Path planning method fusing D*Lite, simulated annealing and genetic algorithm

Abstract

The invention relates to a path planning method fusing D * Lite, simulated annealing and a genetic algorithm. The path planning method comprises the following steps: S1, acquiring map environment information by using an RGB-D camera; s2, modeling is carried out on map environment information; s3, initializing parameters of the D * Lite and a genetic algorithm; s4, performing genetic algorithm population initialization fused with a D * Lite algorithm; s5, the fitness value of the initialized population is calculated, and path redundant point optimization is carried out; s6, applying the generated population to a selection operator of a fusion simulated annealing algorithm, and introducing an elitism strategy; s7, applying the population subjected to operator selection to adaptive crossover and mutation operators; s8, the fitness values of all individuals of the population are calculated again, and path redundant points are optimized again; and S9, judging whether the path meets an optimal condition or not, and outputting an optimal solution after iteration. The method not only has strong global search capability in a complex environment, but also improves local search capability, has good universality and robustness, and meets actual requirements.

Description

Technical Field

[0001] The present invention belongs to the field of autonomous navigation of mobile robots, and specifically relates to a path planning method integrating D*Lite, simulated annealing and genetic algorithm. Background Art

[0002] The statements in this part only provide background technical information related to the present disclosure, and these statements may constitute prior art. In the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art.

[0003] With the continuous development of science and technology, mobile robots have become increasingly important in people's lives and work, such as floor cleaning robots, drones and water surface navigators. The prerequisite for a mobile robot to perform various tasks is to have the ability of autonomous navigation, that is, to use its own sensors to sense information such as the working environment, obstacles and its own state position, establish a corresponding environmental model, and plan the best path to the target point. The navigation technology of mobile robots includes environmental perception technology, autonomous positioning and map building, path planning algorithms, power and motion control, etc.

[0004] Path planning is one of the key technologies of mobile robots and is an optimization problem with complexity, non-linearity and multi-objective constraints. Path planning algorithms have always been a hot topic of concern for researchers. After long-term research and development, new and different types of path planning algorithms have been continuously proposed and improved, such as ant colony algorithm, artificial potential field method, A* algorithm, dynamic window method, genetic algorithm (GA), grey wolf algorithm, golden jackal optimization algorithm, neural network algorithm, particle swarm optimization algorithm, rapidly-exploring random tree star (RRT*), tidal algorithm, whale algorithm. However, most of the algorithm optimization objectives do not consider enough factors, and there are situations such as being easily trapped in local optimal solutions and having poor global optimal solution effects.

[0005] The genetic algorithm is a global optimization method based on natural inheritance and natural selection. Due to its advantages such as strong adaptability to different spatial environments and high robustness, it can be widely used to solve path planning problems. However, the local search ability of this algorithm is poor, and the search efficiency is low in the later stage of evolution. However, in a complex environment, when a mobile robot faces road risks, different road factors need to be considered more.

[0006] To address the deficiencies of genetic algorithms, there is a method of integrating a function model that correlates task danger levels with road factors into path planning, called the Hybrid Adaptive Genetic Algorithm - HAGA. HAGA combines the optimized quadratic optimization 2-opt* operator with an enhanced adaptive genetic algorithm to perform efficient path planning under different tasks and complex road conditions. First, the initial path is pre-optimized by introducing a new operator based on domain knowledge to remove duplicate paths in the path to avoid redundant paths; then it is divided into the TOP layer and the ordinary layer, and the adaptive 2-opt* operator that satisfies the hyperbolic tangent function TANH is used to optimize the TOP layer (elite layer), and the Sigmoid and TANH functions are respectively used to optimize the crossover and mutation of the ordinary layer, establishing a robot path planning algorithm applicable to complex environments.

[0007] A similar patent application No. 202410778247.X with the patent name "Robot Path Planning Method Based on Ant Colony and Adaptive Genetic Algorithm" uses the above method to solve the problems of low search efficiency, uneven paths, many redundant nodes, and obstacle avoidance in complex environments of traditional genetic algorithms. To a certain extent, it improves the search efficiency, and the path length and smoothness are also optimized. However, it does not incorporate too many judgment factors regarding road safety, and the calculation process is relatively complex, reducing the algorithm running time.

