Air-sea cross-domain unmanned intelligent cluster cooperative path planning method

Abstract

The invention provides an air-sea cross-domain unmanned intelligent cluster collaborative path planning method, and belongs to the technical field of collaborative planning. The method comprises the following steps: S1, carrying out kinematics modeling on a UAV unmanned cluster, a USV unmanned cluster and an AUV unmanned cluster; s2, by taking multiple straight lines and few curves of the path, the length of the total path and the short total consumed time as optimization objectives, constructing a collaborative navigation planning objective function through weighted fusion; and S3, carrying out path planning by adopting a collaborative path planning algorithm based on an improved particle swarm. According to the air-sea cross-domain unmanned intelligent cluster collaborative path planning method, the advantages of different platforms can be fully utilized, various collaborative tasks such as target searching, tracking and disposal are executed in space areas such as low altitude, water surface and underwater, the application range is further expanded, the resource utilization rate is further increased, and the method is suitable for large-scale popularization and application. The method has the advantages of finer modeling and performance, prevents the algorithm from falling into local optimum, and accelerates the convergence speed.

Description

Technical Field

[0001] The present invention relates to the field of collaborative planning technology, and in particular to an air-sea cross-domain unmanned intelligent cluster collaborative path planning method. Background Art

[0002] Common cross-domain heterogeneous unmanned platforms in the air and sea include UAV clusters, USV clusters and AUV clusters. These clusters perform task allocation and path planning through collaborative decision-making, and make full use of the advantages of different platforms to perform various collaborative tasks such as target search, tracking, and disposal in low-altitude, surface, underwater and other spatial areas, further expanding the scope of application and resource utilization.

[0003] Due to the characteristics of cross-domain collaborative tasks such as multiple dimensions, high complexity and strong coupling, large-scale complex tasks may be difficult to decouple and optimize the sequencing. Various unmanned intelligent systems have huge differences in mobility, endurance, payload capacity, communication distance and detection methods in different spatial domains. The complex cross-domain environment leads to problems such as insufficient instant decision-making and dynamic adjustment capabilities, making it very difficult for various unmanned intelligent clusters to plan and adjust paths in real time under conditions of limited information resources.

[0004] Currently, researchers have proposed a variety of methods for unmanned cluster collaborative path planning tasks, such as RTT algorithm, A* algorithm, PSO algorithm, etc.

[0005] For example, in a paper published in ICMA in 2018 by Zu Wei et al., by improving the RTT algorithm, UAV clusters can generate optimized path algorithms in real time for various dynamic and static obstacles. This method only focuses on the problem of coordinated path planning for UAV clusters, and does not involve the cross-domain coordination of unmanned clusters at sea.

[0006] For example, in Example 2, in a paper published in Information Sciences: An International Journal in 2019, Liu Yuanchang et al. proposed a USV cluster collaborative task allocation and path planning method considering constraints such as endurance and communication distance. This method only focuses on the collaborative path planning problem of unmanned boat clusters, and does not involve the cross-domain collaboration problem of aerial unmanned clusters.

[0007] For example, in the paper published by Wu Yu et al. in Ocean Engineering in 2019, the PSO algorithm was used to study the collaborative path planning of UAAV and AUV. This method involves the cross-domain collaborative path planning problem of two types of heterogeneous clusters on the surface and underwater, but the PSO algorithm is not easy to converge during the optimization iteration process. In the early stage of optimization, it is necessary to set more constraints and collocation point layouts according to different mission stages, and the adaptability is relatively weak.

[0008] The current research algorithms mostly focus on the task coordination of homogeneous clusters in a certain combat field. The optimization algorithms involved in the task coordination of heterogeneous clusters across domains have large computational complexity and slow convergence speed, resulting in weak adaptive performance. Therefore, the present invention proposes an air-sea cross-domain unmanned intelligent cluster collaborative path planning method to address the deficiencies of the above-mentioned prior art. Summary of the invention

[0009] The purpose of the present invention is to provide an air-sea cross-domain unmanned intelligent cluster collaborative path planning method, which can make full use of the advantages between different platforms, and perform various collaborative tasks such as target search, tracking, and disposal in low-altitude, water surface, underwater and other spatial areas, further expand the scope of application and resource utilization, have more sophisticated modeling and performance advantages, avoid the algorithm from falling into local optimality, and accelerate the convergence speed.

