A simulation system and method for quickly generating optimal planning paths for unmanned aerial vehicles

Through the particle swarm optimization algorithm and six-axis semi-physical simulation mathematical model combined with the gradient descent parameter optimization module, the problems of fuel consumption constraints and local global optimization in the traditional path planning method are solved, and the rapid and efficient generation of drone path planning is achieved, improving the efficiency of the generation of drone path planning. The accuracy and computational efficiency of simulation results are achieved.

CN115562074BActive Publication Date: 2024-11-26FUZHOU UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202211339465.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-10-29
Publication Date
2024-11-26
Estimated Expiration
2042-10-29

AI Technical Summary

Technical Problem

Traditional UAV path planning methods lack consideration of fuel consumption and other constraints, resulting in low confidence in simulation results and defects of local and global optimization, making it difficult to generate efficient optimal paths in complex military environments.

Method used

The particle swarm optimization algorithm is used to combine the six-axis semi-physical simulation mathematical model and the gradient descent parameter optimization module to generate the optimal path through the data input module and the simulation map module, consider the fuel constraints and optimize the adaptive factors in the algorithm to achieve dual optimization of the path and closed-loop optimization.

Benefits of technology

Rapidly generate the optimal planning path of the drone while taking into account fuel constraints, improves the accuracy and response speed of path planning, enhances the confidence of simulation results, and realizes efficient computing and wide application.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115562074B_ABST
    Figure CN115562074B_ABST
Patent Text Reader

Abstract

The invention relates to a simulation system for rapidly generating an optimal planning path for an unmanned aerial vehicle, comprising a data input module, a simulation map module, an unmanned aerial vehicle path planning module, a gradient descent parameter optimization module and a six-axis semi-physical simulation module; the unmanned aerial vehicle path planning module comprises a path planning module, an adaptive factor unit, a six-axis semi-physical simulation mathematical model and a particle swarm optimization algorithm; the unmanned aerial vehicle path planning module uses a particle swarm optimization algorithm to perform path planning based on data input by a data input module and a three-dimensional map imported by a simulation map module, performs simulated flight according to a path obtained by the path planning module through the six-axis semi-physical simulation mathematical model, optimizes the particle swarm optimization algorithm after importing path information into an adaptive factor, and obtains an initial optimal path; the six-axis semi-physical simulation module performs simulation based on the initial optimal path to obtain motion information, optimizes the particle swarm optimization algorithm through a gradient descent parameter optimization module, and obtains a final optimal path based on the final particle swarm optimization algorithm.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle (UAV) path planning, and in particular to a simulation system and method for rapidly generating an optimal planned path for an unmanned aerial vehicle (UAV). Background Art

[0002] With the development of unmanned aerial technology, unmanned aerial vehicles have been increasingly used in complex and dangerous military environments. As an important part of the flight control and autonomous navigation system of unmanned aerial vehicles, the response speed and planning accuracy of unmanned aerial vehicle path planning methods have a huge impact on the navigation results. Traditional path planning methods include simulated annealing methods, A* methods, etc. These path planning methods are not suitable for complex military environments with various constraints.

[0003] Traditional path planning methods lack constraints on fuel consumption and other conditions. The above method omits the constraints in the actual system, which greatly reduces the confidence of the simulation results. At the same time, there are defects in the local optimization and global optimization decision in the UAV path planning algorithm. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a simulation system and method for quickly generating the optimal planning path of a UAV, which can quickly generate the optimal planning path of a UAV under consideration of fuel consumption constraints.

[0005] To achieve the above object, the present invention adopts the following technical solution:

[0006] A simulation system for rapidly generating an optimal planned path for an unmanned aerial vehicle comprises a data input module, a simulation map module, an unmanned aerial vehicle path planning module, a gradient descent parameter optimization module and a six-axis semi-physical simulation module; the unmanned aerial vehicle path planning module comprises a six-axis semi-physical simulation mathematical model and a particle swarm optimization algorithm; the unmanned aerial vehicle path planning module uses a particle swarm optimization algorithm to perform path planning based on data input by the data input module and a three-dimensional map imported by the simulation map module, and performs simulated flight according to the path obtained by the path planning module through the six-axis semi-physical simulation mathematical model, optimizes the particle swarm optimization algorithm after importing the path information into an adaptive factor, and further obtains an initial optimal path; the six-axis semi-physical simulation module performs simulation based on the initial optimal path to obtain motion information, further optimizes the particle swarm optimization algorithm through the gradient descent parameter optimization module, and obtains a final optimal path based on the final particle swarm optimization algorithm.