[0008] How to better ensure population diversity, so as to comprehensively consider the optimal path in multiple aspects and reduce the operation time of the algorithm, thus obtaining the current optimal path in a shorter time, is an urgent problem to be solved in this field. Summary of the Invention

[0009] In view of the above problems, the purpose of the present invention is to solve a part of the problems in the prior art, or at least alleviate these problems.

[0010] A path planning method that integrates D*Lite, simulated annealing, and genetic algorithms includes the following steps:

[0011] S1: Obtain map environment information;

[0012] S2: Model the map environment information;

[0013] S3: Initialize the parameters of D*Lite and genetic algorithms;

[0014] S4: Initialize the population of the genetic algorithm integrated with the D*Lite algorithm to obtain the initial population;

[0015] S5: Calculate the fitness values of the initial population and optimize the redundant path points to generate the population;

[0016] S6: Apply the generated population to the selection operator of the fusion simulated annealing algorithm and introduce the elite retention strategy to obtain the population after the selection operator.

[0017] S7: Apply the population after the selection operator to the adaptive crossover and mutation operators to obtain the population after the crossover and mutation operators.

[0018] S8: Calculate the fitness values of all individuals in the population again for the population after the crossover and mutation operators, and optimize the path redundant points again to obtain the path after re-optimization.

[0019] S9: Determine whether the path after re-optimization meets the optimal conditions; if it meets, output the optimal solution after iteration; if it does not meet, enter the loop of step S6.

[0020] In step S2, the map environment is modeled, including the following steps:

[0021] The grid numbers increase in ascending order from bottom to top and from left to right.

[0022] Establish a formula for converting coordinates to numbers and a formula for converting numbers to coordinates.

[0023] Set the code value of free grids and the code value of obstacle grids.

[0024] Formulate the obstacle inflation rule for map modeling: Obstacles less than one grid are calculated as a whole grid.

[0025] In step S3, initialize the parameters of D*Lite and the genetic algorithm, specifically including:

[0026] Initialization of D*Lite algorithm parameters: The grid map data built in step S2, the path start point, the path end point, the movement cost, and the heuristic function value.

[0027] Initialization of genetic algorithm parameters: The grid map data built in step S2, the path start point, the path end point, the population size, the maximum number of iterations, the fitness function parameter value, and the crossover and mutation probability values.

[0028] In step S4, initialize the population of the genetic algorithm integrated with the D*Lite algorithm, including the following steps:

[0029] Traverse all free grids on the grid map, and randomly select a free grid as a path necessary point from all free grids in each row.

[0030] Connect all path necessary points, the start point, and the end point by the path generation function to form a path.

[0031] Integrate the excellent paths generated by the D*Lite algorithm as elite individuals into the initial population of the genetic algorithm and add them to subsequent crossover and mutation.

[0032] In step S5, the fitness value of the initialized population is calculated and the path redundant points are optimized, which includes the following steps:

[0033] Calculate the path length using the classic Euclidean distance:

[0034]

[0035] (x i ,y i ) is the current grid position, (x i+1 ,y i+1 ) is the position of the next grid, and n is the total number of individuals in the population;

[0036] Calculate the path smoothness, which is defined as follows:

[0037] C i =(x i+2 -x i ) 2 +(y i+2 -y i ) 2 ,i∈[1,n-2]

[0038]

[0039] (x i ,y i ) is the current grid position, (x i+2 ,y i+2 ) is the position of the second grid, according to C i The size of S is used to judge the smoothness of the path. i is the turning penalty value of the current position;

[0040] At the same time, an obstacle distance penalty function is added, which is defined as follows:

[0041]

[0042] (x i ,y i ) is the current grid position, (x o ,y o ) is the grid position of the nearest obstacle, O i is the Euclidean distance between the current position and the nearest obstacle, D i is the obstacle distance penalty value at the current position;

[0043] Path length, path smoothness, and obstacle distance penalty function together constitute the evaluation indicators of the fitness function;

[0044]

[0045] F(k) is the fitness function value, and ε1 and ε2 are user-defined variables;

[0046] The RDP algorithm is introduced to eliminate the redundant parts of the path after calculating the path smoothness.

