Path planning method based on the fusion of improved firefly algorithm and dynamic window method

Abstract

The invention discloses a path planning method based on the fusion of improved firefly algorithm and dynamic window method. Firstly, a chaotic sequence is generated by using Skew Tent chaotic mapping to initialize the firefly population, so as to enhance the diversity of the firefly population and improve the global convergence speed of the firefly algorithm. Secondly, the fixed step length in FA is changed to an adaptive step length to limit the premature convergence of the algorithm and improve the situation that FA is prone to fall into the local optimum. Then, a differential evolution algorithm is used to enhance the search capability of the firefly algorithm through mutation, crossover and selection operations. Finally, the improved firefly algorithm is combined with the dynamic window method, so that a mobile robot can perform real-time dynamic path planning on the basis of a global optimal path and effectively avoid unknown obstacles on the basis of ensuring the global optimal path.

Description

Technical Field

[0001] The invention relates to path planning, in particular to a path planning method based on the fusion of an improved firefly algorithm and a dynamic window method. Background Art

[0002] Robots are the "crown jewel of the manufacturing industry". Their research and development, manufacturing and application are important indicators of a country's scientific and technological innovation and high-end manufacturing level. The development of robot technology and industry has become a major strategic need for my country. The "Outline of the Development Plan for New Generation Artificial Intelligence" places special emphasis on promoting the development of intelligent robots. Path planning, as an important component of the automatic navigation of intelligent mobile robots, has always been the focus and hot spot of research in the field of robotics engineering. Effective path planning is a prerequisite for robots to complete various operations and tasks.

[0003] According to whether the global environment information is known or the global environment information is unknown or the local environment information is unknown, path planning can be divided into global path planning methods and local path planning methods. Firefly Algorithm (FA) is one of the global path planning methods, while Dynamic Window Approach (DWA) is one of the local path planning methods.

[0004] FA was proposed by XS Yang in 2008 based on the movement law of fireflies. In FA, fireflies with low brightness will be attracted by fireflies with high brightness and continue to move towards fireflies with high brightness, thus guiding the entire firefly population to move to a better area. The principle of FA is intuitive, easy to operate, simple to implement, and the parameters have little effect on the algorithm. However, it also has the following disadvantages: (1) Since it mainly relies on the brightness attraction between fireflies to seek the optimal solution, in the early stage of the iteration of the firefly population, the brightest firefly in the population will attract other fireflies to gather quickly around it, which greatly reduces the diversity of the firefly population; (2) The fixed step size of FA greatly reduces the search ability of the algorithm, making it impossible to obtain a better path; (3) Some fireflies may fall into a state of aimless movement, resulting in poor FA search performance and slow convergence speed. DWA has received widespread attention due to its low computational complexity and real-time obstacle avoidance. However, the path planned by DWA is only the optimal path at the current moment, not the global optimal path, and it is easy to fall into the local minimum.

[0005] In practical applications, the environment is often unknown or partially unknown, and simply using global path planning can no longer meet the requirements of practical applications. For example: in a warehousing scenario, if the objects are fixed, the mobile robot can plan a path from the starting point to the target point based on the known information. However, if a dynamic obstacle (such as a person, object, etc.) appears on the robot's pre-planned path, the mobile robot will not be able to move according to the pre-planned global path, resulting in the mobile robot being unable to reach the target point. Summary of the invention

[0006] Purpose of the invention: The purpose of the present invention is to provide a path planning method based on the fusion of improved firefly algorithm and dynamic window method, thereby combining the global path planning method with the local path planning method, having the advantages of both planning methods, and better meeting the actual needs of path planning.