[0010] To achieve the above object, the present invention provides an air-sea cross-domain unmanned intelligent cluster collaborative path planning method, comprising the following steps: Step S1, kinematic modeling of the UAV unmanned cluster, the USV unmanned cluster and the AUV unmanned cluster; Step S2, taking the path with more straight lines and less twists, shorter total route length and shorter total time as the optimization goals, constructing the collaborative navigation planning objective function through weighted fusion; Step S3: Use the collaborative path planning algorithm based on improved particle swarm to perform path planning: The particle swarm algorithm is improved by introducing nonlinear decreasing inertia weight and asymmetric learning factor to obtain the IPSO algorithm. The UAV unmanned cluster, USV unmanned cluster and AUV unmanned cluster are respectively used as the particle swarms to be optimized, and are substituted into the IPSO algorithm for iterative calculation. The function fitness of each particle is calculated according to the objective function, and the speed and position of the particle are updated until the stopping condition is met. Output the optimal path planning results for various types of unmanned swarms.

[0011] Preferably, in step S1, kinematic modeling of the UAV unmanned cluster is performed as follows: ; in, Represents the UAV position vector, including the three-dimensional coordinates of the UAV And heading angle ; Represents the velocity vector of the UAV, including the movement speed of the UAV , angular velocity , acceleration ; represents transpose; The constraints are as follows: ; in, Represents the constraints of the UAV unmanned cluster; represents the detection space of the UAV, Indicates the upper limit of its detection space, Indicates the lower limit of its flight altitude. Indicates the upper limit of its crawl / dive angle; Indicates the speed of the UAV, Indicates the upper limit of its movement speed. represents the upper limit of its angular velocity, Indicates its upper acceleration limit; Indicates the distance of one-way communication from UAV to USA, Indicates the upper limit of its communication distance; represents the target detection range of the UAV, Indicates the upper limit of its detection distance.

[0012] Preferably, in step S1, kinematic modeling of the USV unmanned cluster is performed as follows: ; in, Represents the USV position vector, including the three-dimensional coordinates of the USV And heading angle ; Represents the velocity vector of the USV, including the movement speed of the USV , angular velocity , acceleration ; The constraints are as follows: ; in, Represents the constraints of the USV unmanned cluster; represents the detection space of USV, Indicates the upper limit of its detection space, Indicates the height of the boat. Indicates the upper limit of its heading angle; represents the movement speed of USV, Indicates the upper limit of its movement speed. represents the upper limit of its angular velocity, Indicates its upper acceleration limit; Indicates that the USV itself can bear the load, Indicates its load limit; Indicates the distance of one-way communication from USV to AUV, Indicates the upper limit of its communication distance; Indicates the target detection distance of USV, Indicates the upper limit of its detection distance.

[0013] Preferably, in step S1, kinematic modeling of the AUV unmanned cluster is performed as follows: ; in, Represents the AUV position vector, including the 3D coordinates of the AUV And heading angle ; Represents the velocity vector of the AUV, including the movement speed of the AUV , angular velocity , acceleration ; The constraints are as follows: ; in, Represents the constraints of the AUV unmanned cluster; represents the detection space of AUV, Indicates the upper limit of its detection space, Indicates the deepest working altitude underwater. Indicates the upper limit of its heading angle; represents the movement speed of the AUV, Indicates the upper limit of its movement speed. represents the upper limit of its angular velocity, Indicates its upper acceleration limit; Indicates the endurance of the AUV itself. Indicates the upper limit of its endurance; represents the target detection distance of the AUV, Indicates the upper limit of its detection distance.

[0014] Preferably, in step S2, the optimization objectives are to have more straight paths and fewer twists and turns, shorter total distances, and shorter total time consumption, and the cooperative navigation planning objective function is constructed through weighted fusion, as follows: ; in, Indicates the type of unmanned cluster, with a value range of 1-3, representing UAV unmanned cluster, USV unmanned cluster and AUV unmanned cluster respectively; j Indicates the first j individual; Represents the turning energy consumption function, and each turning radius , Total number of turns related, represents its weighting coefficient; Represents the collision penalty function, the closest distance to each obstacle encountered , the radius of the obstacle itself , Total number of obstacles related, represents its weighting coefficient; Represents the time limit function, and the total distance , real-time motion speed related, Represents its weighting coefficient.

