A UAV path planning method based on multi-strategy improved ant colony algorithm

Abstract

The present invention discloses a UAV path planning method based on a multi-strategy improved ant colony algorithm, and relates to the technical field of UAV path planning. In view of the problems existing in the traditional ant colony algorithm when performing UAV path planning, such as slow convergence speed, easy to fall into local optimal solution, complex parameter adjustment, etc., the present invention proposes a UAV path planning method based on a multi-strategy improved ant colony algorithm, including: introducing an angle function in the state transition probability to provide directional information, thereby achieving the purpose of accelerating convergence speed and reducing corners; adopting the strategy of inward expansion of pheromone and limiting pheromone concentration in the pheromone update rule to avoid the problem of premature convergence and falling into local optimal solution of the algorithm; dynamically adjusting the pheromone volatilization factor in an adaptive manner, and through pre-experimental processing, designing the objective function to dynamically adjust the value of the corresponding parameter to obtain the optimal parameter combination, thereby improving the convergence effect of the algorithm. The present invention can improve various performances in the UAV path planning process, and is more suitable for large-scale complex environments.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) path planning, and in particular to a UAV path planning method based on a multi-strategy improved ant colony algorithm. Technical Background

[0002] Drones, with their robust autonomy and flexibility, have become a vital tool for meeting human needs and facilitating work in our lives. Current drone research and development has not only driven progress in numerous fields, including industrial automation and intelligent agricultural production, but also provided convenient services in everyday areas such as medical rescue and cargo transportation. Perception and autonomous navigation technologies are core areas of drone research, enabling drones to autonomously plan routes based on environmental information acquired through sensors. Therefore, path planning, as a key component of autonomous navigation, has become a research hotspot in the drone field.

[0003] Path planning refers to the process of finding a path from a starting point to a destination safely while avoiding obstacles, and optimizing its performance as much as possible. Typical path planning algorithms include Dijkstra's algorithm, A* algorithm, artificial potential field method, rapidly expanding random tree algorithm, genetic algorithm, ant colony algorithm, particle swarm algorithm, etc. However, during the path planning process, drones need to consider information such as static known obstacles, map environment complexity, and target point location to plan the global path. As a global path planning algorithm that uses distributed computing, the ant colony algorithm has the advantage of being highly robust and having lower requirements for the initial route compared to other algorithms. It is a self-organizing algorithm with strong adaptability to different problem environments. It is parallel and has strong global search capabilities, and it is easy to integrate with other algorithms.

[0004] However, the traditional ant colony algorithm still has the following shortcomings in the global path planning process:

[0005] (1) Slow convergence: In the initial stage of the algorithm, the probability of ants choosing all paths is the same, so the search process is relatively random and blind. Especially when dealing with complex or large-scale problems, the search efficiency is too low.

[0006] (2) Too many inflection points: The traditional ant colony algorithm does not consider the size of the corners and the number of inflection points, which will lead to excessive energy loss and low safety factor in actual application.

[0007] (3) Easy to fall into local optimal solution: Due to the positive feedback mechanism of the algorithm, the subsequent search path is concentrated on some candidate solutions, which reduces the diversity of the population and is not conducive to finding the global optimal solution.

[0008] (4) Complex parameter adjustment: The setting of parameters has a great impact on the performance of the algorithm, but there is no perfect theory to determine the value of the ant colony algorithm parameters. Usually, we can only rely on a large number of experiments and experience to give the optimal parameter combination. Summary of the Invention

[0009] In order to solve the above-mentioned problems of the ant colony algorithm from multiple angles, the main purpose of the present invention is to propose a drone path planning method based on a multi-strategy improved ant colony algorithm. By improving the state transition probability, pheromone update rules and parameter adjustment, the convergence speed of the ant colony algorithm is improved, the path length is reduced, and its various performances are optimized.