[0047] In step S6, the generated population is applied to the selection operator of the fusion simulated annealing algorithm, and an elite retention strategy is introduced, including the following steps:

[0048] The roulette wheel method used in the selection operator of the optimized genetic algorithm is as follows:

[0049]

[0050] ε is a user-defined parameter for smoothing the fitness value, f(x) is the fitness value of the corresponding individual, and x j , x k , x i are all individuals in the population;

[0051] An elite individual retention strategy is introduced. Before selecting an individual by the roulette wheel method each time, the individuals with fitness values before the user-defined percentage are stored as elite individuals;

[0052] The fusion simulated annealing algorithm retains individuals with relatively poor fitness values with a certain probability; the probability formula is as follows:

[0053]

[0054] Among them, F(W1) and F(W2) are the fitness values of the corresponding individuals, T0 is the initial temperature, q is the temperature reduction coefficient, and GEN is the number of iterations.

[0055] In step S7, the population after the selection operator is applied to the adaptive crossover and mutation operators, including the following steps:

[0056] Set the range and initial value of the crossover and mutation probabilities;

[0057] Introduce the sigmoid function to flexibly adjust the probability value, and the formula is as follows:

[0058]

[0059] Among them, p is the final crossover probability value, p max and p min are the set maximum and minimum crossover probability values respectively, f max and f min are the maximum and minimum fitness values in the population respectively, x o is the set fitness difference center point, and k is the parameter controlling the steepness of the curve.

[0060] In step S8, the fitness values of all individuals in the population after the cross mutation operator are recalculated, and the path redundant points are optimized again. The specific implementation is the same as that in step S5.

[0061] In step S9, it is judged whether the path after the re - optimization meets the optimal conditions; if it meets, the optimal solution after iteration is output, including the following steps:

[0062] First, it is judged whether the fitness value meets the expected requirements;

[0063] Then, it is judged whether the population iteration times reach the limited maximum times;

[0064] Finally, the fitness values of all individuals are compared, and the individual with the highest fitness value is output as the current optimal solution.

[0065] An electronic device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the path planning method integrating D*Lite, simulated annealing, and genetic algorithm are implemented.

[0066] The present invention has the following beneficial effects:

[0067] 1. The present invention integrates the D*Lite algorithm and the elite retention strategy, and adaptively adjusts the crossover and mutation probabilities of the genetic algorithm, which can reduce the running time of the algorithm to a certain extent;

[0068] 2. The present invention integrates the simulated annealing algorithm, which can enhance the diversity of the initial population of the genetic algorithm, and uses the roulette wheel method with enhanced diversity in the selection operator of the genetic algorithm to further enhance the diversity of the initial population;

[0069] 3. The present invention introduces a new evaluation mechanism into the fitness function, which not only considers the path length, but also evaluates the path smoothness and the route safety factor, and comprehensively considers the optimal path in multiple aspects;

[0070] 4. The present invention optimizes the path while calculating the fitness function value each time, removes redundant path points, greatly reduces the actual iteration times of the algorithm, and can quickly obtain the current optimal path.

[0071] The above - mentioned technical solutions work together, so that the present invention can not only meet the population diversity to achieve the purpose of comprehensively considering the optimal path in multiple aspects, but also reduce the operation time of the algorithm, has good versatility and robustness, and better meets the actual needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1The figure shows a flowchart of a simultaneous localization and mapping method based on a lightweight PSPNet network provided by the present invention;

[0073] Figure 2 These are experimental effect diagrams; among them, (a) is the path effect of the genetic algorithm, and (b) is the path effect of the present invention. Detailed implementation manners

[0074] The following further describes the present invention with reference to the accompanying drawings. The embodiments of the present invention are only used to illustrate the present invention and not to limit it. Without departing from the technical idea of the present invention, various substitutions and changes made according to common general knowledge and conventional means in the art shall all be included within the scope of the present invention.

[0075] To solve the above technical problems, the present invention designs a path planning method that combines D*Lite, simulated annealing, and genetic algorithm. As Figure 1 shown, it includes the following steps:

[0076] S1: Use an RGB-D camera (a type of camera that combines color information and depth information) to obtain map environment information;

[0077] S2: Model the map environment information;

[0078] S3: Initialize the parameters of D*Lite and genetic algorithm; D*lite is the name of a path planning algorithm.

[0079] S4: Initialize the population of the genetic algorithm combined with the D*Lite algorithm to obtain an initial population;

[0080] S5: Calculate the fitness values of the initial population and optimize the path redundant points to generate a population;

[0081] S6: Apply the generated population to the selection operator of the simulated annealing algorithm combined, and introduce an elite retention strategy to obtain the population after the selection operator;

[0082] S7: Apply the population after the selection operator to the adaptive crossover and mutation operators to obtain the population after the crossover and mutation operators;

[0083] S8: Calculate the fitness values of all individuals in the population again for the population after the crossover and mutation operators, and optimize the path redundant points again to obtain the path after the second optimization;

[0084] S9: Determine whether the path after the second optimization meets the optimal conditions; if it meets, output the optimal solution after iteration; if it does not meet, enter the loop of step S6.