[0007] Technical solution: The path planning method based on the fusion of improved firefly algorithm and dynamic window method described in the present invention has the following principles: first, Skew Tent chaotic mapping is used to generate chaotic sequence to initialize the firefly population, so as to improve the global convergence speed of the firefly algorithm; second, an adaptive step size is introduced to balance the global and local optimal of the firefly algorithm; then, a differential evolution algorithm is used to enhance the search capability of the firefly algorithm through mutation, crossover and selection operations; finally, the improved firefly algorithm is combined with the dynamic window method, so that the mobile robot can perform real-time dynamic path planning on the basis of the global optimal path, and effectively avoid unknown obstacles on the basis of ensuring the global optimal path.

[0008] The path planning method based on the fusion of improved firefly algorithm and dynamic window method described in the present invention comprises the following steps:

[0009] (1) Use the grid method to establish a grid environment map.

[0010] The grid method refers to decomposing the space of the environmental map into independent cells to obtain grids, and the grids are interconnected but not overlapping; after rasterization, a pass factor is assigned to each grid, and the path planning problem of the mobile robot in the grid map is transformed into the problem of finding the shortest path between two grid nodes; if there are no obstacles in the grid, the robot can pass freely in the grid, and this grid is called a free grid, otherwise, it is called an obstacle grid; after the grid is marked, the mobile robot searches and displays the path according to the coordinates or serial number.

[0011] (2) Initialize the improved firefly algorithm and dynamic window method.

[0012] (2.1) Initialize the parameters of the improved firefly algorithm and the dynamic window method, including the population size n, the initial step size factor α 0, light absorption coefficient γ, attraction β0, maximum number of iterations T max , maximum linear velocity v max and the maximum angular velocity ω max .

[0013] (2.2) Randomly initialize the initial position of the firefly in the search space.

[0014] (3) Use Skew Tent chaotic mapping to initialize the firefly population.

[0015] (3.1) During the search process, the firefly individual is located at a certain position in the three-dimensional solution space. The firefly individual is mapped to the chaotic space [0,1], and then chaotic operations are performed according to the mathematical model of the Skew Tent chaotic mapping. The search is performed based on the chaotic variables generated by the operation, and the new firefly individual after the operation is obtained based on the search results. The obtained chaotic variable sequence is then restored to the original D-dimensional solution space. In this process, if a better solution is obtained after the search, the position of the original firefly can be replaced by the position of the better solution, thereby enhancing the diversity of the firefly algorithm.

[0016] (3.2) The mathematical model of Skew Tent chaos mapping is:

[0017]

[0018] Among them, k represents the number of mappings, x k Represents the kth function mapping value. When u∈(0,1) and x∈[0,1], the system is in a chaotic state.

[0019] (4) Let the current number of iterations t = 0, the improved firefly algorithm starts iterating, and an adaptive step size is introduced.

[0020] (4.1) In the firefly algorithm, assuming that the brightness of firefly i is less than that of firefly j, the iterative update formula for the position of firefly j attracted by and moving towards firefly i is:

[0021]

[0022] Where t is the current iteration number, is the spatial position of firefly i and firefly j at the tth iteration, where i≠j, is the spatial position of firefly i at the t+1th iteration, α∈[0,1] is the step size factor, ε i is a random number that follows a uniform or Gaussian distribution, β ij is the attraction between fireflies i and j:

[0023]

[0024] Among them, β0 is the maximum attraction, γ is the light absorption coefficient, r ij is the Cartesian distance from firefly i to firefly j in space.

[0025] (4.2) When the step size α is large, the algorithm will expand its search range. At this time, the algorithm's global optimization ability is stronger and the optimization speed is faster, but the search accuracy will be greatly reduced; when the step size α is small, the algorithm will narrow its search range. At this time, the algorithm's local optimization ability is stronger, the optimization speed is slower, and there is a possibility of falling into the local optimum, but the search accuracy will also be improved.

