Mobile robot path planning method and system fusing improved A* algorithm and adaptive genetic algorithm

Abstract

The invention discloses a mobile robot path planning method and system fusing an improved A * algorithm and an adaptive genetic algorithm, and the method comprises the steps: reducing redundant search in a dense obstacle region and enhancing path exploration in a sparse region through improving a dynamic weight mechanism of the A * algorithm, and giving consideration to the search efficiency and path optimality; the problems that a traditional algorithm is slow in convergence and prone to local optimum are solved in combination with an adaptive genetic algorithm, meanwhile, a path repairing mechanism is introduced, A * search and replacement point adjustment are dynamically fused, environment changes are dealt with in real time, the average path is shortened, and the path planning success rate under environment disturbance is increased. The advantages of the two algorithms are complementary, the improved A * algorithm provides a high-quality initial path, the adaptive genetic algorithm optimizes multi-target parameters, and finally the optimal comprehensive performance is realized in the aspects of path length, smoothness and obstacle avoidance adaptability.

Description

Technical Field

[0001] The present invention relates to the field of mobile perception technology, and more specifically, to a mobile robot path planning method and system that integrates an improved A* algorithm and an adaptive genetic algorithm. Background Art

[0002] In the field of autonomous mobile robot navigation, path planning methods directly impact the efficiency and safety of mobile robots. Currently, mobile robot path planning primarily relies on two methods: the A* algorithm and the genetic algorithm (GA). The A* algorithm and the genetic algorithm (GA) are two commonly used methods for path planning. The A* algorithm is a classic heuristic search algorithm that gradually expands the optimal path by evaluating the cost function from the current node to the target node. However, both the A* algorithm and the genetic algorithm face significant technical bottlenecks in terms of real-time performance, path optimization capabilities, and environmental adaptability. In complex environments, the A* algorithm requires traversing a large number of nodes, resulting in high computational complexity and reduced search efficiency. Due to the limitations of the search strategy, the generated path may contain an excessive number of turning points, affecting the path's smoothness and practical feasibility. The traditional A* algorithm uses a fixed heuristic function and lacks adaptability to environmental changes, which can lead to reduced search efficiency and path quality. While expanding the neighborhood search improves path smoothness, it also dramatically increases the number of nodes traversed, making it difficult to meet the real-time response requirements of dynamic scenarios. The weighted optimization strategy lacks dynamic environmental awareness, resulting in insufficient robustness in path planning results when obstacle distribution changes suddenly. The angle search algorithm relies on fixed threshold constraints and cannot adaptively adjust the search direction, resulting in local path redundancy. The adaptive parameter mechanism relies on preset rules and is prone to falling into local optimality or convergence lags in complex obstacle environments. The failure to incorporate prior environmental information (such as obstacle density and motion trends) leads to an increase in invalid path iterations. The single initialization strategy generates a large number of infeasible paths, prolonging the algorithm's warm-up period.

[0003] Existing improvement solutions are often limited to single-algorithm optimization, failing to address the fundamental contradiction between the weak global planning capabilities of the A* algorithm and the low local refinement efficiency of the genetic algorithm. Particularly in dynamic and complex environments, traditional methods are unable to simultaneously achieve the core requirements of low computational latency, high path quality, and strong environmental adaptability, severely hindering improvements in mobile robot navigation performance. Summary of the Invention

[0004] In order to overcome the defects of existing path planning methods, such as low real-time calculation efficiency, low path generation quality and weak adaptability to dynamic environments, the present invention provides a mobile robot path planning method and system that integrates an improved A* algorithm and an adaptive genetic algorithm.

[0005] In order to solve the above technical problems, the technical solutions of the present invention are as follows:

[0006] The present invention provides a mobile robot path planning method that integrates an improved A* algorithm and an adaptive genetic algorithm, the method comprising:

[0007] Collect the working environment information of the mobile robot, build a two-dimensional grid map based on the working environment information of the mobile robot, and set the initial point and target point of the mobile robot;

[0008] Generate an initial path according to the improved A* algorithm, and dynamically adjust the weight coefficient of the A* algorithm according to the obstacle density;

[0009] Optimize the initial path based on the adaptive genetic algorithm to generate the optimal path;

[0010] If there are still obstacles in the optimal path, the optimal path is repaired according to the preset path repair mechanism, and the final optimal path is output; if there are no obstacles in the optimal path, the optimal path is directly output;

[0011] A path planning route of the mobile robot is generated based on the optimal path.