[0085] In step S2, modeling the map environment specifically includes:

[0086] The grid numbers increase in the order from bottom to top and from left to right;

[0087] Establish a formula for converting coordinates to numbers and a formula for converting numbers to coordinates;

[0088] Set the code value of free grids and the code value of obstacle grids;

[0089] Formulate the obstacle inflation rule for map modeling: obstacles less than one grid are calculated as a whole grid.

[0090] In step S3, initialize the parameters of D*Lite and genetic algorithm, specifically including:

[0091] Initialization of D*Lite algorithm parameters: the grid map data built in step S2, the path start point, the path end point, the movement cost, and the heuristic function value;

[0092] Initialization of genetic algorithm parameters: the grid map data built in step S2, the path start point, the path end point, the population size, the maximum number of iterations, the parameter values of the fitness function, and the crossover and mutation probability values.

[0093] In step S4, initialize the population of the genetic algorithm integrating the D*Lite algorithm, specifically including:

[0094] Initialization of the genetic algorithm population: First, traverse all free grids on the grid map, randomly select a free grid from all free grids in each row as a path necessary point; then use the path generation function to connect all path necessary points, the start point, and the end point to form a path; then integrate the excellent paths generated by the D*Lite algorithm as elite individuals into the initial population of the genetic algorithm for subsequent crossover and mutation.

[0095] In step S5, calculate the fitness values of the initialized population and optimize the redundant path points, specifically including:

[0096] First, calculate the path length using the classic Euclidean distance:

[0097]

[0098] (x i , y i ) is the current grid position, (x i+1 , y i+1 ) is the position of the next grid, and n is the total number of individuals in the population;

[0099] Calculate the path smoothness, defined as follows:

[0100] C i =(x i+2 -xi ) 2 +(y i+2 -y i ) 2 , i ∈ [1, n - 2]

[0101]

[0102] (x i , y i ) is the current grid position, (x i+2 , y i+2 ) is the position of the second next grid. The smoothness of the path is judged according to the size of C i . S i is the turning penalty value at the current position;

[0103] At the same time, an obstacle distance penalty function is added, which is defined as follows:

[0104]

[0105] (x i , y i ) is the current grid position, (x o , y o ) is the grid position of the nearest obstacle. O i is the Euclidean distance between the current position and the nearest obstacle, and D i is the obstacle distance penalty value at the current position.

[0106] The path length, path smoothness, and obstacle distance penalty function together constitute the evaluation index of the fitness function;

[0107]

[0108] F(k) is the value of the fitness function, and ε1 and ε2 are user-defined variables.

[0109] At the same time, the RDP algorithm (Ramer-Douglas-Peucker) is introduced to remove the redundant part of the path after calculating the path smoothness to optimize the path.

[0110] In step S6, the generated population is applied to the selection operator of the fusion simulated annealing algorithm, and an elite retention strategy is introduced, including the following steps:

[0111] First, optimize the roulette wheel method used in the selection operator of the genetic algorithm to make it capable of selecting more diverse individuals. The formula is as follows:

[0112]

[0113] ε is a user-defined parameter used to smooth the fitness value, f(x) is the fitness value of the corresponding individual, and x j , x k , x i are all individuals in the population;

[0114] Then, an elite individual retention strategy is introduced. Before selecting individuals by roulette wheel method each time, the individuals with fitness values before the user-defined percentage are stored as elite individuals to prevent excellent individuals from being eliminated;

[0115] Finally, the simulated annealing algorithm is integrated to retain individuals with relatively poor fitness values with a certain probability to ensure the diversity of the population. The probability formula is as follows:

[0116]

[0117] where F(W1) and F(W2) are the fitness values of the corresponding individuals, T0 is the initial temperature, q is the cooling coefficient, and GEN is the number of iterations.

[0118] In step S7, the population after the selection operator is applied to the adaptive crossover and mutation operators, which specifically include:

[0119] First, set the range and initial value of the crossover and mutation probabilities;

[0120] Then, introduce the sigmoid function to flexibly adjust the probability value. The formula is as follows:

[0121]

[0122] where p is the final crossover probability value, p max and p min are the set maximum and minimum crossover probability values respectively, f max and f min are the maximum and minimum fitness values in the population respectively, x o is the set fitness difference center point, and k is the parameter controlling the steepness of the curve. By adjusting x o and k, the change curve of the crossover probability can be flexibly adjusted.