[0026] (4.3) For the firefly algorithm, a larger step size α should be used in the early stage of algorithm iteration to increase the search range of the algorithm, avoid the algorithm from falling into the local optimum, and try to find the global optimal solution; as the number of iterations increases, a smaller step size α should be used to gradually improve the search accuracy and convergence speed of the algorithm; in view of this, a decreasing function can be used to meet the requirements of the firefly algorithm for different step sizes in different iteration periods:

[0027]

[0028] Among them, α t is the step size of the algorithm at the tth iteration, α t+1 is the step size of the algorithm at the t+1th iteration, T max is the maximum number of iterations; when the number of iterations t is small, the step factor is large, and as the number of iterations increases, the step factor gradually decreases to 0.

[0029] (4.4) Add the adaptive step size of the decreasing function, and first select a larger initial step size value α at the beginning of the iteration 0 =0.9, in order to increase the search range of the firefly algorithm, avoid the algorithm from falling into the local optimum, and thus seek the global optimal solution as much as possible; as the number of iterations increases, the distance between each firefly will continue to decrease, and the step size factor in the adaptive step size strategy of the firefly algorithm will also decrease with the increase in the number of iterations, so as to gradually improve the search accuracy and convergence speed of the algorithm; as the number of iterations increases until the maximum number of iterations is reached, the step size factor gradually decreases to the minimum value; therefore, the defect of the fixed step size in the traditional FA is improved by adding the adaptive step size of the decreasing function, and the global and local optimal of the firefly algorithm are balanced in the entire iterative process.

[0030] (5) Perform differential evolution operations on the firefly individuals in the firefly population in turn, while continuously updating the position of the corresponding fireflies and the information of moving towards brighter fireflies.

[0031] (5.1) First, according to the mutation operation in the differential evolution algorithm, three firefly individuals r1, r2, and r3 are randomly selected from the firefly population. Vector difference scaling is performed between these three fireflies, and then the mutated individuals and the unmutated individuals are combined to form new individuals. The mutated firefly individuals are generated by the following formula:

[0032]

[0033] Where t is the current iteration number; For the mutant firefly individuals; and are the positions of the three firefly individuals of the tth generation; and i=1,2,…,n, n is the population size; F m ∈(0,1] is the scaling factor, which is used to scale the position vectors of three randomly selected fireflies to a certain extent; when F m When F is large, it will reduce the ability of FA to search for the local optimal solution. m When it is smaller, it will speed up the convergence of FA and cause premature convergence of FA.

[0034] (5.2) The position of the target individual in the firefly population The position of the variant individuals obtained Perform crossover operation to generate intermediate individuals The crossover operation is:

[0035]

[0036] Where, d = 1, 2, ..., D; rand (0, 1) is a random number between 0 and 1; d rand is a random number between {1,2,…,D}; represents the position of the i-th firefly individual in the t-th generation in d dimension; is the position of the new firefly individual of the tth generation in the d-dimensional space generated by the crossover operation; CR is the crossover probability, which is a random number in [0,1]. By reasonably selecting the crossover probability CR, it is beneficial to improve the convergence speed and the diversity of solutions.

[0037] (5.3) The excellent individuals after mutation and crossover operations are retained and copied to the next generation through selection operation; the selection operation can be simplified as follows: when the new individual position u i and the target individual position x i In comparison, if the new individual position u i better, the new individual is retained. Otherwise, if the target individual position x iIf the target individual is better, the target individual will be retained; after the selection operation, whether the new individual or the target individual is retained, the population state will be better than or equal to the original state, and will not make its state worse; the selection operation is:

[0038]

[0039] Among them, f(·) is the fitness function, which is used to calculate the fitness value of the current firefly individual; is the position of the target individual in the t+1th generation.

[0040] (6) Determine whether the corresponding fireflies gather at the brightest firefly or whether the number of iterations reaches the maximum value: If one of the conditions is met, the global optimal path can be obtained and go to step (7); otherwise, go to step (5).

[0041] (7) Extract multiple local sub-target points and detect environmental information to see if there are new obstacles. If new obstacles are detected, determine whether the mobile robot can avoid the new obstacles and reach the local target points based on the detected environmental information. If the new obstacles can be avoided, go to step (9); otherwise, go to step (8); if there are no new obstacles, go to step (10).