[0012] Preferably, the formula of the weight coefficient is:

[0013]

[0014] Where ρ is the obstacle density.

[0015] Preferably, the A* algorithm is specifically:

[0016] f(n)=g(n)+α(ρ)h(n)

[0017] Among them, g(n) is the cost function; h(n) is the heuristic function; α(ρ) is the weight coefficient.

[0018] Preferably, the formula of the heuristic function is:

[0019] h(n)=(|x n -x g |+|y n -y g |)

[0020] Among them, (x n ,y n ) is the coordinate of the current node; (x g ,y g ) is the coordinate of the target node;

[0021] The formula of the cost function is:

[0022] g(n)=g(p)+d(p,n)

[0023] Where p is the predecessor node of n; d(p,n) is the cost of moving from p to n.

[0024] Preferably, the initial path is generated according to the improved A* algorithm, including:

[0025] Establishing an open list and a closed list, wherein the open list is used to store nodes to be searched and the nodes in the open list are sorted according to the evaluation function value; the closed list is used to store visited nodes;

[0026] Select the node with the minimum evaluation function value in the open list as the current node. If the current node is the target node, stop searching and return the path; otherwise, traverse to the next node;

[0027] Recalculate the cost function value of the current node based on all feasible neighbor nodes; if the neighbor node is not in the open list or its calculated cost function value is greater than the initial cost function value; then update the evaluation function value and heuristic function value of the current node and add all feasible neighbor nodes to the open list;

[0028] If the feasible neighbor node of the current node is the target node, stop searching and return the path; if the open list is empty at this time, the target is unreachable and return an empty path; if none of the above conditions are met, reselect a node in the open list;

[0029] Generate an initial path based on the return path.

[0030] Preferably, the initial path is optimized based on an adaptive genetic algorithm to generate an optimized path, including:

[0031] The population is initialized according to the initial path, where each individual is a path sequence. The improved A algorithm adjusts the path search direction through a dynamic heuristic factor and introduces a collision penalty mechanism when generating the path to ensure that the path avoids obstacles.

[0032] The path length, smoothness, number of turning points and collision penalty of the initial path population are optimized by the fitness function in the adaptive genetic algorithm;

[0033] Sort in descending order according to the fitness function value, and use the elite retention strategy to filter out the top N ranking result values ​​of each generation e Individuals directly enter the next generation, N e calculated based on elite retention rates and population size dynamics;

[0034] Dynamically adjust the adaptive crossover rate and adaptive mutation rate of the next generation according to individual fitness;

[0035] The parent individuals are selected through roulette wheel, crossover and mutation operations are performed to generate offspring, and the heuristic function of the improved A algorithm is used to guide the mutation direction of the offspring;

[0036] If there is an obstacle conflict on the child path, the improved A algorithm is called to repair the path;

[0037] Output the optimized path based on the repaired path.

[0038] Preferably, the fitness function is:

[0039] F(X)=-(ω1L(X)+ω2S(X)+ω3T(X)+P(X))

[0040] Among them, ω1 is the first weight coefficient; ω2 is the second weight coefficient; ω3 is the third weight coefficient; L(X) is the path length; S(X) is the smoothness of the path; T(X) is the number of turning points in the path; P(X) is the collision penalty term of the path.

[0041] Preferably, the adaptive crossover rate is:

[0042]

[0043] in, is the maximum value of the crossover rate; is the minimum value of the crossover rate; f i is the fitness function value of the current individual; f max is the maximum fitness function value in the current population; f avg is the average fitness function value of the current population;

[0044] The adaptive mutation rate is:

[0045]

[0046] in, is the maximum value of the mutation rate; is the minimum value of the mutation rate.

[0047] Preferably, the preset path repair mechanism includes:

[0048] Detect illegal points in the path and perform an improved A* algorithm search on all detected illegal points;

[0049] If the search is successful, the path is changed and a detour is generated; if the search fails, an alternative point is found and the nearest legal point is searched to generate a new path;

[0050] Verify the connectivity of the generated new path and the detour path. If the generated path is connected, replace the original path segment with the current path segment; if the generated path is not connected, return to continue the improved A* algorithm search.