[0015] Preferably, in step S3, the inertia weight is nonlinearly decreased as follows: ; in, , They are respectively represented as the maximum and minimum values ​​of the inertia weight, , They represent the steepness and midpoint of the curve respectively. represents the maximum number of iterations, represents the nonlinear decreasing inertia weight, Indicates the current iteration number.

[0016] Preferably, in step S3, the asymmetric learning factor is as follows: ; ; in, Represents the iteration t +1 individual learning factor, , Respectively represent the maximum and minimum extreme values ​​of individual learning factors; Represents the iteration t +1 Social Learning Factor; , Respectively represent the maximum and minimum extreme values ​​of the social learning factor.

[0017] Therefore, the present invention adopts the above-mentioned air-sea cross-domain unmanned intelligent cluster collaborative path planning method, and the beneficial technical effects are as follows: (1) More sophisticated modeling: This approach fully combines the characteristics of UAV’s wide detection range, USV’s large payload, and AUV’s high maneuverability, and takes into account constraints such as UAV’s climbing angle, USV’s navigation altitude, and AUV’s endurance time, to complete the refined kinematic modeling of various unmanned systems.

[0018] (2) Performance advantages: An IPSO optimization algorithm is proposed, which introduces nonlinear decreasing inertia weight and asymmetric learning factor in the iterative update process to avoid the algorithm falling into local optimality and accelerate the convergence speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of a method for collaborative path planning of an air-sea cross-domain unmanned intelligent cluster in the present invention; Figure 2 It is a collaborative path planning algorithm based on improved IPSO. DETAILED DESCRIPTION

[0020] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0021] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0022] Embodiment 1 like Figure 1 As shown, it is a flow chart of a method for collaborative path planning of an air-sea cross-domain unmanned intelligent cluster in the present invention, which specifically includes the following steps: 1. Kinematic modeling of UAV swarm.

[0023] UAV is flying in a two-dimensional plane at low altitude, and its motion model can be described as follows: ; in, is the position vector of the UAV, including the three-dimensional coordinates of the UAV And heading angle ; is the velocity vector of the UAV, including the velocity of the UAV , angular velocity , acceleration ; Indicates transpose.

[0024] The UAV's position vector With velocity vector The following relationship exists: ; in, Represents the first-order derivative of the horizontal displacement of the UAV, i.e., the horizontal velocity component; It represents the first derivative of the vertical displacement of the UAV, i.e., the vertical velocity component; Represents the first derivative of the UAV heading angle, i.e., angular velocity; Represents the first derivative of the UAV velocity, i.e., acceleration.

[0025] The UAV follows the following constraints during movement: ; in, Represents the constraints of the UAV unmanned cluster; represents the detection space of the UAV, Indicates the upper limit of its detection space, Indicates the lower limit of its flight altitude. Indicates the upper limit of its crawl / dive angle; Indicates the speed of the UAV, Indicates the upper limit of its movement speed. represents the upper limit of its angular velocity, Indicates its upper acceleration limit; Indicates the distance of one-way communication from UAV to USA, Indicates the upper limit of its communication distance; represents the target detection range of the UAV, Indicates the upper limit of its detection distance; Among them, the first line stipulates the upper limit of the UAV detection space, the lower limit of the flight altitude, and the upper limit of the crawl / dive angle. The second line stipulates the upper limit of the UAV motion performance. The third line stipulates the upper limit of the distance of the UAV's one-way communication to the USA. The fourth line stipulates the upper limit of the UAV's detection target distance.

[0026] 2. Kinematic modeling of USV unmanned cluster.

[0027] USV is sailing in a two-dimensional plane on the water surface, and its motion model can be expressed as follows: ; in, is the position vector of the USV, including the three-dimensional coordinates of the USV And heading angle ; is the velocity vector of the USV, including the movement speed of the USV , angular velocity , acceleration .

