Mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm

Abstract

The invention provides a mobile robot path planning method based on a dynamic information entropy ant colony optimization algorithm. The mobile robot path planning method comprises the following steps: S1, environment modeling: setting a starting point and an ending point of a robot and a position of an obstacle; s2, setting initial parameters of an ant algorithm; s3, performing non-uniform pheromone initialization: performing non-uniform initialization on the pheromone according to the obstacle distance, for each non-obstacle node (i, j), calculating the Euclidean distance minDist from the non-obstacle node (i, j) to the nearest obstacle, and calculating and initializing the pheromone concentration tauij; and S4, constructing a path. According to the invention, through a non-uniform pheromone initialization method based on the obstacle distance, ants are guided to preferentially select nodes far away from the obstacle, the death rate of the ants is reduced, and the diversity and efficiency of initial path search are improved; according to the method, a dynamic pheromone updating strategy based on information entropy and a self-adaptive pheromone volatilization mode are utilized, the search process is flexibly controlled, the convergence speed and the exploration capability of the algorithm are balanced, and the global search capability and efficiency of the algorithm are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of mobile robot path planning, and in particular relates to a mobile robot path planning method based on an ant colony optimization algorithm with dynamic information entropy. Background Art

[0002] Mobile robot path planning is one of the key research directions in the field of intelligent robots and is widely used in many fields such as automated warehousing, intelligent logistics, and unmanned driving. With the growing demand for intelligent manufacturing and service robots, the ability of mobile robots to navigate autonomously in complex environments has become particularly important. Effective path planning can not only reduce the risk of collision between robots and obstacles, but also shorten task execution time and reduce energy consumption, thereby significantly improving the overall performance of robot navigation. Therefore, how to design efficient and reliable path planning algorithms in complex and changing environments has become the focus of current research.

[0003] At present, researchers at home and abroad have proposed a variety of optimization algorithms for path planning of mobile robots, including genetic algorithms, simulated annealing algorithms, particle swarm algorithms, neural networks, etc. Among them, the Ant Colony Optimization (ACO) algorithm has been widely used in the field of path planning due to its good global search capabilities and ability to adapt to complex environments. The algorithm simulates the behavior of ants relying on pheromones during foraging, continuously searches and optimizes the path, and finally finds the optimal solution.

[0004] However, the traditional ant colony algorithm still has some shortcomings in path planning. First, when ants initially build paths, the uniform distribution of pheromones leads to a lack of effective guidance for the ant colony, which reduces the efficiency of subsequent searches. Second, the fixed pheromone evaporation coefficient (ρ value) leads to an unbalanced search and development. If the ρ value is too large, the algorithm will converge to the local optimum too early, while if it is too small, the pheromone will disappear too quickly, reducing the algorithm's memory and development capabilities, resulting in poor overall adaptability and difficulty in dynamically responding to the needs of different stages, thus affecting the optimization effect of the algorithm; in addition, the number of iterations of the algorithm is usually fixed, which may terminate prematurely before the optimal solution is found, or continue to iterate after the optimal solution is found, wasting computing resources. Summary of the invention

[0005] In view of the above technical problems, the present invention provides a mobile robot path planning method based on an ant colony optimization algorithm with dynamic information entropy, which improves the global optimal solution search capability and path planning efficiency and saves computing resources.

[0006] The present invention is helpful to overcome the problems of traditional ant colony algorithm in path planning, such as easy to fall into local optimum, low search efficiency, missed detection of optimal solution and waste of resources caused by fixed number of iterations.

[0007] Note that the description of these objectives does not prevent the existence of other objectives. One embodiment of the present invention does not need to achieve all of the above objectives. Objectives other than the above objectives can be extracted from the description of the specification, drawings, and claims.

[0008] The present invention achieves the above technical objectives through the following technical means.

[0009] A mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm comprises the following steps:

[0010] S1. Environment modeling: setting the robot’s starting point, end point, and obstacle locations;

[0011] S2. Set the initial parameters of the ant algorithm: set the maximum number of iterations T max , the number of ants k, the heuristic factor η, the pheromone factor α, the initial pheromone evaporation coefficient ρ 0 , initial maximum information entropy threshold H max , minimum information entropy threshold H min and entropy decay factor ω;

[0012] S3, non-uniform pheromone initialization: According to the obstacle distance, the pheromone is initialized non-uniformly. For each non-obstacle node (i, j), the Euclidean distance minDist to the nearest obstacle is calculated, and the pheromone concentration τ is calculated and initialized. ij ;