[0051] The present invention also provides a mobile robot path planning system that integrates an improved A* algorithm and an adaptive genetic algorithm, which is used to implement the above method, including:

[0052] The environment construction module collects the working environment information of the mobile robot, constructs a two-dimensional grid map based on the working environment information of the mobile robot, and sets the initial point and target point of the mobile robot;

[0053] An initial path generation module generates an initial path according to an improved A* algorithm and dynamically adjusts the weight coefficient of the A* algorithm according to the density of obstacles;

[0054] The optimization path generation module optimizes the initial path based on the adaptive genetic algorithm and generates the optimized path;

[0055] A path repair module, which performs path repair on the optimal path according to a preset path repair mechanism if there are still obstacles in the optimal path and outputs the final optimal path; if there are no obstacles in the optimal path, it directly outputs the optimal path;

[0056] A path planning module generates a path planning route for the mobile robot based on the optimal path.

[0057] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0058] This paper proposes a mobile robot path planning method and system that integrates an improved A* algorithm and an adaptive genetic algorithm. By improving the dynamic weighting mechanism of the A* algorithm, this method reduces redundant searches in obstacle-dense areas and enhances path exploration in sparse areas, balancing search efficiency and path optimality. The adaptive genetic algorithm addresses the slow convergence and local optimality issues of traditional algorithms. Furthermore, a path repair mechanism is introduced, dynamically integrating A* search with alternative point adjustment to respond to environmental changes in real time, shorten the average path, and improve the path planning success rate under environmental disturbances. The two algorithms complement each other: the improved A* algorithm provides a high-quality initial path, while the adaptive genetic algorithm optimizes multiple objective parameters, ultimately achieving optimal overall performance in terms of path length, smoothness, and obstacle avoidance adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a flow chart of the mobile robot path planning method that integrates the improved A* algorithm and the adaptive genetic algorithm described in Example 1;

[0060] Figure 2 This is a schematic structural diagram of a flow chart of a mobile robot path planning method that integrates an improved A* algorithm and an adaptive genetic algorithm as described in Example 2;

[0061] Figure 3 This is a flow chart of the improved A* algorithm described in Example 2;

[0062] Figure 4 This is a flow chart of the adaptive genetic algorithm described in Example 2;

[0063] Figure 5 This is a flow chart of the path repair mechanism described in Example 2;

[0064] Figure 6 Schematic diagram of the mobile robot path planning simulation described in Example 2;

[0065] Figure 7 This is a structural diagram of the mobile robot path planning system that integrates the improved A* algorithm and the adaptive genetic algorithm described in Example 3. DETAILED DESCRIPTION

[0066] The accompanying drawings are for illustrative purposes only and are not to be construed as limiting this patent;

[0067] In order to better illustrate this embodiment, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product size;

[0068] It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0069] The technical solution of the present invention is further described below with reference to the accompanying drawings and embodiments.

[0070] Example 1

[0071] This embodiment provides a mobile robot path planning method that integrates the improved A* algorithm and the adaptive genetic algorithm. Figure 1 As shown, the method includes:

[0072] Collect the working environment information of the mobile robot, build a two-dimensional grid map based on the working environment information of the mobile robot, and set the initial point and target point of the mobile robot;

[0073] Generate an initial path according to the improved A* algorithm, and dynamically adjust the weight coefficient of the A* algorithm according to the obstacle density;

[0074] Optimize the initial path based on the adaptive genetic algorithm to generate the optimal path;

[0075] If there are still obstacles in the optimal path, the optimal path is repaired according to the preset path repair mechanism, and the final optimal path is output; if there are no obstacles in the optimal path, the optimal path is directly output;

[0076] A path planning route of the mobile robot is generated based on the optimal path.

[0077] The formula of the weight coefficient is:

[0078]

[0079] Where ρ is the obstacle density.

[0080] The A* algorithm is specifically:

[0081] f(n)=g(n)+α(ρ)h(n)

[0082] Among them, g(n) is the cost function; h(n) is the heuristic function; α(ρ) is the weight coefficient.

[0083] The formula of the heuristic function is:

[0084] h(n)=(|x n -x g |+|y n -y g |)

[0085] Among them, (x n ,y n ) is the coordinate of the current node; (x g ,y g ) is the coordinate of the target node;

[0086] The formula of the cost function is:

[0087] g(n)=g(p)+d(p,n)

[0088] Where p is the predecessor node of n; d(p,n) is the cost of moving from p to n.