[0010] To achieve the above objectives, the present invention provides a drone path planning method based on a multi-strategy improved ant colony algorithm. To address the blindness of the search process and the high number of turning points in the planned path in traditional ant colony algorithms, the present invention introduces an angle function containing two factors, the steering angle and the corner angle, into the state transition probability. This provides the ants with a direction to the target point, reducing unnecessary random searches, accelerating convergence, and simultaneously reducing turning points and corners.

[0011] Furthermore, in order to improve the algorithm search effect, the present invention adjusts the pheromone volatility factor into the form of an adaptive parameter, so that it changes with the number of iterations and obeys a normal distribution, thereby ensuring the stability of the algorithm while increasing the diversity and flexibility of the search.

[0012] Furthermore, based on the pheromone update rules of the traditional algorithm, the present invention introduces a strategy of expanding pheromones inward and limiting pheromone concentration. While narrowing the search range of the optimal path, it is beneficial for ants to find better paths in subsequent searches, avoiding the accumulation of pheromones and directly converging on a non-optimal path to form a local optimal solution.

[0013] Finally, in order to address the problem that the ant colony algorithm has a strong dependence on parameters but a complex parameter adjustment process, the present invention conducts a preliminary experiment to dynamically adjust the values ​​of relevant parameters according to the objective function, and uses the parameter values ​​obtained at the end of the experimental run as the optimal parameter combination of the present invention.

[0014] The technical solution adopted in the present invention is:

[0015] A UAV path planning method based on a multi-strategy improved ant colony algorithm mainly includes the following steps:

[0016] Step 1: Create a two-dimensional grid map and a three-dimensional space map to determine the positions of the starting point S, the target point G, and all obstacles;

[0017] Step 2: Use the traditional ant colony algorithm to perform preliminary path planning, providing an algorithm platform for the UAV path planning method based on the multi-strategy improved ant colony algorithm;

[0018] Step 3: Introduce the angle function into the state transition probability and its weight γ, reducing unnecessary random searches and effectively solving the problem of blind search process and many turning points in the planning path;

[0019] Step 4: Dynamically and adaptively adjust the pheromone volatility factor ρ, so that the algorithm can continuously adjust the influence of pheromone on path selection as the number of iterations changes, improve the search effect at different stages of the algorithm, and increase the diversity of search while maintaining stability;

[0020] Step 5: Introduce the inward expansion strategy of pheromone into the pheromone update rule, and perform a secondary update on the pheromone concentrations between feasible points in some preferred areas to avoid the algorithm from converging prematurely and falling into local optimality, and to ensure the quality of subsequent search paths.

[0021] Step 6: limit the updated residual pheromone concentration to [τ min ,τ max ] range, to avoid excessive pheromone concentration on a certain path, which would cause the path to be suboptimal but the algorithm would converge because the pheromone concentration is much greater than that of other paths;

[0022] Step 7: Based on the above steps, a preliminary experiment is performed on the parameters α, β and γ, and dynamic adjustments are made according to the objective function. The values ​​of the three parameters at the end of the operation are used as the values ​​of the weight coefficients of the present invention, avoiding the complex process of determining the parameters through multiple experiments.

[0023] Furthermore, in order to speed up the convergence of the algorithm, the present invention introduces the angle function in step 3 The definition is as follows:

[0024]

[0025] Which includes the steering angle θ jg and the corner angle θ sj The steering angle is formed by connecting the current node with the next node and the target point respectively. The smaller the angle, the closer the direction of movement to the next node is to the target point, providing a certain directionality for the ant, thereby reducing unnecessary random searches and accelerating convergence. The corner angle is formed by connecting the current node with the previous node and the next node respectively. The larger the angle, the smoother the path, which can achieve the purpose of reducing turning points and corners.

[0026] Furthermore, after the present invention introduces the angle function and its weight, the state transition probability calculation formula is as follows:

[0027]

[0028] Among them, τ ij (t) is the pheromone concentration between node i and node j at time t, α is the pheromone weight, η ij (t) is the heuristic information representing the expected degree, β is the heuristic information weight, γ is the angle function weight, J k (i) is the set of nodes that ant k can visit in the next step.