[0042] (8) The dynamic window method is used for local path planning. The mobile robot uses the dynamic window method for real-time speed sampling and uses the speed combination (v, ω) that meets the constraints for trajectory prediction. The speed combination (v, ω) corresponding to the optimal trajectory is selected according to the evaluation function for local path planning.

[0043] (9) Determine whether the local sub-target point reached is the global target point. If it is the global target point, it means that the mobile robot has reached the end point, and go to step (10); otherwise, if the local sub-target point is not the global target point, return to step (7).

[0044] (10) Output the final hybrid algorithm optimal planning path and the algorithm ends.

[0045] A computer storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned path planning method based on the fusion of an improved firefly algorithm and a dynamic window method.

[0046] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the path planning method based on the fusion of an improved firefly algorithm and a dynamic window method is implemented.

[0047] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0048] 1. Use Skew Tent chaotic mapping to initialize the firefly population, enhance the diversity of the firefly population, and improve the global convergence speed of FA;

[0049] 2. Change the fixed step size in FA to an adaptive step size to limit the premature convergence of the algorithm and improve the situation where FA is prone to fall into the local optimum;

[0050] 3. Enhance the search capability of FA through the mutation, crossover and selection operations of the differential algorithm;

[0051] 4. Combine the Improved Firefly Algorithm (IFA) with DWA, add real-time dynamic path planning on the basis of satisfying global path planning, and further improve the real-time obstacle avoidance capability of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is a flowchart of the steps of the method of the present invention;

[0053] Figure 2 It is the model diagram of the environment map;

[0054] Figure 3 This is the path planning result graph. DETAILED DESCRIPTION

[0055] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0056] like Figure 1 As shown, the path planning method based on the fusion of improved firefly algorithm and dynamic window method described in the present invention comprises the following steps:

[0057] (1) Use the grid method to establish a grid environment map.

[0058] The grid method refers to decomposing the space of the environmental map into independent cells to obtain grids, and the grids are interconnected but not overlapping; after rasterization, a pass factor is assigned to each grid, and the path planning problem of the mobile robot in the grid map is transformed into the problem of finding the shortest path between two grid nodes; if there are no obstacles in the grid, the robot can pass freely in the grid, and this grid is called a free grid, otherwise, it is called an obstacle grid; after the grid is marked, the mobile robot searches and displays the path according to the coordinates or serial number.

[0059] The grid environment map established by the present invention is as follows Figure 2As shown in the figure, its environment map model contains 20×20 grids, each of which is 1m in size; the grid is dotted in the lower left corner, and is counted from bottom to top and from left to right. In the grid, (0.5, 0.5) is the starting point of the mobile robot, and (19.5, 19.5) is the end point of the mobile robot. The black squares in the figure represent the area where the obstacles are located and are not passable, and the white squares represent the feasible area.

[0060] (2) Initialize the improved firefly algorithm and dynamic window method.

[0061] (2.1) Initialize the parameters of the improved firefly algorithm and the dynamic window method, including the population size n, the initial step size factor α 0 , light absorption coefficient γ, attraction β0, maximum number of iterations T max , maximum linear velocity v max and the maximum angular velocity ω max .

[0062] (2.2) Randomly initialize the initial position of the firefly in the search space. If d path points are to be determined, the position vector of the firefly is a d-dimensional vector.

[0063] (3) Use Skew Tent chaotic mapping to initialize the firefly population.

[0064] (3.1) During the search process, the firefly individual is located at a certain position in the D-dimensional solution space. The firefly individual is mapped to the chaotic space [0,1], and then the chaotic operation is performed according to the mathematical model of the Skew Tent chaotic mapping. The search is performed according to the chaotic variables generated by the operation, and then the new firefly individual after the operation is obtained according to the search results. Then the obtained chaotic variable sequence is restored to the original D-dimensional solution space. In this process, if a better solution is obtained after the search, the position of the better solution can replace the original position of the firefly, thereby enhancing the diversity of the firefly algorithm.