[0089] Generate the initial path based on the improved A* algorithm, including:

[0090] Establishing an open list and a closed list, wherein the open list is used to store nodes to be searched and the nodes in the open list are sorted according to the evaluation function value; the closed list is used to store visited nodes;

[0091] Select the node with the minimum evaluation function value in the open list as the current node. If the current node is the target node, stop searching and return the path; otherwise, traverse to the next node;

[0092] Recalculate the cost function value of the current node based on all feasible neighbor nodes; if the neighbor node is not in the open list or its calculated cost function value is greater than the initial cost function value; then update the evaluation function value and heuristic function value of the current node and add all feasible neighbor nodes to the open list;

[0093] If the feasible neighbor node of the current node is the target node, stop searching and return the path; if the open list is empty at this time, the target is unreachable and return an empty path; if none of the above conditions are met, reselect a node in the open list;

[0094] Generate an initial path based on the return path.

[0095] The initial path is optimized based on the adaptive genetic algorithm to generate the optimized path, including:

[0096] The population is initialized according to the initial path, where each individual is a path sequence. The improved A algorithm adjusts the path search direction through a dynamic heuristic factor and introduces a collision penalty mechanism when generating the path to ensure that the path avoids obstacles.

[0097] The path length, smoothness, number of turning points and collision penalty of the initial path population are optimized by the fitness function in the adaptive genetic algorithm;

[0098] Sort in descending order according to the fitness function value, and use the elite retention strategy to filter out the top N ranking result values ​​of each generation e Individuals directly enter the next generation, N e calculated based on elite retention rates and population size dynamics;

[0099] Dynamically adjust the adaptive crossover rate and adaptive mutation rate of the next generation according to individual fitness;

[0100] The parent individuals are selected through roulette wheel, crossover and mutation operations are performed to generate offspring, and the heuristic function of the improved A algorithm is used to guide the mutation direction of the offspring;

[0101] If there is an obstacle conflict on the child path, the improved A algorithm is called to repair the path;

[0102] Output the optimized path based on the repaired path.

[0103] The fitness function is:

[0104] F(X)=-(ω1L(X)+ω2S(X)+ω3T(X)+P(X))

[0105] Among them, ω1 is the first weight coefficient; ω2 is the second weight coefficient; ω3 is the third weight coefficient; L(X) is the path length; S(X) is the smoothness of the path; T(X) is the number of turning points in the path; P(X) is the collision penalty term of the path.

[0106] The adaptive crossover rate is:

[0107]

[0108] in, is the maximum value of the crossover rate; is the minimum value of the crossover rate; f i is the fitness function value of the current individual; f max is the maximum fitness function value in the current population; f avg is the average fitness function value of the current population;

[0109] The adaptive mutation rate is:

[0110]

[0111] in, is the maximum value of the mutation rate; is the minimum value of the mutation rate.

[0112] The preset path repair mechanism includes:

[0113] Detect illegal points in the path and perform an improved A* algorithm search on all detected illegal points;

[0114] If the search is successful, the path is changed and a detour is generated; if the search fails, an alternative point is found and the nearest legal point is searched to generate a new path;

[0115] Verify the connectivity of the generated new path and the detour path. If the generated path is connected, replace the original path segment with the current path segment; if the generated path is not connected, return to continue the improved A* algorithm search.

[0116] Example 2

[0117] This embodiment provides a mobile robot path planning method that integrates an improved A* algorithm and an adaptive genetic algorithm, the method comprising:

[0118] Collect the working environment information of the mobile robot, build a two-dimensional grid map based on the working environment information of the mobile robot, and set the initial point and target point of the mobile robot;

[0119] Generate an initial path according to the improved A* algorithm, and dynamically adjust the weight coefficient of the A* algorithm according to the obstacle density;

[0120] Optimize the initial path based on the adaptive genetic algorithm to generate the optimal path;

[0121] If there are still obstacles in the optimal path, the optimal path is repaired according to the preset path repair mechanism, and the final optimal path is output; if there are no obstacles in the optimal path, the optimal path is directly output;

[0122] A path planning route of the mobile robot is generated based on the optimal path.

[0123] In a specific embodiment, Figure 2 As shown in step 1: map construction. According to the slope characteristics of the three-dimensional terrain environment and the robot motion geometry characteristics, a two-dimensional grid map is established to simulate the environment, and the starting point S = (x s ,y s ) and the end point G=(x g ,y g ).

[0124] Step 2: Path search. Use the improved A* algorithm to generate the initial path P inital , where a dynamically adjusted heuristic weight α(ρ) is used, and the weight value changes adaptively according to the obstacle density to enhance search flexibility. The cost function g(n) and the heuristic estimation function h(n) of the current node are calculated, and the weight coefficient is dynamically adjusted to adapt to different environments.

