Unmanned target vehicle path planning method based on improved artificial fish swarm algorithm
By adaptively adjusting the field of view and step size of the artificial fish school algorithm in the path planning of unmanned target vehicles, the problem of low path planning efficiency and local optimal solution caused by fixed field of view and step size is solved, and more efficient path planning is achieved result.
Patent Information
- Application Number
- CN202111490926.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-08
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2041-12-08
Smart Images

Figure CN114169607B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of unmanned target vehicles and relates to a path planning method for an unmanned target vehicle, and in particular to a path planning method for an unmanned target vehicle based on an improved artificial fish swarm algorithm. Background Art
[0002] With the advancement of science and technology, driverless cars have developed rapidly and have been widely used in the target vehicle industry. After simple modification, driverless cars can be used as unmanned target vehicles. Unmanned target vehicles are a kind of movable equipment that trainers need to use when conducting shooting training at the shooting range, and path planning is a key link in unmanned target vehicle technology. The path planning of unmanned target vehicles is to obtain the path with the least time for the unmanned target vehicle to start from the starting point, pass through all target positions and reach the end point. In the traditional unmanned target vehicle path planning method, the unmanned target vehicle travels on a pre-laid track, which cannot realize autonomous path planning, and the cost of laying tracks is high. The unmanned target vehicle path planning method based on magnetic guidance or camera guidance can realize the autonomous planning path of the unmanned target vehicle, but this method can only realize linear motion and the motion mode is single. The typical method of the existing unmanned target vehicle path planning method is to realize the multi-directional movement of the unmanned target vehicle based on the improved ant colony algorithm, and finally obtain the path planning result through the transfer probability and pheromone update, which improves the speed of the unmanned target vehicle's autonomous path planning. However, the improved ant colony algorithm has a slow convergence speed due to the same pheromone initial value and is prone to fall into local optimality during pheromone updating, which affects the efficiency of the unmanned target vehicle path planning.
[0003] The artificial fish school algorithm is an intelligent algorithm for solving optimization problems based on the study of fish behavior in nature. It means that in a body of water, fish can often find places with abundant nutrients on their own or by following other fish. The artificial fish school algorithm determines the solution to the problem by comparing the fitness values of fish after they perform grouping, tail-chasing or foraging behaviors. When the algorithm is initialized, it is necessary to fix parameters such as the field of view and step length of the artificial fish school, which will affect the algorithm's solution process and calculation amount, and thus affect the efficiency of obtaining the optimal solution.
[0004] In the prior art, no path planning method for unmanned target vehicles based on artificial fish swarm algorithm has been retrieved, and only the path planning method for unmanned aerial vehicles can be realized. For example, the application was published as CN112650287A, entitled "A path planning method for unmanned aerial vehicles based on artificial fish swarm", which discloses a path planning method for unmanned aerial vehicles based on the artificial fish swarm method. The method is based on the artificial fish swarm algorithm. The unmanned aerial vehicle adjusts its own step length through the visual step coefficient at each iteration, obtains the fitness value of the bulletin board after making a behavior selection, and judges whether the number of iterations or the fitness value reaches the preset value, and finally obtains the path planning result with the shortest flight time for the unmanned aerial vehicle. Although the method uses the artificial fish swarm algorithm with adjusted step length to improve the convergence speed and accuracy of the unmanned aerial vehicle in autonomous path planning, the field of view of the unmanned aerial vehicle in the iterative process adopts a fixed value, which makes the path planning of the unmanned aerial vehicle easy to fall into the local optimal solution, and the calculation amount is large, which further affects the further improvement of the efficiency of the unmanned aerial vehicle path planning. Summary of the invention
[0005] The purpose of the present invention is to address the deficiencies of the above-mentioned prior art and provide an unmanned target vehicle path planning method based on an improved artificial fish swarm algorithm, so as to solve the technical problem of low efficiency in unmanned target vehicle path planning caused by the use of fixed fish swarm field of view and step length in the existing artificial fish swarm algorithm.