[0013] S4. Path construction:

[0014] S41, setting the starting point of the robot as the initial node of the ant, adding it to the taboo list, and calculating the path selection probability between the current node i and the candidate node j;

[0015] S42, using the roulette algorithm to select the next node based on the path selection probability, and adding the current node to the taboo list, repeating S41 and S42 until all ants reach the end point;

[0016] S43, calculating the information entropy of this iteration;

[0017] S44. Calculate the entropy guidance factor and entropy difference of this iteration:

[0018] S45, adaptive adjustment of pheromone evaporation coefficient;

[0019] S46, dynamic pheromone update: determine the pheromone update method based on the comparison between the information entropy of this iteration and the entropy guidance factor;

[0020] S47, iteration termination mechanism: calculate the entropy threshold, check whether the information entropy of this iteration of the algorithm is greater than the entropy threshold. If the condition is not met, return to S41 and start the iteration again from the starting point. If the condition is met, output the optimal solution found and the algorithm ends.

[0021] In the above scheme, the pheromone concentration τ in S3 ij Calculated by formula (1):

[0022]

[0023] Among them, maxPheromone represents the maximum pheromone concentration of the entire map, which is set to 1;

[0024] obstacleMap(i,j) indicates whether there is an obstacle at the position (i,j) on the map. If there is an obstacle, the pheromone concentration is 0;

[0025] minDist is the Euclidean distance from the current (i, j) node to its nearest obstacle;

[0026] σ is the standard deviation based on the distance.

[0027] In the above scheme, the S41 path selection probability is based on equation (2) and equation (3):

[0028]

[0029]

[0030] in, represents the probability of ant k moving from node i to node j;

[0031] d ij Represents the Euclidean distance between the current node i and the candidate node j;

[0032] τ ij and η ij are the pheromone concentration and heuristic information on the path respectively;

[0033] α and β are two weight parameters, which respectively measure the influence of pheromone concentration and heuristic information on path selection;

[0034] allowed k represents the set of nodes that ant k has not visited yet;

[0035] s is the node that the ant is about to go to;

[0036] τ is and η is They represent the pheromone concentration and heuristic information on the path from node i to node s respectively.

[0037] In the above scheme, step S43 calculates the information entropy of this iteration according to formula (4):

[0038]

[0039] Among them, H(X) is the path information entropy, X represents the set of paths, including all possible paths, and p(xi) represents the specific path x i The probability of being selected, occurrence(x i ) is the path x in the group i The number of times selected is the historical selection frequency of the ant on the path, and n is the number of paths generated in this iteration. This formula is used to measure the uncertainty or amount of information of the random variable, reflecting the degree of exploration of the path by the entire ant colony in the search space, and quickly locating the optimal solution.

[0040] In the above scheme, step S44 calculates the entropy guidance factor of this iteration according to formula (5):

[0041] H t =min{L iter (t-1),L iter (t-2),…,L iter (1)}-L best (t) (5)

[0042] Among them, H t represents the entropy guiding factor of the tth iteration, L best is the global optimal path so far, L iter (t-1) represents the best path of t-1 iteration, L iter (t-2) represents the best path of t-2 iterations..., L iter (1) represents the best path of the first iteration.

[0043] In the above scheme, step S44 calculates the entropy difference of this iteration according to formula (9):

[0044] ΔH=H x (t)-H x (t-1) (9)

[0045] H x (t) is the information entropy of this generation, H x (t-1) is the information entropy of the previous generation, and ΔH is the difference between the information entropy of this generation and the previous generation, that is, the entropy difference.

[0046] In the above scheme, the step S45 calculates the pheromone evaporation coefficient according to formula (10):

[0047] ρ=ρ 0(1+μΔH) (10)

[0048] Where ρ is the evaporation coefficient of pheromone, ρ 0 is the initialization evaporation coefficient, μ is the evaporation influence factor, and this formula is used to balance the convergence speed and exploration ability of the algorithm to achieve an efficient balance of the algorithm.

[0049] In the above scheme, the pheromone in step S46 is updated according to equations (6), (7) and (8):

[0050]

[0051]

[0052]

[0053] Among them, L best is the global optimal path length, L iter is the optimal path length of the current iteration, and are the pheromone increments released by the ants that find the global optimal path and the optimal path of the current iteration in this iteration, respectively, ij (t+1) represents the updated pheromone concentration. edge(i,j) is usually used to indicate whether a certain edge is used in the process of ants building solutions. Q is usually a positive constant used to adjust the absolute amount of pheromone released on the path. H(x) is the information entropy of this iteration. H t is the entropy-guiding factor of this iteration, τ ij and η ij are the pheromone concentration and heuristic information on the path respectively.