[0029] Furthermore, in order to improve the search effect at different stages of the algorithm, the present invention introduces an adaptive pheromone volatility factor in step 4, which is defined as follows:

[0030]

[0031] Among them, σ is the scale parameter, a is the adjustment coefficient, N c is the current iteration number, N max is the maximum number of iterations. In the early stages of the algorithm, a small ρ value prevents excessive pheromone volatilization and weakening of the pheromone effect. In the middle stages of the algorithm, a large ρ value increases the diversity of the search and prevents premature convergence due to pheromone accumulation. In the later stages, a small ρ value increases pheromone guidance, allowing the algorithm to converge to the shortest path.

[0032] Furthermore, after the present invention introduces the adaptive pheromone volatilization factor, the pheromone update rules are as follows:

[0033]

[0034] Where m is the number of ants, is the pheromone concentration released by ant k between nodes i and j in this iteration, and its expression is

[0035]

[0036] Among them, Q is the total amount of pheromone released by ants during one traversal, L k is the length of the path traversed by ant k in this iteration.

[0037] Furthermore, to prevent the algorithm from falling into a local optimal solution, the present invention introduces an inward expansion strategy for pheromones in step 5. First, the ideal path is defined as the shortest path from the starting point to the target point without obstacles. The preferred path is the path that is shorter than the average path length in this iteration. The area where the preferred path is expanded inward to the ideal path is called the preferred area. Based on the pheromone update rules of the traditional algorithm, the present invention performs a secondary update, increasing the pheromone concentration between feasible points within some preferred areas. The calculation formula is as follows:

[0038]

[0039] Among them L avg is the average length of all paths in this iteration, L i The path length in this iteration is less than the average path. This narrows the optimal path search range and helps ants find better paths in the area during subsequent searches, avoiding the formation of local optimal solutions on non-optimal preferred paths due to the accumulation of pheromones.

[0040] Furthermore, considering that there may be too many obstacles within the preferred area, which may prevent the formation of a feasible path and cause ants to stagnate or bypass obstacles to form a longer path, the present invention determines whether to update the internal pheromone by judging the proportion of obstacles within the preferred area. If the proportion is small, indicating that there may be a shorter path within it, the pheromone in this area is updated; otherwise, no update is performed. In other words, the secondary update of pheromones must simultaneously meet the following constraints:

[0041]

[0042] Where n is the limit of the number of iterations, N obs N is the number of grids occupied by obstacles in the formed preferred area. all is the total number of grids in the preferred area, and δ is the threshold of the obstacle ratio.

[0043] Furthermore, in order to avoid excessive accumulation of pheromones leading to non-optimal path convergence, the present invention limits the updated residual pheromone concentration in step 6 to the following range:

[0044]

[0045] where τ min is the minimum pheromone concentration, τ max is the maximum pheromone concentration.

[0046] Furthermore, since there is currently no perfect theory to determine the values ​​of the ant colony algorithm parameters, the optimal parameter combination is usually given based on multiple experimental comparisons. Therefore, in step 7, the present invention conducts a preliminary experiment on the weight parameters α, β, and γ. By designing the objective function, the values ​​of the relevant parameters are dynamically adjusted. At the end of the operation, the obtained parameter values ​​are used as the optimal parameter combination of the improved ant colony algorithm. First, the weight parameters are set to α = 1, β = 1, and γ = 1, and then the corresponding parameters are adjusted according to the following conditions:

[0047] The pheromone weight α determines the degree of influence of the residual pheromone concentration on the path selection. Therefore, the present invention dynamically adjusts the parameter α according to the variance change of the path length after each iteration. When the variance of this iteration is larger than that of the previous one, it indicates that the search process is more random, and the value of α is increased at this time; when the variance is smaller, it indicates that there are fewer search paths and it is easy to produce local optimal solutions, and the value of α is reduced at this time. That is, the objective function is