[0028] Position vector of USV With velocity vector There is a relationship that is consistent with the second formula.

[0029] USV follows the following constraints during movement: ; in, Represents the constraints of the USV unmanned cluster; represents the detection space of USV, Indicates the upper limit of its detection space, Indicates the height of the boat. Indicates the upper limit of its heading angle; represents the movement speed of USV, Indicates the upper limit of its movement speed. represents the upper limit of its angular velocity, Indicates its upper acceleration limit; Indicates that the USV itself can bear the load, Indicates its load limit; Indicates the distance of one-way communication from USV to AUV, Indicates the upper limit of its communication distance; Indicates the target detection distance of USV, Indicates the upper limit of its detection distance; Among them, the first line stipulates the upper limit of the USV's detection space, the upper and lower limits of the surface navigation altitude (not exceeding half of its own height), and the upper limit of the heading angle. The second line stipulates the upper limit of the UAV's motion performance. The third line stipulates the upper limit of the USV's own load. The fourth line stipulates the upper limit of the distance for one-way communication between the USV and the AUV. The fifth line stipulates the upper limit of the USV's detection target distance.

[0030] 3. Kinematic modeling of AUV unmanned swarm.

[0031] The AUV navigates in a two-dimensional plane on the water surface, and its motion model can be expressed as follows: ; in, is the position vector of the AUV, including the 3D coordinates of the AUV And heading angle ; is the velocity vector of the AUV, including the movement speed of the AUV , angular velocity , acceleration .

[0032] The AUV's position vector With velocity vector There is a relationship that is consistent with the second formula.

[0033] The AUV follows the following constraints during movement: ; in, Represents the constraints of the AUV unmanned cluster; represents the detection space of AUV, Indicates the upper limit of its detection space, Indicates the deepest working altitude underwater. Indicates the upper limit of its heading angle; represents the movement speed of the AUV, Indicates the upper limit of its movement speed. represents the upper limit of its angular velocity, Indicates its upper acceleration limit; Indicates the endurance of the AUV itself. Indicates the upper limit of its endurance; represents the target detection distance of the AUV, Indicates the upper limit of its detection distance; Among them, the first line stipulates the upper limit of the AUV's detection space, the upper and lower limits of the underwater navigation altitude (not less than the maximum working depth and not higher than the water surface), and the upper limit of the heading angle. The second line stipulates the upper limit of the AUV's motion performance. The third line stipulates the upper limit of the AUV's own endurance. The fourth line stipulates the upper limit of the AUV's detection target distance.

[0034] 4. Objective function construction.

[0035] For the planning paths of three types of unmanned intelligent swarms at different stages, the optimization goals are to have more straight paths and fewer twists, shorter total distances, and shorter total time. The influencing factors of the above goals are analyzed as turning energy consumption function, collision penalty function, and time limit function. The cooperative navigation planning objective function can be formed through weighted fusion. i The first in the unmanned cluster j The objective function formula of each individual is as follows: ; in, The value range is 1-3, representing UAV unmanned cluster, USV unmanned cluster and AUV unmanned cluster respectively; j This is the first unmanned cluster in this category. j individual; is the turning energy consumption function, and each turning radius , Total number of turns related, is its weighting coefficient; is the collision penalty function, the closest distance to each obstacle encountered , the radius of the obstacle itself , Total number of obstacles related, is its weighting coefficient; is the time limit function, and the total distance , real-time motion speed related, Its weighting coefficient.

[0036] 5. Collaborative path planning algorithm based on improved particle swarm optimization (IPSO).

[0037] It is known that the particle swarm optimization (PSO) is an optimization algorithm that imitates the bird hunting process. In the algorithm iterative optimization process, the position and speed of each particle are updated to determine the direction of the next iterative optimization. The first j The speed of the particle ,Location The update equation is as follows: ; ; in, is the inertia weight, , They are individual learning factor and social learning factor. , is a random real number in the interval [0,1]. For the The first j The optimal solution for a particle is No. Global optimal solution for particle-like populations; For the The first j The position of a particle at the last moment.

[0038] The following improvements are made to the above PSO algorithm: Nonlinear decreasing inertia weight .