[0125] Step 3: Path optimization. Adaptive genetic algorithm is used to optimize P inital Optimize and generate the optimized path P optimize The fitness function comprehensively considers the path length L, the number of turning points T, the smoothness S, and the obstacle collision penalty P. The crossover probability P c and mutation probability P m Dynamically adjust based on population fitness to improve search diversity. Adopt goal-oriented mutation strategy to avoid falling into local optimum.

[0126] Step 4: Path repair. In the adaptive genetic algorithm iteration, after the crossover operation, the infeasible paths in the adaptive genetic algorithm population are corrected by detecting path breakpoints or obstacle crossing points. The improved A* algorithm search and the path repair mechanism assisted by the alternative point search are used to ensure that the adaptive genetic algorithm finally obtains a feasible path P. final .

[0127] like Figure 3 Specifically, step 2 includes:

[0128] S1. The calculation method of dynamic weight coefficient α(ρ) is:

[0129]

[0130] The obstacle density G i,j= 1 indicates that the grid cell is an obstacle; otherwise, it is 0. m and n are the number of rows and columns of the obstacle in the two-dimensional grid map, respectively. 1(·) is a 0-1 indicator function, taking the value 1 when the given condition holds and 0 otherwise. When there are no obstacles (ρ = 0), α = 1. As the obstacle density increases, α gradually increases, increasing the influence of the heuristic function and reducing the exploration of potential dead ends. Reducing the weight of α in areas with sparse obstacles facilitates a wider search space and improves path optimality. The change in α is constrained by the square of ρ to ensure that it does not increase too quickly.

[0131] S2. Calculate Manhattan distance as a heuristic function:

[0132] h(n)=(|x n -x g |+|y n -y g |)

[0133] Among them, (x n ,y n ) and (x g ,y g ) are the coordinates of the current node and the target node respectively.

[0134] S3. Actual cost function

[0135] g(n)=g(p)+d(p,n)

[0136] Among them, g(n) is the actual cost from the starting point to node n; p is the predecessor node of n, that is, the node on the path from p to n; d(p,n) is the moving cost from p to n.

[0137] S4. The evaluation function of the improved A* algorithm search consists of the actual cost and the heuristic cost:

[0138] f(n)=g(n)+α(ρ)h(n)

[0139] Among them, f(n) is the algorithm's indicator for evaluating whether the current node has a promising future; g(n) is the actual cost from the starting point to n; h(n) is the heuristic function, which estimates the shortest distance from n to the target.

[0140] S5. The entire improved A* algorithm performs searches by maintaining two sets: an open list O and a closed list C. The open list O stores the nodes to be explored and is sorted by the f(n) value (i.e., the total cost); the closed list C stores the nodes that have been visited to avoid repeated searches.

[0141] S6. Select the next expansion node: select the node with the minimum f(n) value in the open list O as the current node c; if the current node c is the target goal, the search ends and the path is returned.

[0142] S7. Update neighbor nodes: For all feasible neighbors n of the current node c, calculate the new cost g′(n) = g(c) + d(c,n); if neighbor n is not in the open list or the new cost is better, update g(n) and f(n), record n's predecessor node c (for backtracking path), and then add n to the open list.

[0143] S8. The conditions for the termination of the improved A* algorithm are: if the current node n = goal, stop searching and backtrack the path; if the open list O is empty, it means that the goal is unreachable and return an empty path. If none of the above conditions are met, return to S6.

[0144] like Figure 4 As shown, further, in step 3, the adaptive genetic algorithm optimizes the selection path, specifically including the following:

[0145] S1. Use the path generated by the improved A* algorithm to initialize the population. Initial population size where N p represents the population size, each individual X i Is a path sequence, initialized based on the improved A* algorithm:

[0146] X i =A*(G,s,g,α),α∈[1.0,2.0]

[0147] Where F is the grid environment, s and g are the starting point and end point respectively, and α is the dynamic heuristic factor.