[0065] (3.2) The mathematical model of Skew Tent chaos mapping is:

[0066]

[0067] Among them, k represents the number of mappings, x k Represents the kth function mapping value. When u∈(0,1) and x∈[0,1], the system is in a chaotic state.

[0068] (4) Let the current number of iterations t = 0, the improved firefly algorithm starts iterating, and an adaptive step size is introduced.

[0069] (4.1) In the firefly algorithm, assuming that the brightness of firefly i is less than that of firefly j, the iterative update formula for the position of firefly j attracted by and moving towards firefly i is:

[0070]

[0071] Where t is the current iteration number, is the spatial position of firefly i and firefly j at the tth iteration, where i≠j, is the spatial position of firefly i at the t+1th iteration, α∈[0,1] is the step size factor, ε i is a random number that follows a uniform or Gaussian distribution, β ij is the attraction between fireflies i and j:

[0072]

[0073] Among them, β0 is the maximum attraction, γ is the light absorption coefficient, r ij is the Cartesian distance from firefly i to firefly j in space.

[0074] (4.2) When the step size α is large, the algorithm will expand its search range. At this time, the algorithm's global optimization ability is stronger and the optimization speed is faster, but the search accuracy will be greatly reduced; when the step size α is small, the algorithm will narrow its search range. At this time, the algorithm's local optimization ability is stronger, the optimization speed is slower, and there is a possibility of falling into the local optimum, but the search accuracy will also be improved.

[0075] (4.3) For the firefly algorithm, a larger step size α should be used in the early stage of the algorithm iteration to increase the search range of the algorithm, avoid the algorithm from falling into the local optimum, and try to find the global optimal solution; as the number of iterations increases, a smaller step size α should be used to gradually improve the search accuracy and convergence speed of the algorithm. In view of this, a decreasing function can be used to meet the requirements of the firefly algorithm for different step sizes in different iteration periods:

[0076]

[0077] Among them, α t is the step size of the algorithm at the tth iteration, α t+1 is the step size of the algorithm at the t+1th iteration, T max is the maximum number of iterations. When the number of iterations t is small, the step factor is large, and as the number of iterations increases, the step factor gradually decreases to 0.

[0078] (4.4) Add the adaptive step size of the decreasing function, and first select a larger initial step size value α at the beginning of the iteration 0=0.9, in order to increase the search range of the firefly algorithm, avoid the algorithm from falling into the local optimum, and thus seek the global optimal solution as much as possible; as the number of iterations increases, the distance between each firefly will continue to decrease, and the step factor in the adaptive step strategy of the firefly algorithm will also decrease with the increase in the number of iterations, so as to gradually improve the search accuracy and convergence speed of the algorithm. As the number of iterations increases until the maximum number of iterations is reached, the step factor gradually decreases to the minimum value. Therefore, the defect of the fixed step size in the traditional FA is improved by adding the adaptive step size of the decreasing function, and the global and local optimal of the firefly algorithm are balanced in the entire iterative process.

[0079] (5) Perform differential evolution operations on the firefly individuals in the firefly population in turn, while continuously updating the position of the corresponding fireflies and the information of moving towards brighter fireflies.

[0080] (5.1) First, according to the mutation operation in the differential evolution algorithm, three firefly individuals r1, r2, and r3 are randomly selected from the firefly population. Vector difference scaling is performed between these three fireflies, and then the mutated individuals and the unmutated individuals are combined to form new individuals. The mutated firefly individuals are generated by the following formula:

[0081]

[0082] Where t is the current iteration number; For the mutant firefly individuals; and are the positions of the three firefly individuals of the tth generation; and i=1,2,…,n, n is the population size; F m ∈(0,1] is the scaling factor, which is used to scale the position vectors of three randomly selected fireflies. m When F is large, it will reduce the ability of FA to search for the local optimal solution. m When it is smaller, it will speed up the convergence of FA and cause premature convergence of FA.