[0006] To achieve the above object, the technical solution adopted by the present invention includes the following steps:
[0007] (1) Initialize the unmanned target vehicle path planning scenario:
[0008] Initialization includes K unmanned target vehicles with labels A=(A1,A2,...,A k ,...,A K ) and M targets B=(B1,B2,...,B m ,...,B M ) unmanned target vehicle path planning scenario, where K ≥ 20, A k represents the kth unmanned target vehicle with a mark, M ≥ 30, B m represents the mth target;
[0009] (2) Initialize the parameters of the improved artificial fish swarm algorithm:
[0010] The number of initialization iterations is t, the maximum number of iterations is T, T ≥ 50, the number of attempts and the maximum number of attempts are c and u respectively, u ≥ 10, and each unmanned target vehicle A k Current status is the current position of M targets B, The fitness value is , the current status of the bulletin board is the fitness value of K unmanned target vehicles A with labels The current state of the unmanned target vehicle corresponding to the minimum value, X t The fitness value is Y t ,in, Indicates the mth target B m The current position of , and let t = 0, c = 0;
[0011] (3) Each unmanned target vehicle makes adaptive adjustments to the current field of view and step length:
[0012] Each unmanned target vehicle A k Calculate the average distance between itself and other K-1 unmanned target vehicles , and As the field of view of its own t-th iteration, Calculate the step size of its own t-th iteration
[0013] (4) Each unmanned target vehicle obtains the current optimal path status:
[0014] (4a) Each unmanned target vehicle A k Explore the current field of view The target vehicle A with the smallest fitness value among multiple unmanned target vehicles z , and judge its current state The fitness value of With A z Current Status The fitness value of Is it satisfied? , if so, to the unmanned target vehicle A z If the vehicle moves in the same direction, the rear-end collision is successful, and step (5) is executed; otherwise, step (4b) is executed;
[0015] (4b) Each unmanned target vehicle A k Select the field of view Unmanned target vehicle A at the center of multiple unmanned target vehicles e , and judge A k Current Status The fitness value of With A e Current Status The fitness value of Is it satisfied? , if so, to the unmanned target vehicle A e If the cluster moves in the same direction, the clustering is successful and step (5) is executed. Otherwise, step (4c) is executed.
[0016] (4c) Each unmanned target vehicle A k Randomly select a An unmanned target vehicle A j , and judge A k Current Status The fitness value of With A j Current Status The fitness value of Is it satisfied? If so, send the unmanned target vehicle A j If the foraging is successful, execute step (5); otherwise, execute step (4d);
[0017] (4d) Each unmanned target vehicle A k Determine whether c≥u is true. If so, within its field of view Randomly select an unmanned target vehicle A v Afterwards, to A v Move in the same direction and execute step (5); otherwise, set c = c + 1 and execute step (4c);
[0018] (5) Obtain the path planning results of the unmanned target vehicle:
[0019] Determine whether t≥T is satisfied. If so, set the state of the bulletin board after the Tth iteration X T The connection of the M states is used as the optimal planning path of the K marked unmanned target vehicles A. Otherwise, let t = t + 1 and execute step (6);
[0020] (6) Each unmanned target vehicle obtains the current status of the bulletin board:
[0021] Each unmanned target vehicle A k Opposing unmanned target vehicle A z Direction, A e Direction, A j Direction or A v The state after the direction moves and fitness value Update and determine its own status after the update The fitness value of and the fitness value Y of the bulletin board at time t-1 t-1 Is it satisfied? If so, update the state of itself As the current state of the bulletin board, the state of the bulletin board is updated and step (3) is executed; otherwise, the state of the bulletin board remains unchanged and step (3) is executed.