[0007] Furthermore, an optimal path output module is provided, which is connected to the UAV path planning simulation system to output the final optimal path;

[0008] Further, the data input module includes a waypoint given input unit and a fuel constraint input unit;

[0009] The waypoint given input unit is connected to the UAV path planning module to set the starting point, end point and other expected waypoints of the path;

[0010] The fuel constraint input unit is connected to the UAV path planning module, and the fuel constraint input specifies the fuel consumption in the UAV path planning.

[0011] Furthermore, the simulation map module includes a map unit and a threat / obstacle unit; the map unit generates a three-dimensional map of the simulated flight of the drone, and the threat / obstacle input unit generates obstacles and emergencies.

[0012] Furthermore, the path planning module uses a particle swarm optimization algorithm to optimize the path; the six-axis semi-physical simulation mathematical model uses the obtained optimal path to simulate flight, and then imports the obtained path information into the adaptive factor; the adaptive factor is used to measure the weight of the global or local search in the simulated flight, optimize the parameters of the weight factor in the algorithm, strengthen the global / local search, and realize a local closed loop; the UAV path planning module repeatedly runs to obtain the initial optimal path.

[0013] Furthermore, the six-axis semi-physical simulation module includes a conversion module, a motion controller and a six-axis simulation platform; the conversion module is connected to the motion controller; the conversion module is connected to the path planning module to obtain the initial optimal path for setting the specific motion and trajectory information of the unmanned aerial vehicle, including the position, speed and attitude angle corresponding to each moment, and then converting the motion and trajectory information into a control signal to control the six-axis simulation platform through the motion controller; the position, speed and attitude information collected by the six-axis simulation platform is received by the motion controller and the gradient descent parameter optimization module is used to optimize the adaptive factor.

[0014] Furthermore, the gradient descent parameter optimization module includes a loss function and a gradient descent method; the loss function is a function for optimizing the parameters of the adaptive factor; the loss function is respectively connected to the motion controller and the gradient descent method to receive path information; the path information is imported into the loss function, and the parameters of the adaptive factor are optimized by the gradient descent method.

[0015] A path generation method for a simulation system for quickly generating an optimal planning path for a drone comprises the following steps:

[0016] Step S1: Set the motion trajectory, scene, obstacles and emergencies of the UAV through the simulation map module, add fuel consumption constraints to the particle swarm optimization algorithm, and set the required waypoints;

[0017] Step S2: After the scenario, fuel constraints and waypoints are determined, the path planning is performed using the particle swarm optimization algorithm. The six-axis semi-physical simulation mathematical model uses the obtained path to simulate flight, and the path information is imported into the adaptive factor to optimize the initial algorithm. After repeating the above steps, the initial optimal path is output;

[0018] Step S3: the conversion module obtains the optimal path, issues a motion control instruction, transmits a control signal through the motion controller, and controls the six-axis simulation platform to achieve corresponding linear and rotational motions;

[0019] Step S4: the six-axis simulation platform collects motion information and feeds it back to the motion controller, and transmits it to the loss function in the gradient descent parameter optimization module through the motion controller;

[0020] Step S5: The loss function receives the path information such as the position, speed and posture of the semi-physical simulation motion of the UAV, and optimizes the parameters of the adaptive factor by the gradient descent method; the adaptive factor updates the optimized parameters to achieve a global closed loop;

[0021] Step S6: Run the optimized UAV path planning module to obtain the final optimized path and output it.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] 1. The present invention takes fuel constraints and optimizes adaptive factors into consideration, which is closer to reality and aims to quickly generate optimal path planning. The system can continuously update and iterate during path planning simulation of fast flight, optimize both the algorithm and adaptive factors, and achieve closed-loop optimization.

[0024] 2. The present invention has strong practicability, high computational efficiency, supports expansion and updating, and has wide applications. It can become a highly practical semi-physical simulation system for UAVs to quickly generate optimal path planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 1 is a system block diagram of an embodiment of the present invention, wherein 1 is a waypoint given input unit, 2 is a fuel constraint input unit, 3 is a simulation map module, 4 is a particle swarm optimization algorithm, 5 is a path planning module, 6 is a six-axis semi-physical simulation mathematical model, 7 is an adaptive factor unit, 8 is a conversion module, 9 is a motion controller, 10 is a six-axis simulation platform, and 11 is a gradient descent parameter optimization module;

[0026] Figure 2 4 is a system working flow diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0028] Please refer to Figure 1 The present invention provides a semi-physical simulation system for quickly generating the optimal planning path of a UAV under fuel constraint conditions, including: a waypoint given input unit 1, a fuel constraint input unit 2, a map unit 3, a UAV path planning module, a six-axis semi-physical simulation module, and a gradient descent parameter optimization module 11.