[0123] In step S8, the fitness values of all individuals in the population after the crossover and mutation operators are recalculated, and the path redundant points are optimized again. The specific implementation is the same as that in step S5.

[0124] Step S9 is to judge whether the path meets the optimal conditions and output the optimal solution after iteration, which specifically includes:

[0125] First, judge whether the fitness value meets the expected requirements;

[0126] Then, judge whether the number of population iterations reaches the limited maximum number;

[0127] Finally, compare the fitness values of all individuals and output the individual with the highest fitness value as the current optimal solution.

[0128] Appendix Figure 1 Whether the termination condition is satisfied. For specific conditions, see step S9. When the algorithm is initialized, the initial population size M and the maximum number of iterations are set.

[0129] GEN: Represents that the population has completed one iteration.

[0130] i: An individual in the population has undergone operations such as selection, crossover, and mutation once.

[0131] GEN is the outer loop, and i is the inner loop. The inner loop is contained in the outer loop.

[0132] Assume that the maximum number of iterations is set to 50 (meeting the termination condition), and the population size is 20 (M). The count represented by the outer loop of 50 is GEN, and the count represented by the inner loop of 20 is i. The inner loop is within the outer loop.

[0133] Experimental data:

[0134] Such as Figure 2 As shown, each algorithm conducts no less than 20 repeated experiments to obtain the distance values of each optimal path, and 20 consecutive distance values are taken for data analysis:

[0135]

[0136]

[0137] It can be clearly seen from the above that the present invention has a significant improvement compared with the genetic algorithm.

[0138] The present invention adds a new judgment factor to the fitness function to introduce a judgment factor regarding road safety. At the same time, a simple and efficient method is used to separate elite individuals from ordinary individuals, and it can better ensure population diversity, and can obtain the current optimal solution in a shorter time. Thus, this method not only has strong global search ability in complex environments, but also improves the local search ability, and has good generality and robustness, meeting the actual needs.

[0139] Only one electronic device is required, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the path planning method that combines D*Lite, simulated annealing, and genetic algorithm are implemented.

[0140] The systems, devices, modules or units described in the above embodiments can be specifically implemented by computer chips or entities, or by products with certain functions.

[0141] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media, such as modulated data signals and carrier waves.

[0142] It should also be noted that the term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0143] The above embodiments should be understood as only illustrative of the present invention and not limiting the scope of protection of the present invention. After reading the content described in the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent changes and modifications also fall within the scope defined by the claims of the present invention.

Claims

1. A path planning method integrating D*Lite, simulated annealing and genetic algorithm, characterized in that It includes the following steps: S1: Obtain map environment information; S2: Model the map environment information; S3: Initialize the parameters of D*Lite and genetic algorithm; S4: Initialize the population of the genetic algorithm integrated with D*Lite algorithm to obtain the initialized population; S5: Calculate the fitness values of the initialized population and optimize the redundant path points to generate a population; S6: Apply the generated population to the selection operator of the integrated simulated annealing algorithm and introduce the elite retention strategy to obtain the population after the selection operator; S7: Apply the population after the selection operator to the adaptive crossover and mutation operators to obtain the population after the crossover and mutation operators; S8: Calculate the fitness values of all individuals in the population again for the population after the crossover and mutation operators and optimize the redundant path points again to obtain the path after re-optimization; S9: Judge whether the path after re-optimization meets the optimal conditions; if it meets, output the optimal solution after iteration; if it does not meet, enter the loop of step S6.

2. The path planning method integrating D*Lite, simulated annealing and genetic algorithm according to claim 1, characterized in that In step S2, modeling the map environment includes the following steps: The grid numbers increase in the order from bottom to top and from left to right; Establish a formula for converting coordinates to numbers and a formula for converting numbers to coordinates; Set the code values for free grids and the code values for obstacle grids; Formulate the obstacle expansion rule for map modeling: Obstacles less than one grid are calculated as a whole grid.