[0148] S2, using adaptive genetic algorithm to inital Optimize and generate the optimized path P optimize , the quality of any given path X is measured based on the following fitness function F(X):

[0149] F(X)=-(ω1L(X)+ω2S(X)+ω3T(X)+P(X))

[0150] ω1, ω2, ω3 are weight coefficients, Represents the length of path X; represents the smoothness of path X; T(X) = |X| represents the total number of steps in path X, that is, the number of turning points; P(X) is the collision penalty for path X, where χ is a sufficiently large positive number

[0151]

[0152] S3, adopt the elite retention strategy, retain the top N fitness in each generation e Individuals directly enter the next generation population, and N is determined by rounding down according to the following formula e :

[0153]

[0154] Where η∈(0.05,0.2) is the elite retention rate, M p is the population size.

[0155] S4, adaptive cross rate P c It is used to control the probability of crossover operation and is dynamically adjusted according to fitness. The calculation method is as follows:

[0156]

[0157] in, and are the maximum and minimum values ​​of the crossover rate respectively; f i is the fitness value of the current individual; f max is the maximum fitness value in the current population; f avg is the average fitness value of the current population. i When the number of individuals is large (excellent individuals), the crossover rate P c Smaller, to reduce the probability of destroying good individuals; when the individual fitness f i When it is smaller (poor individual), the crossover rate P c near Increase opportunities to explore new solutions;

[0158] S5, adaptive mutation rate P m Based on the fitness, the probability of mutation is adjusted so that individuals with better fitness have a lower mutation rate, while individuals with poorer fitness have a higher mutation rate, thereby avoiding premature algorithm convergence and enhancing the global optimization capability. The calculation method is as follows:

[0159]

[0160] in, and are the maximum and minimum values ​​of the mutation rate respectively; if f max -f avg If is too small, it is considered that the population has converged, the population diversity is extremely low, and it may fall into a local optimum. In this case, the method of forcibly increasing the mutation rate can be adopted, directly making This increases the search range and breaks the stagnation of the population.

[0161] S6, selection: The fitness value is converted into a probability distribution through the softmax function, and the parent is selected using roulette wheel Parent∝softmax(F(X))

[0162] S7, crossover and mutation: generate offspring Child = Mutate (Crossover (Parent1, Parent2), p m After the crossover operation, the path repair mechanism is activated as needed to ensure that the generated path is feasible (i.e., it avoids obstacles) and meets the requirements of safety and effectiveness. The mutation operation is then performed, and the heuristic function of the improved A* algorithm is introduced to guide the mutation direction, thereby enhancing the targetedness and effectiveness of the mutation.

[0163] S8. Output the optimized path. If there are still obstacles on the optimal path, the improved A* algorithm is also used to correct them:

[0164]

[0165] like Figure 5 As shown, further, the path repair mechanism in step 4 is as follows:

[0166] S1. Assume that a given path Q = {q0,q1,...,q k}, where q i Denotes the i-th point on the path. Let the set of illegal points in the path be Q invalid ={q i |grid(q i )=1,1≤i≤k-1}.

[0167] S2, perform the improved A* algorithm search. For each illegal point q i ∈Q invalid Perform improved A* algorithm search;

[0168]

[0169] If the search is successful, the update path is

[0170]

[0171] And enter S4, otherwise enter the alternative point search.

[0172] S3, perform alternative point search as needed. Only when the improved A* algorithm fails to search, it will search at the illegal point q. i Find the nearest legal point in the neighborhood:

[0173]

[0174] Satisfy: B(q i,r)={q|||qq i ||≤r,grid(q)=0}, that is, all i A feasible point within the radius r. The new point q i ′The path must be continuous: valid(q i-1 ,q′ i )=True,valid(q′ i ,q i+1 )=True, where If you find q i ′, then the update path is Q′={q0,...,q i-1 ,q i ′,q i+1 ,...,q k}

[0175] S4. Verify the connectivity of the new path. After fixing all illegal points, it is required that for any point on the new path Q′, the following conditions are met:

[0176]

[0177] If the new path is still discontinuous, return to S2 and re-execute the repair process until a feasible path is found.