[0029] In this embodiment, the waypoint given input unit is connected to the UAV path planning module, and the waypoint given input unit 1 is used to set the starting point, end point and other expected waypoints of the path. The fuel constraint input unit 2 is connected to the UAV path planning module, and the fuel constraint input unit 2 stipulates the fuel consumption in the UAV path planning; the fuel constraint conditions generated by the fuel constraint input unit 2 are input into the UAV path planning module, and are used to optimize the algorithm under the fuel constraint conditions; the simulation map module 3 is connected to the UAV path planning module, and the simulation map module 3 includes a map unit and a threat / obstacle unit; the map unit generates a three-dimensional map of the UAV simulated flight, and the threat / obstacle unit generates obstacles and possible threats and other emergencies; the UAV path planning module is respectively connected to the optimal path output and the six-axis semi-physical simulation module, and the UAV path planning module The optimal path is quickly generated by optimizing the algorithm; the optimal path output is connected to the UAV path planning simulation system to output the final optimal path; the six-axis semi-physical simulation module is respectively connected to the UAV path planning module and the gradient descent parameter optimization module 11 to input the obtained optimal path into the six-axis semi-physical simulation module for semi-physical simulated flight, and then obtain path information in a simulated flight relative to the real situation; the gradient descent parameter optimization module 11 is connected to the UAV path planning module, and the gradient descent parameter optimization module is used to receive the path information obtained by the semi-physical simulation, and optimize the UAV path planning module through these relatively real flight path information.

[0030] In this embodiment, the UAV path planning module includes four parts: a particle swarm optimization algorithm 4, a path planning module 5, a six-axis semi-physical simulation mathematical model 6, and an adaptive factor 7; the UAV path planning module path planning algorithm selects the particle swarm optimization algorithm 4; the particle swarm optimization algorithm 4 is imported into the path planning module 5 to obtain the optimal path; the six-axis semi-physical simulation module mathematical model 6 uses the obtained optimal path to simulate flight, and then imports the obtained path information into the adaptive factor 7; the adaptive factor 7 is used to measure the weight of the global or local search in the simulated flight, optimize the parameters of the weight factor in the algorithm, strengthen the global / local search, and realize the local closed loop; the UAV path planning module is repeatedly operated to obtain the initial optimal path.

[0031] In this embodiment, the six-axis semi-physical simulation module includes a conversion module 8 , a motion controller 9 and a six-axis simulation platform 10 .

[0032] Preferably, the motion controller 9 adopts the British TRIO sixteen-axis motion controller; the six-axis simulation platform 10 is built with aviation aluminum, and includes a three-axis truss platform and a three-axis rotating platform; the six-axis simulation platform is connected to the motion controller 9 using the conversion module 8; the conversion module 8 is connected to the path planning module 5 to obtain the initial optimal path for setting the specific motion and trajectory information of the unmanned aerial vehicle, including the position, speed and attitude angle corresponding to each moment, and then converts the motion and trajectory information into a control signal, and controls the six-axis simulation platform 10 through the motion controller 9; the position, speed and attitude information collected by the six-axis simulation platform 10 is received by the motion controller 9 and transmitted to the gradient descent parameter optimization module 11 to optimize the adaptive factor 7.

[0033] In this embodiment, the gradient descent parameter optimization module 11 includes a loss function and a gradient descent method; the loss function is a function for optimizing the parameters of the adaptive factor 7; the loss function is connected to the motion controller 9 and the gradient descent method respectively, and is used to receive the collected path information such as position, speed, and posture. The path information is imported into the loss function, and the parameters of the adaptive factor 7 are optimized by the gradient descent method.