[0048] f(k)=S(k)-S(k-1), k≥2

[0049] Where S(k) is the variance of all path lengths at the kth iteration. The dynamic change rule of the corresponding parameter α is as follows:

[0050]

[0051] Similarly, the heuristic information weight β determines the degree of influence of the next moving distance on the path selection. Therefore, the present invention dynamically adjusts β according to the shortest path length after each iteration. The corresponding objective function is

[0052] g(k)=L min (k)-L min (k-1),k≥2

[0053] Among them L min (k) is the shortest path length in the kth iteration. The dynamic change rule of the corresponding parameter β is as follows:

[0054]

[0055] The angle function weight γ determines the degree of influence of the angle information of the next node on the path selection. Therefore, the present invention dynamically adjusts γ according to the angle deviation between the shortest path and the ideal path after each iteration. The corresponding objective function is

[0056] h(k)=θ(k)-θ(k-1), k≥2

[0057] Where θ(k) is the sum of the angles formed by the target point and the lines connecting the turning points and the starting point of the shortest path in the kth iteration. The dynamic change rules of the corresponding parameter γ are as follows:

[0058]

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] The present invention adopts the above technical solution, so that when the UAV performs path planning in a global static environment, it provides the algorithm with directional information pointing to the target point, reduces unnecessary random searches, speeds up the convergence speed compared to traditional algorithms, and has greater advantages in large-scale complex environments. Secondly, the secondary update and the introduction of the adaptive pheromone volatilization factor in the pheromone update rule solve the problem that the traditional ant colony algorithm is prone to falling into local optimality, while optimizing the search effect of the algorithm at different stages. Finally, the problem of complex parameter adjustment is solved through pre-experimental processing. Therefore, the present invention improves the performance of the ant colony algorithm in many aspects, enables the UAV to reach the target location quickly and accurately, and reduces its energy loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0062] Figure 1 This is a flowchart of a UAV path planning method based on a multi-strategy improved ant colony algorithm;

[0063] Figure 2 This is a diagram showing the preliminary experimental results of the improved ant colony algorithm based on the present invention;

[0064] Figure 3(a) and Figure 3(b) are simulation diagrams of two-dimensional path planning results based on the traditional ant colony algorithm and the improved ant colony algorithm;

[0065] Figure 4 This is a comparison diagram of the convergence of two-dimensional path planning based on the traditional ant colony algorithm and the improved ant colony algorithm;

[0066] Figure 5 The simulation diagram of the three-dimensional path planning results based on the traditional ant colony algorithm and the improved ant colony algorithm;

[0067] Figure 6 This is a comparison diagram of the convergence of three-dimensional path planning based on traditional ant colony algorithm and improved ant colony algorithm. DETAILED DESCRIPTION

[0068] In order to make the improved purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the following embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0069] At present, when performing path planning in a global static environment, the traditional ant colony algorithm has the following shortcomings: the early search process of the algorithm is highly random, resulting in slow convergence, long search time, and many turning points; the algorithm is prone to concentrating the search path on some candidate solutions under the action of the positive feedback mechanism, thereby reducing the diversity of the population and causing the algorithm to fall into a local optimal solution; the weight parameters in the algorithm have a great influence on the experimental results, but usually can only rely on a large number of experiments and experience to determine the value, and the parameter adjustment process is relatively complicated. Based on this, the main purpose of this application is to propose a UAV path planning method based on a multi-strategy improved ant colony algorithm, which solves the problems of slow convergence, many turning points, easy to fall into local optimal solutions, and complex parameter adjustment when using the traditional ant colony algorithm for path planning by improving the state transition probability, pheromone update rules, and parameter adjustment.

[0070] To facilitate understanding of this embodiment, a drone path planning method based on a multi-strategy improved ant colony algorithm disclosed in an embodiment of the present invention is first introduced in detail.