[0083] (5.2) Secondly, the position of the target individual in the firefly population The position of the variant individuals obtained Perform crossover operation to generate intermediate individuals The crossover operation is:

[0084]

[0085] Where, d = 1, 2, ..., D; rand (0, 1) is a random number between 0 and 1; d rand is a random number between {1,2,…,D}; represents the position of the i-th firefly individual in the t-th generation in d dimension; is the position of the new firefly individual of the tth generation generated by the crossover operation; CR is the crossover probability, which is a random number in [0,1]. By reasonably selecting the crossover probability CR, it is helpful to improve the convergence speed and the diversity of solutions.

[0086] (5.3) Finally, the excellent individuals after mutation and crossover operations are retained and copied to the next generation through selection. The selection operation can be simplified as follows: when the new individual position u i and the target individual position x i In comparison, if the new individual position u i better, the new individual is retained. Otherwise, if the target individual position x i If the target individual is better, the target individual will be retained. After the selection operation, whether the new individual or the target individual is retained, the population state will be better than or equal to the original state, and will not make it worse. The selection operation is:

[0087]

[0088] Among them, f(·) is the fitness function, which is used to calculate the fitness value of the current firefly individual; is the position of the target individual in the t+1th generation.

[0089] (6) Determine whether the corresponding fireflies gather at the brightest firefly or whether the number of iterations reaches the maximum value: If one of the conditions is met, the global optimal path can be obtained and go to step (7); otherwise, go to step (5).

[0090] (7) Extract multiple local sub-target points and detect environmental information to see if there are new obstacles

[0091] If a new obstacle is detected, the mobile robot is judged whether it can avoid the new obstacle and reach the local target point based on the detected environmental information. If it can avoid the new obstacle, the process goes to step (9); otherwise, the process goes to step (8). If there is no new obstacle, the process goes to step (10).

[0092] (8) Using dynamic window method for local path planning

[0093] The mobile robot uses the dynamic window method for real-time velocity sampling, adopts the speed combination (v, ω) that meets the constraints for trajectory prediction, and selects the speed combination (v, ω) corresponding to the optimal trajectory according to the evaluation function for local path planning.

[0094] (9) Determine whether the local sub-target point reached is the global target point

[0095] If it is the global target point, it means that the mobile robot has reached the end point, and go to step (10); otherwise, if the local sub-target point is not the global target point, return to step (7).

[0096] (10) Output the final hybrid algorithm optimal planning path and the algorithm ends.

[0097] The path planned by the mobile robot in this embodiment is as follows Figure 3 As shown in the figure, the two circular obstacles are new obstacles, the solid line path is the global path planning path of the improved firefly algorithm in a static environment, and the dotted line path is the path of the improved firefly algorithm combined with the dynamic window method proposed in the present invention, where the circle in the lower left corner is the starting point of the fusion algorithm, and the circle in the upper right corner is the local sub-target point of the fusion algorithm. Figure 3 It means that the mobile robot has realized collision-free dynamic path planning from the starting point to the end point on the basis of ensuring the global optimal path.