[0034] like Figure 2 As shown, preferably, this embodiment also provides a working method of a semi-physical simulation system for quickly generating an optimal planning path for a UAV under fuel constraints, comprising the following steps:

[0035] Step S1: Set the motion trajectory, scene, obstacles and possible threats of the UAV through the simulation map module, add fuel consumption constraints to the particle swarm optimization algorithm 4, and set the required waypoints;

[0036] Step S2: After the scenario, fuel constraints and waypoints are determined, the particle swarm optimization algorithm 4 is introduced into the path planning module 5 for path planning, the six-axis semi-physical simulation mathematical model 6 uses the obtained path to simulate flight, and the path information is introduced into the adaptive factor 7 to optimize the initial algorithm. After repeating the above steps, the initial optimal path is output;

[0037] Step S3: the conversion module 8 obtains the optimal path, issues a motion control instruction, transmits a control signal through the motion controller 9, and controls the six-axis simulation platform 10 to achieve corresponding linear and rotational motions;

[0038] Step S4: the six-axis simulation platform 10 collects motion information and feeds it back to the motion controller 9, and transmits it to the loss function in the gradient descent parameter optimization module 11 through the motion controller 9;

[0039] Step S5: The loss function receives the path information such as the position, speed and posture of the semi-physical simulation motion of the UAV, and optimizes the parameters of the adaptive factor 7 by the gradient descent method; the adaptive factor 7 updates the optimized parameters to achieve a global closed loop;

[0040] Step S6: Run the optimized UAV path planning module to obtain the final optimized path and output it.

[0041] Preferably, in this embodiment, the local closed loop means that, on the one hand, the particle swarm optimization algorithm 4 in the UAV path planning module can realize simulated flight in the six-axis semi-physical simulation mathematical model 6, and on the other hand, the algorithm can optimize the parameters of the algorithm through the adaptive factor 7 after the six-axis semi-physical simulation mathematical model 6 is completed and then feed the parameters back to the particle swarm optimization algorithm 4.

[0042] Preferably, in this embodiment, the global closed loop means that the UAV path planning module can import the obtained initial optimal planning path into the conversion module 8, and the path information is converted into a control signal through the conversion module 8 and transmitted to the motion controller 9. The motion controller 9 receives the control signal to make the six-axis simulation platform 10 move according to the motion instructions of the optimal planning path. On the other hand, the position, speed, posture and path information of the six-axis simulation platform 10 can be fed back to the motion controller 9. The motion controller 9 imports the information obtained by the six-axis simulation platform 10 into the gradient descent optimization module to optimize the parameters of the adaptive factor 7. The optimized parameters of the adaptive factor 7 are then updated and imported into the UAV path planning module to form a closed loop.

[0043] Preferably, the six-axis semi-physical simulation module of this embodiment uses a 6-degree-of-freedom motion simulation mechanism, which can achieve high-precision simulation of the position change and attitude angle change of the drone in all directions, and the simulation effect is closer to the actual situation. The method adds fuel constraints and optimization of adaptive factors, which is closer to reality and is aimed at fast flight optimization algorithms. The system can continuously update and iterate during path planning simulation of fast flight, dually optimize the algorithm and adaptive factors, and achieve closed-loop optimization. The system of the present invention has strong practicality, high computational efficiency, high computational efficiency, and supports expansion and updating, and is widely used. It can become a highly practical semi-physical simulation system for drones to quickly generate optimal path planning.

[0044] The function of the map module is to set a variety of different scenes for the simulated flight of the drone, such as high-rise buildings, mountains, etc.

[0045] Preferably, in this embodiment, the workflow of the method is as follows:

[0046] Step 1: Set experimental conditions from aspects such as simulated flight scenarios, fuel constraints, waypoints, obstacles, threats and attitude information, formulate experimental plans, and select one combination of conditions to conduct experiments to optimize the particle swarm optimization algorithm 4.

[0047] Step 2: The particle swarm optimization algorithm 4 obtains the optimal path through the path planning module 5, inputs the optimal path into the six-axis semi-physical simulation mathematical model 6 for simulated flight, and then optimizes the parameters of the path planning algorithm through the adaptive factor 7. An optimization algorithm suitable for quickly generating the optimal planning path of a UAV with fuel constraints is obtained.

[0048] Step 3: The UAV path planning module outputs the optimal path, and the conversion module 8 obtains the optimal planned path and converts the motion information into a control signal and transmits it to the motion controller 9. The motion controller 9 receives the control signal and controls the six-axis simulation platform 10 to simulate the flight state of the UAV.

[0049] Step 4: The motion controller 9 imports the semi-physical simulation information obtained by the six-axis simulation platform 10 into the gradient descent parameter optimization module 11 to optimize the parameters of the adaptive factor 7. The optimized adaptive factor 7 parameters are updated into the drone path planning module to form a closed loop.

