3rd visual angle true to nature, in real time for operating personnel provide virtual visual effect.

Interface control document (Interface Control Document, ICD) be the technological document of the function of interface signal, technical characteristic and operation instruction between definition with each subsystem describing the digital aircraft system of composition, it is the important component part of digital aircraft system Top-layer Design Method and system specifications.Virtual aircraft emulation is divided into 6 modules, and builds with reference to ICD agreement, and be conducive to unmanned plane function modoularization and progressively encapsulate and expand, each integration module can carry out the emulation of normal condition and special feelings.

It should be noted that; the special feelings fault simulation training method of unmanned plane of the present invention and system thereof comprise any one and combination in any thereof in above-described embodiment; but embodiment recited above is only be described the preferred embodiment of the present invention; not the scope of the invention is limited; do not departing under the present invention designs spiritual prerequisite; the various distortion that the common engineering technical personnel in this area make technical scheme of the present invention and improvement, all should fall in protection domain that claims of the present invention determine.

Claims (7)

1. a unmanned plane special feelings fault simulation training method, is characterized in that, comprising:

Using special for unmanned plane feelings fault as top event, the bottom event that all and described top event directly or indirectly associates is listed by traversal, described traversal is, find out likely directly cause the intermediate event of described top event, to each intermediate event, find out likely directly cause the secondary one-level event of described intermediate event, find out all bottom events until follow the trail of, described bottom event is the fundamental cause causing the special feelings fault of described unmanned plane;

The whole process of special for described unmanned plane feelings fault simulation training is divided into multiple training stage, and in each training stage, the flight status parameter of extract real-time current training stage is preserved, and sets up snapshot database;

When unmanned plane arbitrary training stage carries out the training of special feelings fault simulation, read the snapshot data in the snapshot database corresponding with this training stage, original state is it can be used as to be loaded in digital aircraft emulation system, choose bottom event to join this training stage as special feelings fault and carry out simulated training, described digital aircraft emulation system is the sequencing simulation of true aircraft simultaneously;

Simulated training process graphicalization is shown.

2. unmanned plane according to claim 1 special feelings fault simulation training method, is characterized in that: the described intermediate event of top event that directly causes comprises sensor fault, flight control computer fault, actuator fault, Mechatronic Systems fault, power system fault and Data-Link fault.

3. unmanned plane according to claim 1 special feelings fault simulation training method, it is characterized in that: the fundamental cause of the special feelings fault of described unmanned plane refers to the bottom fault in the special feelings fault of unmanned plane, this bottom fault cannot continue to be split as other fault, or lower one deck fault of this bottom fault does not belong to maintenance range when the special feelings fault simulation of unmanned plane is trained.

4. unmanned plane according to claim 1 special feelings fault simulation training method, is characterized in that, the training stage add bottom event carry out simulated training as fault time, comprising:

By Matlab, all events of the special feelings fault of unmanned plane are emulated and verified, build the realistic model of digital aircraft;

Above-mentioned realistic model is converted into the binary code of dynamic link library form, this dynamic link library makes each bottom event realistic model corresponding with it be associated;

In VC, call above-mentioned dynamic link library carry out simulated training.

5. unmanned plane according to claim 4 special feelings fault simulation training method, it is characterized in that: first event model is changed into ANSI-C code by described comprising when event model being converted into dynamic link library, again this ANSI-C code compilation is generated dynamic link library afterwards.

6. unmanned plane according to claim 5 special feelings fault simulation training method, is characterized in that, call dynamic link library and carry out simulated training, comprising in VC:

Call initialize routine to digital aircraft emulation system initialization;

Call periodic function and the control command arranging instruction and land station of instructor station is passed to digital aircraft emulation system;

Numeral aircraft emulation system is resolved above-mentioned instruction, and calculation result is passed to instructor station;

Instructor station drives vision simulation program to carry out vision simulation;

Numeral aircraft emulation system repeats above-mentioned periodic function and subsequent step, until receive stopping or the reset instruction of instructor station.

7. a unmanned plane special feelings fault simulation training system, is characterized in that, comprising:

Numeral aircraft module, for emulating the special feelings fault of unmanned plane and verify, and sets up its dynamic link library;

Instructor station's module, for in each training stage, the flight status parameter of extract real-time current training stage is preserved, set up snapshot database, and for by the middle of snapshot database and special feelings direct fault location to the training stage of digital aircraft emulation system, described by special feelings direct fault location to the digital aircraft emulation system training stage time, described instructor station's module calls the dynamic link library in the middle of described digital aircraft module, thus calls the special feelings fault emulated;

What comes into a driver's module, it has built model bank and terrain lib, for showing simulated training process graphicalization.