Claims

1. A path planning method based on the fusion of improved firefly algorithm and dynamic window method, characterized in that: The following steps are involved: (1) Use the grid method to establish a grid environment map; (2) Initialize the improved firefly algorithm and dynamic window method; (3) Use Skew Tent chaotic mapping to initialize the firefly population; specifically: (3.1) During the search process, the firefly individual is located at a certain position in the three-dimensional solution space. The firefly individual is mapped to the chaotic space [0, 1], and then a chaotic operation is performed according to the mathematical model of the Skew Tent chaotic mapping. The search is performed according to the chaotic variables generated by the operation, and then the new firefly individual after the operation is obtained according to the search results. Then, the obtained chaotic variable sequence is restored to the original D-dimensional solution space. In this process, if a better solution than the original one is obtained after the search, the position of the better solution can be replaced by the original position of the firefly, thereby enhancing the diversity of the firefly algorithm. (3.2) The mathematical model of Skew Tent chaos mapping is: Among them, k represents the number of mappings, x k represents the kth function mapping value. When u∈(0, 1) and x∈[0, 1], the system is in a chaotic state. (4) Let the current iteration number t=0, the improved firefly algorithm starts to iterate, and the adaptive step size is introduced. Specifically: (4.1) In the firefly algorithm, assuming that the brightness of firefly i is less than that of firefly j, the iterative update formula for the position of firefly j attracted by and moving towards firefly i is: Where t is the current iteration number, is the spatial position of firefly i and firefly j at the tth iteration, where i≠j, is the spatial position of firefly i at the t+1th iteration, α∈[0,1] is the step size factor, ε i is a random number that follows a uniform or Gaussian distribution, β ij is the attraction between fireflies i and j: Among them, β0 is the maximum attraction, γ is the light absorption coefficient, r ij is the Cartesian distance from firefly i to firefly j in space; (4.2) When the step size α is large, the algorithm will expand its search range. At this time, the algorithm has stronger global optimization ability and faster optimization speed, but its search accuracy will be greatly reduced; when the step size α is small, the algorithm will narrow its search range. At this time, the algorithm has stronger local optimization ability, slower optimization speed, and may fall into local optimality, but its search accuracy will also be improved; (4.3) For the firefly algorithm, a larger step size α should be used in the early stage of algorithm iteration to increase the search range of the algorithm, avoid the algorithm from falling into the local optimum, and try to find the global optimal solution; as the number of iterations increases, a smaller step size α should be used to gradually improve the search accuracy and convergence speed of the algorithm; in view of this, a decreasing function is used to meet the requirements of the firefly algorithm for different step sizes in different iteration periods: Among them, α t is the step size of the algorithm at the tth iteration, α t+1 is the step size of the algorithm at the t+1th iteration, T max is the maximum number of iterations; when the number of iterations t is small, the step factor is large, and as the number of iterations increases, the step factor gradually decreases to 0; (4.4) Add the adaptive step size of the decreasing function, and first select a larger initial step size value α at the beginning of the iteration 0 =0.9, in order to increase the search range of the firefly algorithm, avoid the algorithm from falling into the local optimum, and thus try to find the global optimal solution; as the number of iterations increases, the distance between each firefly will continue to decrease, and the step factor in the adaptive step strategy of the firefly algorithm will also decrease with the increase in the number of iterations, so as to gradually improve the search accuracy and convergence speed of the algorithm; as the number of iterations increases until the maximum number of iterations is reached, the step factor gradually decreases to the minimum value; therefore, the defect of the fixed step size in the traditional FA is improved by adding the adaptive step size of the decreasing function, and the global and local optimal of the firefly algorithm are balanced in the entire iterative process; (5) Perform differential evolution operations on the firefly individuals in the firefly population in turn, while continuously updating the position of the corresponding fireflies and the information of moving towards brighter fireflies; (6) Determine whether the corresponding fireflies gather at the brightest firefly or whether the number of iterations reaches the maximum value: If one of the conditions is met, the global optimal path can be obtained and go to step (7); otherwise, go to step (5); (7) Extract multiple local sub-target points, detect environmental information to see if there are new obstacles, if new obstacles are detected, determine whether the mobile robot can avoid the new obstacles and reach the local target point based on the detected environmental information, if the new obstacles can be avoided, go to step (9); otherwise, go to step (8); if there are no new obstacles, go to step (10); (8) The dynamic window method is used for local path planning. The mobile robot uses the dynamic window method to perform real-time speed sampling, uses the speed combination (v, ω) that meets the constraints for trajectory prediction, and selects the speed combination (v, ω) corresponding to the optimal trajectory according to the evaluation function for local path planning; (9) Determine whether the local sub-target point reached is the global target point. If it is the global target point, it means that the mobile robot has reached the end point, and go to step (10); otherwise, if the local sub-target point is not the global target point, return to step (7); (10) Output the final hybrid algorithm optimal planning path and the algorithm ends.