[0178] The robot path planning algorithm flow chart proposed in this embodiment is shown in the following experimental simulation to demonstrate the effect of the method. Figure 6 As shown, the path planning results of the patented method in a 30×30 grid environment are shown. First, the environment is modeled into a map containing randomly distributed obstacles, which are presented in the form of black grids. The starting point S is located in the upper left corner of the map with coordinates (0,0), and the end point G is located in the lower right corner of the map with coordinates (29,29). A hybrid strategy combining the improved A* algorithm and the adaptive genetic algorithm is introduced to solve the optimal path from the starting point to the end point. This path is represented by a blue solid line. It avoids all obstacles throughout the entire process, presents a continuous and smooth feature, and significantly reduces the number of turning points. In terms of obstacle avoidance capability, the path achieves rapid detours by dynamically adjusting weights to narrow the search range in areas with dense obstacles, such as the central area of ​​the map; in areas where obstacles are relatively sparse, the genetic algorithm is used to further optimize the smoothness of the path, effectively reducing redundant movements, thereby improving the efficiency and feasibility of the overall path. Table 1 compares the core indicators of the A* algorithm, genetic algorithm and this patent:

[0179] Table 1

[0180]

[0181] The data shows that by improving the A* algorithm, the efficiency of path search in complex environments was improved, and combined with the local optimization of the adaptive genetic algorithm, the path length was shortened by approximately 8.2%. The adaptive mutation strategy effectively reduced directional mutations and improved smoothness by approximately 30.4%, making the path more suitable for robot motion control. The comprehensive path quality index (CPQI) reached 0.70, an increase of approximately 133% compared to the A algorithm, verifying the effectiveness of multi-objective optimization. In terms of energy saving, the optimized path reduced unnecessary movement and energy consumption, improving the robot's operating efficiency. At the same time, the smoothness of the path and the reduction of directional mutations also enhanced the robot's stability and safety during movement, reducing the risk of collision and failure.

[0182] Example 3

[0183] This embodiment also provides a mobile robot path planning system that integrates the improved A* algorithm and the adaptive genetic algorithm. Figure 7 As shown, the method for implementing the method of embodiment 1 and embodiment 2 includes:

[0184] The environment construction module collects the working environment information of the mobile robot, constructs a two-dimensional grid map based on the working environment information of the mobile robot, and sets the initial point and target point of the mobile robot;

[0185] An initial path generation module generates an initial path according to an improved A* algorithm and dynamically adjusts the weight coefficient of the A* algorithm according to the density of obstacles;

[0186] The optimization path generation module optimizes the initial path based on the adaptive genetic algorithm and generates the optimized path;

[0187] A path repair module, which performs path repair on the optimal path according to a preset path repair mechanism if there are still obstacles in the optimal path and outputs the final optimal path; if there are no obstacles in the optimal path, it directly outputs the optimal path;

[0188] A path planning module generates a path planning route for the mobile robot based on the optimal path.

[0189] The same or similar reference numerals correspond to the same or similar components;

[0190] The terms used in the drawings to describe positional relationships are for illustrative purposes only and should not be construed as limiting this patent;

[0191] Obviously, the above embodiments of the present invention are merely examples for the purpose of clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Those skilled in the art will appreciate that other variations or modifications can be made based on the above description. It is not necessary and impossible to enumerate all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A mobile robot path planning method that integrates an improved A* algorithm and an adaptive genetic algorithm, characterized in that: The method comprises: Collect the working environment information of the mobile robot, build a two-dimensional grid map based on the working environment information of the mobile robot, and set the initial point and target point of the mobile robot; Generate an initial path according to the improved A* algorithm, and dynamically adjust the weight coefficient of the A* algorithm according to the obstacle density; Optimize the initial path based on the adaptive genetic algorithm to generate the optimal path; If there are still obstacles in the optimal path, the optimal path is repaired according to the preset path repair mechanism, and the final optimal path is output; if there are no obstacles in the optimal path, the optimal path is directly output; A path planning route of the mobile robot is generated based on the optimal path.

2. The mobile robot path planning method integrating the improved A* algorithm and the adaptive genetic algorithm according to claim 1 is characterized in that: The formula of the weight coefficient is: Where ρ is the obstacle density.

3. The mobile robot path planning method integrating the improved A* algorithm and the adaptive genetic algorithm according to claim 1 is characterized in that: The A* algorithm is specifically: f(n)=g(n)+α(ρ)h(n) Among them, g(n) is the cost function; h(n) is the heuristic function; α(ρ) is the weight coefficient.

4. The mobile robot path planning method integrating the improved A* algorithm and the adaptive genetic algorithm according to claim 3 is characterized in that: The formula of the heuristic function is: h(n)=(|x n -x g |+|y n -y g |) Among them, (x n ,y n ) is the coordinate of the current node; (x g ,y g ) is the coordinate of the target node; The formula of the cost function is: g(n)=g(p)+d(p,n) Where p is the predecessor node of n; d(p,n) is the cost of moving from p to n.