[0039] Traditional PSO algorithm is a constant, which leads to slow convergence of the optimization algorithm. The introduction of nonlinear decreasing factor accelerates the convergence of the algorithm. The specific formula is as follows: ; in, , are the maximum and minimum values ​​of inertia weight, , is the steepness and midpoint position of the curve, is the maximum number of iterations.

[0040] Asymmetric learning factor , .

[0041] Traditional PSO algorithm It reflects the particle's ability to learn and evolve. This will cause the algorithm to fall into a local optimum. It reflects the shared learning ability of the group. This will cause the algorithm to converge very slowly. Introducing an asymmetric learning factor, emphasizing Self-learning ability, emphasized in the later stage of the algorithm Global learning ability. The specific formula is as follows: ; ; Collaborative path planning algorithm.

[0042] The design idea of ​​this algorithm is: first focus on a certain type of unmanned cluster , among which j The individual is used as the first The first j particles, with the objective function is the function fitness of each particle, For the corresponding The optimization constraint boundary of the unmanned cluster is substituted into the IPSO algorithm to iteratively calculate the , , as the first The optimal individual planning path and the optimal cluster collaborative planning path in the unmanned cluster, and then to the next unmanned cluster in sequence Repeat the above process. The specific algorithm flow is as follows Figure 2 shown.

[0043] The specific steps of the above algorithm are as follows: 1. Set , focus on Quasi-unmanned cluster as particle swarm to be optimized i ; 2. Initialize the parameters of the ISPO algorithm, including Kinematic model and constraints of unmanned swarm And other parameters; 3. Determine whether the algorithm meets the stopping condition (exceeding the number of iterations or the threshold range). If yes, jump to step 9, otherwise proceed to the next step; 4. Update particle swarm The speed of each particle in and location ; 5. Update nonlinear decreasing inertia weight ; 6. Update asymmetric learning factors , ; 7. Calculation of particle swarm The function fitness of each particle in ; 8. Calculation of particle swarm The individual optimum , global optimal , return to step 3; 9. Output particle swarm The optimal path planning result; 10. Judgment Is it less than 3? , return to step 2; otherwise the task ends.

[0044] It is worth noting that the contents not elaborated in detail in the present invention are all prior art and are well known to those skilled in the art.

[0045] Therefore, the present invention adopts the above-mentioned air-sea cross-domain unmanned intelligent cluster collaborative path planning method, which can make full use of the advantages between different platforms, and perform various collaborative tasks such as target search, tracking, and disposal in low-altitude, water surface, underwater and other spatial areas, further expand the application scope and resource utilization, have more sophisticated modeling and performance advantages, avoid the algorithm from falling into local optimality, and accelerate the convergence speed.

[0046] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for collaborative path planning of an air-sea cross-domain unmanned intelligent cluster, characterized in that: The following steps are involved: Step S1, kinematic modeling of the UAV unmanned cluster, the USV unmanned cluster and the AUV unmanned cluster; Step S2, taking the path with more straight lines and less twists, shorter total route length and shorter total time as the optimization goals, constructing the collaborative navigation planning objective function through weighted fusion; Step S3: Use the collaborative path planning algorithm based on improved particle swarm to perform path planning: The particle swarm algorithm is improved by introducing nonlinear decreasing inertia weight and asymmetric learning factor to obtain the IPSO algorithm. The UAV unmanned cluster, USV unmanned cluster and AUV unmanned cluster are respectively used as the particle swarms to be optimized, and are substituted into the IPSO algorithm for iterative calculation. The function fitness of each particle is calculated according to the objective function, and the speed and position of the particle are updated until the stopping condition is met. Output the optimal path planning results for various types of unmanned swarms.