2. The path planning method based on the fusion of improved firefly algorithm and dynamic window method according to claim 1 is characterized in that: The step (1) is specifically: The grid method refers to decomposing the space of the environmental map into independent cells to obtain grids, and the grids are interconnected but not overlapping; after rasterization, a pass factor is assigned to each grid, and the path planning problem of the mobile robot in the grid map is transformed into the problem of finding the shortest path between two grid nodes; if there are no obstacles in the grid, the robot can pass freely in the grid, and this grid is called a free grid, otherwise, it is called an obstacle grid; after the grid is marked, the mobile robot searches and displays the path according to the coordinates or serial number.

3. The path planning method based on the fusion of improved firefly algorithm and dynamic window method according to claim 1 is characterized in that: The step (2) is specifically: (2.1) Initialize the parameters of the improved firefly algorithm and the dynamic window method, including the population size n, the initial step size factor α 0 , light absorption coefficient γ, attraction β0, maximum number of iterations T max , maximum linear velocity v max and the maximum angular velocity ω max ; (2.2) Randomly initialize the initial position of the firefly in the search space.

4. The path planning method based on the fusion of improved firefly algorithm and dynamic window method according to claim 1 is characterized in that: The step (5) is specifically as follows: (5.1) First, according to the mutation operation in the differential evolution algorithm, three firefly individuals r1, r2, and r3 are randomly selected from the firefly population. Vector difference scaling is performed between these three fireflies, and then the mutated individuals and the unmutated individuals are combined to form new individuals. The mutated firefly individuals are generated by the following formula: Where t is the current iteration number; For the mutant firefly individuals; and are the positions of the three firefly individuals of the tth generation; and n is the population size; F m ∈(0, 1] is the scaling factor, which is used to scale the position vectors of three randomly selected fireflies to a certain extent; when F m When F is large, it will reduce the ability of FA to search for the local optimal solution. m When it is smaller, it will speed up the convergence of FA, causing premature convergence of FA; (5.2) The position of the target individual in the firefly population The position of the variant individuals obtained Perform crossover operation to generate intermediate individuals The crossover operation is: Where, d = 1, 2, ..., D; rand (0, 1) is a random number between 0 and 1; d rand is a random number between {1, 2, …, D}; represents the position of the i-th firefly individual in the t-th generation in d dimension; is the position of the new firefly individual of the tth generation in the d-dimensional crossover operation; CR is the crossover probability, which is a random number in [0, 1]. By reasonably selecting the crossover probability CR, it is beneficial to improve the convergence speed and the diversity of solutions; (5.3) The excellent individuals after mutation and crossover operations are retained and copied to the next generation through selection operation; the selection operation can be simplified as follows: when the new individual position u i and the target individual position x i In comparison, if the new individual position u i better, the new individual is retained. Otherwise, if the target individual position x i If the target individual is better, the target individual will be retained; after the selection operation, whether the new individual or the target individual is retained, the population state will be better than or equal to the original state, and will not make its state worse; the selection operation is: Among them, f(·) is the fitness function, which is used to calculate the fitness value of the current firefly individual; is the position of the target individual in the t+1th generation.

5. A computer storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, a path planning method based on the fusion of an improved firefly algorithm and a dynamic window method as described in any one of claims 1 to 4 is implemented.

6. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, a path planning method based on the fusion of an improved firefly algorithm and a dynamic window method is implemented as described in any one of claims 1 to 4.

Images