3. The path planning method integrating D*Lite, simulated annealing and genetic algorithm according to claim 1, characterized in that, In step S3, initializing the parameters of D*Lite and genetic algorithm specifically includes: Initialization of D*Lite algorithm parameters: The grid map data, path start point, path end point, movement cost, and heuristic function value built in step S2; Initialization of genetic algorithm parameters: The grid map data, path start point, path end point, population size, maximum number of iterations, fitness function parameter values, and crossover and mutation probability values built in step S2.

4. The path planning method integrating D*Lite, simulated annealing and genetic algorithm according to claim 1, characterized in that, In step S4, initializing the population of the genetic algorithm integrated with D*Lite algorithm includes the following steps: Traverse all free grids on the grid map and randomly select a free grid as a necessary path point in all free grids of each row; Connect all necessary path points, start point, and end point by the path generation function to form a path; Integrate the excellent paths generated by the D*Lite algorithm as elite individuals into the initial population of the genetic algorithm for subsequent crossover and mutation.

5. The path planning method integrating D*Lite, simulated annealing and genetic algorithm according to claim 1, wherein In step S5, calculating the fitness values of the initialized population and optimizing the redundant path points includes the following steps: Calculate the path length using the classical Euclidean distance: (x i , y i ) is the current grid position, (x i+1 , y i+1 ) is the position of the next grid, and n is the total number of individuals in the population; Calculate the path smoothness, defined as follows: C i = (x i+2 - x i ) 2 + (y i+2 - y i ) 2 , i ∈ [1, n - 2] (x i , y i ) is the current grid position, (x i+2 , y i+2 ) is the position of the second grid below. The smoothness of the path is judged according to the size of C i . S i is the turning penalty value at the current position; At the same time, add an obstacle distance penalty function, defined as follows: (x i , y i ) is the current grid position, (x o , y o ) is the grid position of the nearest obstacle, O i is the Euclidean distance between the current position and the nearest obstacle, D i is the obstacle distance penalty value of the current position; The three items of path length, path smoothness, and obstacle distance penalty function together constitute the evaluation index of the fitness function; F(k) is the fitness function value, and ε1 and ε2 are user-defined variables; Introduce the RDP algorithm to remove the redundant part of the path after calculating the path smoothness.

6. The path planning method integrating D*Lite, simulated annealing and genetic algorithm according to claim 1, characterized in that In step S6, applying the generated population to the selection operator of the integrated simulated annealing algorithm and introducing the elite retention strategy includes the following steps: Optimize the roulette wheel method used by the genetic algorithm in the selection operator, and the formula is as follows: ε is a user-defined parameter for smoothing the fitness value, f(x) is the fitness value of the corresponding individual, and x j , x k , x i are all individuals in the population; Introduce the elite individual retention strategy. Before selecting individuals by roulette wheel selection method each time, store the individuals whose fitness values are before the custom percentage as elite individuals; Integrate the simulated annealing algorithm to retain individuals with poor fitness values with a certain probability; the probability formula is as follows: Among them, F(W1) and F(W2) are the fitness values of the corresponding individuals, T0 is the initial temperature, q is the cooling coefficient, and GEN is the number of iterations.

7. The path planning method integrating D*Lite, simulated annealing and genetic algorithm according to claim 1, characterized in that In step S7, apply the population after the selection operator to the adaptive crossover and mutation operators, including the following steps: Set the range and initial value of the crossover and mutation probabilities; Introduce the sigmoid function to flexibly adjust the probability value, and the formula is as follows: Among them, p is the final crossover probability value, p max and p min are the set maximum and minimum crossover probability values respectively, f max and f min are the maximum and minimum fitness values in the population respectively, x o is the set center point of fitness difference, and k is a parameter controlling the steepness of the curve.

8. The path planning method integrating D*Lite, simulated annealing and genetic algorithm according to claim 1, characterized in that In step S8, recalculate the fitness values of all individuals in the population after the crossover and mutation operators, and optimize the path redundant points again. The specific implementation is the same as that in step S5.

9. The path planning method integrating D*Lite, simulated annealing and genetic algorithm according to claim 1, characterized in that, In step S9, judge whether the path after re-optimization meets the optimal conditions; if it meets, output the optimal solution after iteration, including the following steps: First, judge whether the fitness value meets the expected requirements; Then, judge whether the number of population iterations reaches the limited maximum number; Finally, compare the fitness values of all individuals and output the individual with the highest fitness value as the current optimal solution.

10. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the path planning method integrating D*Lite, simulated annealing and genetic algorithm as described in any one of claims 1 to 9.

Images