5. The mobile robot path planning method integrating the improved A* algorithm and the adaptive genetic algorithm according to claim 1 is characterized in that: Generate the initial path based on the improved A* algorithm, including: Establishing an open list and a closed list, wherein the open list is used to store nodes to be searched and the nodes in the open list are sorted according to the evaluation function value; the closed list is used to store visited nodes; Select the node with the minimum evaluation function value in the open list as the current node. If the current node is the target node, stop searching and return the path; otherwise, traverse to the next node; Recalculate the cost function value of the current node based on all feasible neighbor nodes; if the neighbor node is not in the open list or its calculated cost function value is greater than the initial cost function value; then update the evaluation function value and heuristic function value of the current node and add all feasible neighbor nodes to the open list; If the feasible neighbor node of the current node is the target node, stop searching and return the path; if the open list is empty at this time, the target is unreachable and return an empty path; if none of the above conditions are met, reselect a node in the open list; Generate an initial path based on the return path.

6. The mobile robot path planning method integrating the improved A* algorithm and the adaptive genetic algorithm according to claim 1 is characterized in that: The initial path is optimized based on the adaptive genetic algorithm to generate the optimized path, including: The population is initialized according to the initial path, where each individual is a path sequence. The improved A algorithm adjusts the path search direction through a dynamic heuristic factor and introduces a collision penalty mechanism when generating the path to ensure that the path avoids obstacles. The path length, smoothness, number of turning points and collision penalty of the initial path population are optimized by the fitness function in the adaptive genetic algorithm; Sort in descending order according to the fitness function value, and use the elite retention strategy to filter out the top N ranking result values ​​of each generation e Individuals directly enter the next generation, N e calculated based on elite retention rates and population size dynamics; Dynamically adjust the adaptive crossover rate and adaptive mutation rate of the next generation according to individual fitness; The parent individuals are selected through roulette wheel, crossover and mutation operations are performed to generate offspring, and the heuristic function of the improved A algorithm is used to guide the mutation direction of the offspring; If there is an obstacle conflict on the child path, the improved A algorithm is called to repair the path; Output the optimized path based on the repaired path.

7. The mobile robot path planning method integrating the improved A* algorithm and the adaptive genetic algorithm according to claim 6 is characterized in that: The fitness function is: F(X)=-(ω1L(X)+ω2S(X)+ω3T(X)+P(X)) Among them, ω1 is the first weight coefficient; ω2 is the second weight coefficient; ω3 is the third weight coefficient; L(X) is the path length; S(X) is the smoothness of the path; T(X) is the number of turning points in the path; P(X) is the collision penalty term of the path.

8. The mobile robot path planning method integrating the improved A* algorithm and the adaptive genetic algorithm according to claim 6 is characterized in that: The adaptive crossover rate is: in, is the maximum value of the crossover rate; is the minimum value of the crossover rate; f i is the fitness function value of the current individual; f max is the maximum fitness function value in the current population; f avg is the average fitness function value of the current population; The adaptive mutation rate is: in, is the maximum value of the mutation rate; is the minimum value of the mutation rate.

9. The mobile robot path planning method integrating the improved A* algorithm and the adaptive genetic algorithm according to claim 6 is characterized in that: The preset path repair mechanism includes: Detect illegal points in the path and perform an improved A* algorithm search on all detected illegal points; If the search is successful, the path is changed and a detour is generated; if the search fails, an alternative point is found and the nearest legal point is searched to generate a new path; Verify the connectivity of the generated new path and the detour path. If the generated path is connected, replace the original path segment with the current path segment; if the generated path is not connected, return to continue the improved A* algorithm search.

10. A mobile robot path planning system integrating an improved A* algorithm and an adaptive genetic algorithm, used to implement the method of claims 1-9, characterized in that: include: The environment construction module collects the working environment information of the mobile robot, constructs a two-dimensional grid map based on the working environment information of the mobile robot, and sets the initial point and target point of the mobile robot; An initial path generation module generates an initial path according to an improved A* algorithm and dynamically adjusts the weight coefficient of the A* algorithm according to the density of obstacles; The optimization path generation module optimizes the initial path based on the adaptive genetic algorithm and generates the optimized path; A path repair module, which performs path repair on the optimal path according to a preset path repair mechanism if there are still obstacles in the optimal path and outputs the final optimal path; if there are no obstacles in the optimal path, it directly outputs the optimal path; A path planning module generates a path planning route for the mobile robot based on the optimal path.

Images