[0022] Compared with the prior art, the present invention has the following advantages:
[0023] In the process of obtaining the path planning result of the unmanned target vehicle, the first iteration of the present invention is to calculate the field of view and the step length of each unmanned target vehicle through the initialized parameters, and each subsequent iteration is to adaptively adjust the field of view and the step length of each unmanned target vehicle according to the state of each unmanned target vehicle after the unmanned target vehicle moves to a rear-end, a cluster, or a successful foraging, or a random direction, thereby avoiding the defects of the unmanned target vehicle being difficult to get rid of the local extreme value, the slow algorithm convergence speed, and the large amount of calculation caused by using the existing artificial fish school algorithm in the path planning of the unmanned target vehicle due to multiple iterations using a fixed field of view range and adjusting the step length through the view step coefficient, thereby effectively improving the efficiency of the path planning. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 It is a flow chart for implementing the present invention. DETAILED DESCRIPTION
[0025] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.
[0026] Reference Figure 1 , the present invention comprises the following steps:
[0027] Step 1) Initialize the unmanned target vehicle path planning scenario:
[0028] Initialization includes K unmanned target vehicles with labels A=(A1,A2,...,A k ,...,A K ) and M targets B=(B1,B2,...,B m ,...,B M ) unmanned target vehicle path planning scenario, where K ≥ 30, A k represents the kth unmanned target vehicle with a mark, M ≥ 30, B m represents the mth target, and the positions of the M targets are {(x1,y1),...,(x m ,y m ),...,(x M ,y M )}, in this embodiment, K=20, M=35.
[0029] Step 2) Initialize the improved artificial fish swarm algorithm parameters:
[0030] The number of initialization iterations is t, the maximum number of iterations is T, T ≥ 100, the number of attempts and the maximum number of attempts are c and u respectively, u ≥ 10, and each unmanned target vehicle A k Current status is the current position of M targets B, The fitness value is , the current status of the bulletin board is the fitness value of K unmanned target vehicles A with labels The current state of the unmanned target vehicle corresponding to the minimum value, X t The fitness value is Y t ,in, Indicates the mth target B m The current position of , and let t = 0, c = 0, in this embodiment T = 100, u = 10;
[0031] Step 3) Each unmanned target vehicle makes adaptive adjustments to the current field of view and step length:
[0032] Each unmanned target vehicle A k Calculate the average distance between itself and other K-1 unmanned target vehicles , and As the field of view of its own t-th iteration, Calculate the step size of its own t-th iteration , the calculation formula is;
[0033]
[0034]
[0035]
[0036]
[0037] Among them, G = K-1, g ≠ k, Indicates the t-th iteration of unmanned target vehicle A k To unmanned target vehicle A g The distance between them, ∑ represents the accumulation operation, sign(w) represents the sign function, and Respectively represent the t-th iteration of unmanned target vehicle A k and unmanned target vehicle A g The mth target B m The current position of , α represents the visual step coefficient, 0<α≤1;
[0038] The factors that affect the path planning of the artificial fish swarm algorithm mainly include field of view and step size. The field of view determines the range that the unmanned target vehicle can see. If the field of view is too small, it is rare to see other unmanned target vehicles, and multiple explorations are required, which increases the amount of calculation. If the field of view is too large, rear-end collisions and clustering are more frequent, and local searches are easily ignored, which may miss the global optimal solution. The step size determines the local search speed and movement speed of the unmanned target vehicle. If the step size is small, the unmanned target vehicle moves slowly and continuously performs local searches, which is easy to fall into local optimality and increases the amount of calculation. If the step size is large, the unmanned target vehicle may fall outside the field of view, thus missing the global optimal solution.
[0039] Each unmanned target vehicle of the present invention calculates the average distance between the unmanned target vehicles and the field of view according to the current state. and step length Adaptive adjustment is performed to solve the problem that the fixed field of view in the existing artificial fish swarm algorithm makes it difficult to get rid of local extreme values and the amount of calculation is large during path planning, the speed of unmanned target vehicle path planning is improved, and the disadvantage of low efficiency in unmanned target vehicle path planning is overcome. It also avoids the defects of slow convergence speed and easy falling into local optimality during pheromone update in the unmanned target vehicle path planning method based on the improved ant colony algorithm, which affects the efficiency of unmanned target vehicle path planning. In this embodiment, α=0.6.