[0071] Example:

[0072] A UAV path planning method based on a multi-strategy improved ant colony algorithm mainly includes the following steps:

[0073] Step 1: Create a two-dimensional grid map and a three-dimensional space map to determine the positions of the starting point S, the target point G, and all obstacles;

[0074] Step 2: Use the traditional ant colony algorithm to perform preliminary path planning, providing an algorithm platform for the UAV path planning method based on the multi-strategy improved ant colony algorithm;

[0075] Step 3: Introduce the angle function into the state transition probability and its weight γ, reducing unnecessary random searches and effectively solving the problem of blind search process and many turning points in the planning path;

[0076] Step 4: Dynamically and adaptively adjust the pheromone volatility factor ρ, so that the algorithm can continuously adjust the influence of pheromone on path selection as the number of iterations changes, improve the search effect at different stages of the algorithm, and increase the diversity of search while maintaining stability;

[0077] Step 5: Introduce the inward expansion strategy of pheromone into the pheromone update rule, and perform a secondary update on the pheromone concentrations between feasible points in some preferred areas to avoid the algorithm from converging prematurely and falling into local optimality, and to ensure the quality of subsequent search paths.

[0078] Step 6: limit the updated residual pheromone concentration to [τ min ,τ max] range, to avoid excessive pheromone concentration on a certain path, which would cause the path to be suboptimal but the algorithm would converge because the pheromone concentration is much greater than that of other paths;

[0079] Step 7: Based on the above steps, a preliminary experiment is performed on the parameters α, β and γ, and dynamic adjustments are made according to the objective function. The values ​​of the three parameters at the end of the operation are used as the values ​​of the weight coefficients of the present invention, avoiding the complex process of determining the parameters through multiple experiments.

[0080] Furthermore, in step 1, MATLAB R2016b simulation software is used to establish a two-dimensional grid map and a three-dimensional space map, and the positions of the starting point S, the target point G, and all obstacles are determined.

[0081] Furthermore, the two-dimensional path planning result obtained by using the traditional ant colony algorithm in step 2 is shown in Figure 3(a).

[0082] Furthermore, the angle function is introduced in step 3 The definition is as follows:

[0083]

[0084] where θ jg is the steering angle, which is formed by connecting the current node with the next node and the target point respectively; θ sj is the corner angle, which is formed by connecting the current node with the previous node and the next node respectively.

[0085] Furthermore, based on the above improvements, the state transition probability calculation formula is as follows:

[0086]

[0087] Among them, τ ij (t) is the pheromone concentration between node i and node j at time t, α is the pheromone weight, η ij (t) is the heuristic information representing the expected degree, β is the heuristic information weight, γ is the angle function weight, J k (i) is the set of nodes that ant k can visit in the next step.

[0088] Furthermore, in step 4, an adaptive pheromone volatility factor is introduced, which is defined as follows:

[0089]

[0090] Among them, σ is the scale parameter, a is the adjustment coefficient, N c is the current iteration number, N max is the maximum number of iterations, in this embodiment, N max =100.

[0091] Furthermore, based on the above improvements, the pheromone update rules are as follows:

[0092]

[0093] Where m is the number of ants, is the pheromone concentration released by ant k between nodes i and j in this iteration, and its expression is

[0094]

[0095] Among them, Q is the total amount of pheromone released by ants during one traversal, L k is the length of the path traversed by ant k in this iteration. In this embodiment, m=50 and Q=1.

[0096] Furthermore, the strategy of expanding pheromones inward is introduced in step 5. First, it is determined whether the preferred area meets the following constraints:

[0097]

[0098] Where n is the limit of the number of iterations, N obs N is the number of grids occupied by obstacles in the formed preferred area. all is the total number of grids in the preferred area, and δ is the threshold value of the obstacle ratio. In this embodiment, n=40 and δ=0.25.