[0050] Step 5: The UAV path planning module runs again and finally outputs the optimal path.

[0051] The above description is only a preferred embodiment of the present invention. All equivalent changes and modifications made according to the scope of the patent application of the present invention should fall within the scope of the present invention.

Claims

1. A simulation system for rapidly generating optimal planning paths for unmanned aerial vehicles, characterized in that: It includes a data input module, a simulation map module, a UAV path planning module, a gradient descent parameter optimization module and a six-axis semi-physical simulation module; the data input module includes a waypoint given input unit and a fuel constraint input unit; the simulation map module includes a map unit and a threat / obstacle unit; the UAV path planning module includes a path planning module, an adaptive factor unit, a six-axis semi-physical simulation mathematical model and a particle swarm optimization algorithm; the UAV path planning module uses a particle swarm optimization algorithm to plan the path based on the data input by the data input module and the three-dimensional map imported by the simulation map module, and simulates the flight according to the path obtained by the path planning module through the six-axis semi-physical simulation mathematical model, optimizes the particle swarm optimization algorithm after importing the path information into the adaptive factor, and further obtains the initial optimal path; the six-axis semi-physical simulation module simulates and obtains motion information based on the initial optimal path, and further optimizes the particle swarm optimization algorithm through the gradient descent parameter optimization module, and obtains the final optimal path based on the final particle swarm optimization algorithm; The six-axis semi-physical simulation module includes a conversion module, a motion controller and a six-axis simulation platform; the conversion module is connected to the motion controller; the conversion module is connected to the path planning module to obtain the initial optimal path to set the specific motion and trajectory information of the unmanned aerial vehicle, including the position, speed and attitude angle corresponding to each moment, and then convert the motion and trajectory information into a control signal, and control the six-axis simulation platform through the motion controller; The position, speed and posture information collected by the six-axis simulation platform are received by the motion controller and then the adaptive factor is optimized using a gradient descent parameter optimization module; The gradient descent parameter optimization module includes a loss function and a gradient descent method; the loss function is a function for optimizing the parameters of the adaptive factor; the loss function is connected to the motion controller and the gradient descent method respectively to receive path information; the path information is imported into the loss function, and the parameters of the adaptive factor are optimized by the gradient descent method; The waypoint given input unit is connected to the UAV path planning module to set the starting point, end point and other expected waypoints of the path; The fuel constraint input unit is connected to the UAV path planning module, and the fuel constraint input specifies the fuel consumption in the UAV path planning; The map unit generates a three-dimensional map of the simulated flight of the UAV, and the threat / obstacle input unit generates obstacles and emergencies; The path planning module uses a particle swarm optimization algorithm to optimize the path; the six-axis semi-physical simulation mathematical model uses the obtained optimal path to simulate flight, and then imports the obtained path information into the adaptive factor; the adaptive factor is used to measure the weight of the global or local search in the simulated flight, optimize the parameters of the weight factor in the algorithm, strengthen the global / local search, and realize a local closed loop; the UAV path planning module repeatedly runs to obtain the initial optimal path.

2. The path generation method of a simulation system for quickly generating an optimal planning path for a drone according to claim 1, characterized in that: The path generation method comprises the following steps: Step S1: Set the motion trajectory, scene, obstacles and emergencies of the UAV through the simulation map module, add fuel consumption constraints to the particle swarm optimization algorithm, and set the required waypoints; Step S2: After the scenario, fuel constraints and waypoints are determined, the path planning is performed using the particle swarm optimization algorithm. The six-axis semi-physical simulation mathematical model uses the obtained path to simulate flight, and the path information is imported into the adaptive factor to optimize the initial algorithm. After repeating the above steps, the initial optimal path is output; Step S3: the conversion module obtains the initial optimal path, issues a motion control instruction, transmits a control signal through the motion controller, and controls the six-axis simulation platform to achieve corresponding linear and rotational motions; Step S4: the six-axis simulation platform collects motion information and feeds it back to the motion controller, and transmits it to the loss function in the gradient descent parameter optimization module through the motion controller; Step S5: The loss function receives the position, speed and attitude path information of the semi-physical simulation motion of the UAV, and optimizes the parameters of the adaptive factor by the gradient descent method; the adaptive factor updates the optimized parameters to achieve a global closed loop; Step S6: Run the optimized UAV path planning module to obtain the final optimized path and output it.