2. According to claim 1, a method for collaborative path planning of an air-sea cross-domain unmanned intelligent cluster is characterized in that: In step S1, the kinematic modeling of the UAV cluster is performed as follows: （1）； in, Represents the UAV position vector, including the three-dimensional coordinates of the UAV And heading angle ; Represents the velocity vector of the UAV, including the movement speed of the UAV , angular velocity , acceleration ; represents transpose; The constraints are as follows: （2）； in, Represents the constraints of the UAV unmanned cluster; represents the detection space of the UAV, Indicates the upper limit of its detection space, Indicates the lower limit of its flight altitude. Indicates the upper limit of its crawl / dive angle; Indicates the speed of the UAV, Indicates the upper limit of its movement speed. represents the upper limit of its angular velocity, Indicates its upper acceleration limit; Indicates the distance of one-way communication from UAV to USA, Indicates the upper limit of its communication distance; represents the target detection distance of the UAV, Indicates the upper limit of its detection distance.

3. The method for collaborative path planning of an air-sea cross-domain unmanned intelligent cluster according to claim 2 is characterized in that: In step S1, kinematic modeling of the USV unmanned cluster is performed as follows: （3）； in, Represents the USV position vector, including the three-dimensional coordinates of the USV And heading angle ; Represents the velocity vector of the USV, including the movement speed of the USV , angular velocity , acceleration ; The constraints are as follows: （4）； in, Represents the constraints of the USV unmanned cluster; represents the detection space of USV, Indicates the upper limit of its detection space, Indicates the height of the boat. Indicates the upper limit of its heading angle; represents the movement speed of USV, Indicates the upper limit of its movement speed. represents the upper limit of its angular velocity, Indicates its upper acceleration limit; Indicates that the USV itself can bear the load. Indicates its load limit; Indicates the distance of one-way communication from USV to AUV, Indicates the upper limit of its communication distance; Indicates the target detection distance of USV, Indicates the upper limit of its detection distance.

4. According to claim 3, a method for collaborative path planning of an air-sea cross-domain unmanned intelligent cluster is characterized in that: In step S1, kinematic modeling of the AUV unmanned cluster is performed as follows: （5）； in, Represents the AUV position vector, including the 3D coordinates of the AUV And heading angle ; Represents the velocity vector of the AUV, including the movement speed of the AUV , angular velocity , acceleration ; The constraints are as follows: （6）； in, Represents the constraints of the AUV unmanned cluster; represents the detection space of AUV, Indicates the upper limit of its detection space, Indicates the deepest underwater working altitude. Indicates the upper limit of its heading angle; represents the movement speed of the AUV, Indicates the upper limit of its movement speed. represents the upper limit of its angular velocity, Indicates its upper acceleration limit; Indicates the endurance of the AUV itself. Indicates the upper limit of its endurance; represents the target detection distance of the AUV, Indicates the upper limit of its detection distance.

5. The method for collaborative path planning of an air-sea cross-domain unmanned intelligent cluster according to claim 4 is characterized in that: In step S2, the optimization goals are to make the path more straight and less winding, the total route length shorter, and the total time shorter, and the cooperative navigation planning objective function is constructed through weighted fusion, as follows: （7）； in, Indicates the type of unmanned cluster, with a value range of 1-3, representing UAV unmanned cluster, USV unmanned cluster and AUV unmanned cluster respectively; j Indicates the first j individual; Represents the turning energy consumption function, and each turning radius , Total number of turns related, represents its weighting coefficient; Represents the collision penalty function, the closest distance to each obstacle encountered , the radius of the obstacle itself , Total number of obstacles related, represents its weighting coefficient; Represents the time limit function, and the total distance , real-time motion speed related, Represents its weighting coefficient.

6. The method for collaborative path planning of an air-sea cross-domain unmanned intelligent cluster according to claim 1 is characterized in that: In step S3, the inertia weight is nonlinearly decreased as follows: （8）； in, , Respectively represent the maximum and minimum values ​​of inertia weight, , They represent the steepness and midpoint of the curve respectively. represents the maximum number of iterations, represents the nonlinear decreasing inertia weight, Indicates the current iteration number.

7. The method for collaborative path planning of an air-sea cross-domain unmanned intelligent cluster according to claim 6 is characterized in that: In step S3, the asymmetric learning factor is as follows: （9）； （10）； in, Represents the iteration t +1 individual learning factor, , Respectively represent the maximum and minimum extreme values ​​of individual learning factors; Represents the iteration t +1 Social Learning Factor; , Respectively represent the maximum and minimum extreme values ​​of the social learning factor.

Images