[0099] If it is satisfied, the pheromone between the feasible points in the preferred area is updated twice as follows:

[0100]

[0101] Among them L avg is the average length of all paths in this iteration, L i are the paths whose path length is less than the average value in this iteration.

[0102] Furthermore, in step 6, the updated residual pheromone concentration is limited to the following range:

[0103]

[0104] where τ min is the minimum pheromone concentration, τ max is the maximum pheromone concentration.

[0105] Furthermore, in step 7, a preliminary experiment was conducted on the weight parameters α, β and γ. First, the weight parameters were set to α = 1, β = 1, γ = 1, and the objective functions corresponding to the design parameters α, β, and γ were

[0106] f(k)=S(k)-S(k-1), g(k)=L min (k)-L min (k-1),h(k)=θ(k)-θ(k-1)

[0107] Where k ≥ 2, S(k) is the variance of all path lengths at the kth iteration, L min (k) is the length of the shortest path in the kth iteration, and θ(k) is the sum of the angles formed by the lines connecting the target point and each turning point and starting point of the shortest path in the kth iteration.

[0108] The relevant parameters are dynamically adjusted according to the changes in the objective function after each iteration. The dynamic change rules are as follows:

[0109]

[0110] At the end of the run, the obtained parameter values ​​are used as the optimal parameter combination of the improved ant colony algorithm.

[0111] The results obtained based on the above preliminary experiments are shown in the attached Figure 2 shown.

[0112] Furthermore, the two-dimensional path planning result obtained based on the improved ant colony algorithm is shown in Figure 3(b), which is compared with the path planning convergence curve of the traditional ant colony algorithm. Figure 4 The three-dimensional path planning results based on the traditional ant colony algorithm and the improved ant colony algorithm are shown in the attached Figure 5 As shown in the figure, the path planning convergence curve is compared with the attached Figure 6 shown.

[0113] The following table compares the path planning simulation results in two-dimensional and three-dimensional environments based on the traditional ant colony algorithm and the improved ant colony algorithm proposed in this embodiment.

[0114] Table 1 Comparison of two-dimensional path planning simulation results before and after the ant colony algorithm improvement

[0115] algorithm Path length / m Running time / s Number of turns Iteration convergence times Traditional ant colony algorithm 46.6 26.54 15 38 Improved ant colony algorithm 43.3 33.93 7 8

[0116] Table 2 Comparison of 3D path planning simulation results before and after the ant colony algorithm improvement

[0117] algorithm Path length / m Iteration convergence times Number of turns Corner angle and Traditional ant colony algorithm 145.2 56 20 1770.4 Improved ant colony algorithm 131.3 25 19 1391.6

[0118] The simulation results of this embodiment show that when planning paths in a two-dimensional environment, the improved ant colony algorithm proposed in this invention converges faster, requires fewer turns, and achieves a significantly shorter optimal path length than the traditional algorithm. In a three-dimensional environment, the path length planned by the improved ant colony algorithm is 9.6% shorter than that of the traditional algorithm, the number of iterative convergences is reduced by 55.4%, and the number of turns is reduced by 21.4% compared to the traditional algorithm, resulting in a smoother path. Even though the improved algorithm takes longer to run due to the secondary update of pheromones, the above simulation results show that the improved ant colony algorithm proposed in this invention can converge to the global optimal solution more quickly, avoiding the problem of local optimality, and plans a more optimal path with better safety performance.