[0040] Step 4) Each unmanned target vehicle obtains the current optimal path state:
[0041] Step 4a) Each unmanned target vehicle A k Explore the current field of view The target vehicle A with the smallest fitness value among multiple unmanned target vehicles z , and judge its current state The fitness value of With A z Current Status The fitness value of Is it satisfied? , if so, to the unmanned target vehicle A z If the vehicle moves in the same direction, the rear-end collision is successful, and step (5) is executed; otherwise, step (4b) is executed;
[0042] Step 4b) Each unmanned target vehicle A k Select the field of view Unmanned target vehicle A at the center of multiple unmanned target vehicles e , and judge A k Current Status The fitness value of With A e Current Status The fitness value of Is it satisfied? , if so, to the unmanned target vehicle A e If the cluster moves in the same direction, the clustering is successful and step (5) is executed. Otherwise, step (4c) is executed.
[0043] Step 4c) Each unmanned target vehicle A k Randomly select a An unmanned target vehicle A j , and judge A k Current Status The fitness value of With Aj Current Status The fitness value of Is it satisfied? If so, send the unmanned target vehicle A j If the foraging is successful, execute step (5); otherwise, execute step (4d);
[0044] Step 4d) Each unmanned target vehicle A k Determine whether c≥u is true. If so, within its field of view Randomly select an unmanned target vehicle A v Afterwards, to A v Move in the same direction and execute step (5); otherwise, set c = c + 1 and execute step (4c);
[0045] Step 5) Obtain the unmanned target vehicle path planning results:
[0046] Determine whether t≥T is satisfied. If so, set the state of the bulletin board after the Tth iteration X T The connection of the M states is used as the optimal planning path of the K marked unmanned target vehicles A. Otherwise, let t = t + 1 and execute step (6);
[0047] Step 6) Each unmanned target vehicle obtains the current status of the bulletin board:
[0048] Each unmanned target vehicle A k Opposing unmanned target vehicle A z Direction, A e Direction, A j Direction or A v The state after the direction moves and fitness value Update, the update formula is:
[0049]
[0050]
[0051] in, is the unmanned target vehicle A of the tth iteration k Updated status, is the unmanned target vehicle A of the tth iteration h Corresponding to the current state, h takes one of z, e, j and v, and Rand() generates a random number from 0 to 1. For each unmanned target vehicle A in the tth iteration k The step length, Indicates that each unmanned target vehicle A in the tth iteration k The position coordinates, (x m ,y m) is the position coordinate of the mth target;
[0052] Each unmanned target vehicle A k Determine the updated status The fitness value of and the fitness value Y of the bulletin board at time t-1 t-1 Is it satisfied? If so, update the state of itself As the current state of the bulletin board, the state of the bulletin board is updated and step (3) is executed; otherwise, the state of the bulletin board remains unchanged and step (3) is executed.
Claims
1. A path planning method for an unmanned target vehicle based on an improved artificial fish swarm algorithm, characterized in that: The steps include: (1) Initialize the unmanned target vehicle path planning scenario: Initialization includes K unmanned target vehicles with labels A=(A1,A2,...,A k ,...,A K ) and M targets B=(B1,B2,...,B m ,...,B M ) unmanned target vehicle path planning scenario, where K ≥ 20, A k represents the kth unmanned target vehicle with a mark, M ≥ 30, B m represents the mth target; (2) Initialize the parameters of the improved artificial fish swarm algorithm: The number of initialization iterations is t, the maximum number of iterations is T, T ≥ 50, the number of attempts and the maximum number of attempts are c and u respectively, u ≥ 10, and each unmanned target vehicle A k Current status is the current position of M targets B, The fitness value is The current status of the notice board is the fitness value of K unmanned target vehicles A with labels The current state of the unmanned target vehicle corresponding to the minimum value, X t The fitness value is Y t ,in, Indicates the mth target B m The current position of , and let t = 0, c = 0; (3) Each unmanned target vehicle makes adaptive adjustments to the current field of view and step length: Each unmanned target vehicle A k Calculate the average distance between itself and other K-1 unmanned target vehicles and will As the field of view of its own t-th iteration, Calculate the step size of its own t-th iteration (4) Each unmanned target vehicle obtains the current optimal path status: (4a) Each unmanned target vehicle A k Explore the current field of view The target vehicle A with the smallest fitness value among multiple unmanned target vehicles z , and judge its current state The fitness value of With A z Current Status The fitness value of Is it satisfied? If so, send the unmanned target vehicle A z If the vehicle moves in the same direction, the rear-end collision is successful, and step (5) is executed; otherwise, step (4b) is executed; (4b) Each unmanned target vehicle A k Select the field of view Unmanned target vehicle A at the center of multiple unmanned target vehicles e , and judge A k Current Status The fitness value of With A e Current Status The fitness value of Is it satisfied? If so, send the unmanned target vehicle A e If the cluster moves in the same direction, the clustering is successful and step (5) is executed. Otherwise, step (4c) is executed. (4c) Each unmanned target vehicle A k Randomly select a An unmanned target vehicle A j , and judge A k Current Status The fitness value of With A j Current Status The fitness value of Is it satisfied? If so, send the unmanned target vehicle A j If the foraging is successful, execute step (5); otherwise, execute step (4d); (4d) Each unmanned target vehicle A k Determine whether c≥u is true. If so, within its field of view Randomly select an unmanned target vehicle A v Afterwards, to A v Move in the same direction and execute step (5); otherwise, set c = c + 1 and execute step (4c); (5) Obtain the path planning results of the unmanned target vehicle: Determine whether t≥T is satisfied. If so, set the state of the bulletin board after the Tth iteration X T The connection of the M states is used as the optimal planning path of the K marked unmanned target vehicles A. Otherwise, let t = t + 1 and execute step (6); (6) Each unmanned target vehicle obtains the current status of the bulletin board: Each unmanned target vehicle A k Opposing unmanned target vehicle A z Direction, A e Direction, A j Direction or A v The state after the direction moves and fitness value Update and determine its own status after the update The fitness value of and the fitness value Y of the bulletin board at time t-1 t-1 Is it satisfied? If so, update the state of itself As the current state of the bulletin board, the state of the bulletin board is updated and step (3) is executed; otherwise, the state of the bulletin board remains unchanged and step (3) is executed.
2. The unmanned target vehicle path planning method based on improved artificial fish swarm algorithm according to claim 1, characterized in that: Each unmanned target vehicle A described in step (3) k Calculate the average distance between itself and other K-1 unmanned target vehicles And calculate the step size of its own t-th iteration The calculation formulas are: Among them, G = K-1, g ≠ k, Indicates the t-th iteration of unmanned target vehicle A k To unmanned target vehicle A g The distance between them, ∑ represents the accumulation operation, sign(w) represents the sign function, and Respectively represent the t-th iteration of unmanned target vehicle A k and unmanned target vehicle A g The mth target B m The current position of , α represents the visual step coefficient, 0<α≤1.
3. The unmanned target vehicle path planning method based on improved artificial fish swarm algorithm according to claim 1, characterized in that: Each unmanned target vehicle A described in step (6) k Opposing unmanned target vehicle A z Direction, A e Direction, A j Direction or A v The state after the direction moves and fitness value Update, the update formula is: in, is the unmanned target vehicle A of the tth iteration k Updated status, is the unmanned target vehicle A of the tth iteration h Corresponding to the current state, h takes one of z, e, j and v, and Rand() generates a random number from 0 to 1. For each unmanned target vehicle A in the tth iteration k The step length, Indicates that each unmanned target vehicle A in the tth iteration k The position coordinates, (x m ,y m ) is the position coordinate of the mth target.
Citation Information
Patent Citations
Unmanned aerial vehicle path planning method based on artificial fish schools
CN112650287A
Artificial fish swarm algorithm based traffic route searching method
CN104392283A
Global path planning method based on improved quantum particle swarm optimization algorithm
CN111080035A
Intelligent unmanned target vehicle distance control method based on fuzzy control
CN111856943A
Multi-unmanned-target-vehicle path planning method based on improved ant colony algorithm
CN111880542A