Claims

1. A UAV path planning method based on a multi-strategy improved ant colony algorithm, characterized in that: The following steps are involved: Step 1: Create a two-dimensional grid map and a three-dimensional space map to determine the positions of the starting point S, the target point G, and all obstacles; Step 2: Use the traditional ant colony algorithm to perform preliminary path planning, providing an algorithm platform for the UAV path planning method based on the multi-strategy improved ant colony algorithm; Step 3: Introduce the angle function into the state transition probability and its weight γ, reducing unnecessary random searches and effectively solving the problem of blind search process and many turning points in the planning path; Step 4: Dynamically and adaptively adjust the pheromone volatility factor ρ, so that the algorithm can continuously adjust the influence of pheromone on path selection as the number of iterations changes, improve the search effect at different stages of the algorithm, and increase the diversity of search while maintaining stability; Step 5: Introduce the inward expansion strategy of pheromone into the pheromone update rule, and perform a secondary update on the pheromone concentrations between feasible points in some preferred areas to avoid the algorithm from converging prematurely and falling into local optimality, and to ensure the quality of subsequent search paths. Step 6: limit the updated residual pheromone concentration to [τ min ,τ max ] range, to avoid excessive pheromone concentration on a certain path, which would cause the path to be suboptimal but the algorithm would converge because the pheromone concentration is much greater than that of other paths; Step 7: Based on the above steps, a pre-experimental treatment is performed on the three parameters that have a greater impact on the experimental results, namely the pheromone weight α, the heuristic information weight β, and the angle function weight γ. Dynamic adjustments are made according to the objective function, and the values ​​of the three parameters at the end of the run are used as the values ​​of the weight coefficients in the subsequent path planning experiment, avoiding the complex process of determining the parameters through multiple experiments.

2. The UAV path planning method based on the multi-strategy improved ant colony algorithm according to claim 1, characterized in that: In step 3, in order to speed up the convergence of the algorithm and reduce the corner angle, an angle function containing two factors, the steering angle and the corner angle, is introduced into the state transition probability. The definition is as follows: where θ jg is the guiding angle formed by the connection between the current node and the next node and the target point, θ sj The angles of the corners formed by the current node connecting to the previous node and the next node respectively.

3. The UAV path planning method based on the multi-strategy improved ant colony algorithm according to claim 1, characterized in that: In step 4, in order to improve the algorithm search effect, the pheromone volatility factor is adjusted to change with the number of iterations and obey the normal distribution. Its expression is as follows: Where σ is the scale parameter, a is the adjustment coefficient, N c is the current iteration number, N max is the maximum number of iterations.

4. The UAV path planning method based on the multi-strategy improved ant colony algorithm according to claim 1, characterized in that: In step 5, in order to prevent the algorithm from falling into a local optimum and to ensure the quality of subsequent search paths, a strategy of inward expansion of pheromones and limiting pheromone concentration is introduced: when the obstacle ratio in the preferred area formed by the path with a path length less than the average value and the ideal path at the initial iteration is small, then based on the traditional algorithm update rules, the pheromones between the feasible points in the corresponding preferred area are updated twice as follows: In step 6, the residual pheromone concentration is limited to the following range: where τ ij (t+1) is the pheromone concentration between node i and node j after the traditional algorithm pheromone update, Q is the total amount of pheromone released by the ant traversal once, L avg is the average length of all paths in this iteration, L i is the path whose path length is less than the average value in this iteration, τ min is the minimum pheromone concentration, τ max is the maximum pheromone concentration.

5. The UAV path planning method based on multi-strategy improved ant colony algorithm according to claim 1, characterized in that: In step 7, in order to avoid complex manual parameter adjustment, a pre-experimental process is carried out before planning the experiment. By designing the objective function, the relevant parameter values ​​are dynamically adjusted. At the end of the operation, the obtained parameter values ​​are used as the optimal parameter combination of the improved ant colony algorithm, where the objective functions corresponding to the pheromone weight α, the heuristic information weight β, and the angle function weight γ are respectively: f(k)=S(k)-S(k-1), k≥2 g(k)=L min (k)-L min (k-1),k≥2 h(k)=θ(k)-θ(k-1), k≥2 Where S(k) is the variance of all path lengths at the kth iteration, L min (k) is the length of the shortest path in the kth iteration, and θ(k) is the sum of the angles formed by the lines connecting the target point and each turning point and starting point of the shortest path in the kth iteration.

Images