[0054] In the above scheme, step S47 calculates the entropy threshold according to formula (11):

[0055]

[0056] Among them, H threshod represents the entropy threshold of the current iteration, H max is the initial maximum entropy threshold, H min is the acceptable minimum entropy threshold, ω is the entropy decay factor, which is used to balance the termination timing and search range, t is the current iteration number, T max is the maximum number of iterations. As the number of iterations increases, the information entropy threshold decreases linearly to H min .

[0057] In the above solution, the iteration condition in step S47 is:

[0058]

[0059] H x (t) is the information entropy of this generation, H threshod It is a dynamic threshold based on information entropy, which is used to control the termination condition of the ant colony algorithm and terminate the algorithm in time when the algorithm finds the optimal solution.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] 1. The present invention guides ants to preferentially select nodes far away from obstacles through a non-uniform pheromone initialization method based on obstacle distance, thereby reducing ant mortality and improving the diversity and efficiency of initial path search.

[0062] 2. The present invention utilizes a dynamic pheromone update strategy based on information entropy and an adaptive pheromone evaporation method to flexibly control the search process, balance the convergence speed and exploration ability of the algorithm, ensure a higher quality global optimal path, and improve the global search ability and efficiency of the algorithm.

[0063] 3. The present invention introduces an iterative termination mechanism based on information entropy to avoid invalid iterations and waste of computing resources caused by a fixed number of iterations, ensuring that the algorithm continues to iterate when the optimal solution has not been reached, and ensuring that the optimal solution is found; it terminates in time when a satisfactory solution is reached, saving computing resources.

[0064] 4. The simulation results of the present invention show that the path planning algorithm based on the dynamic information entropy ant colony optimization algorithm (Dynamic Information Entropy-based Ant Colony Optimization, DIEACO), namely the DIEACO algorithm, has obvious advantages in the global optimal solution search capability and path planning efficiency compared with other algorithms on grid maps of different scales.

[0065] Note that the description of these effects does not prevent the existence of other effects. One embodiment of the present invention does not necessarily have all of the above effects. Effects other than the above can be clearly seen and extracted from the description of the specification, drawings, claims, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a schematic diagram of a two-dimensional rasterized model of mobile robot path planning according to one embodiment of the present invention.

[0067] Figure 2 It is a schematic diagram of a pheromone initialization distribution model according to one embodiment of the present invention.

[0068] Figure 3 It is a schematic diagram of the effect of information entropy on the variation range of pheromones according to one embodiment of the present invention.

[0069] Figure 4It is a schematic diagram of the DIEACO algorithm flow of the present invention in one embodiment of the present invention.

[0070] Figure 5 It is a schematic diagram of statistical results of different algorithms for a 20*20 grid map according to an embodiment of the present invention.

[0071] Figure 6 It is a schematic diagram of iterative convergence curves under comparison of different algorithms for a 20*20 grid map according to an embodiment of the present invention.

[0072] Figure 7 This is a schematic diagram of the optimal path under the comparison of different algorithms for a 30*30 grid map according to an embodiment of the present invention.

[0073] Figure 8 It is a schematic diagram of iterative convergence curves under comparison of different algorithms for a 30*30 grid map according to an embodiment of the present invention.

[0074] Fig. 9 It is a schematic diagram of the optimal path under the comparison of different algorithms of a 50*50 grid map according to an embodiment of the present invention.

[0075] Fig.10 It is a schematic diagram of iterative convergence curves under comparison of different algorithms for a 50*50 grid map according to an embodiment of the present invention. DETAILED DESCRIPTION

[0076] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0077] A mobile robot path planning method based on an ant colony optimization algorithm with dynamic information entropy comprises the following steps:

[0078] S1. Environment modeling: Use the grid method to model the environment of the mobile robot's motion space, such as Figure 1 As shown, the environment is divided into a G×G grid, where each grid represents a node. By reading the sensor data, the positions of the starting point, end point, and obstacles are determined and recorded in the grid coordinates. Assume that the map size is 20×20, the starting point is set to (0,20), the end point is set to (20,0), and the obstacle distribution is randomly generated to ensure that there are multiple possible paths;

[0079] S2. Set the initial parameters of the ant colony algorithm: the maximum number of iterations Tmax is set to 100, the number of ants is set to 100, the heuristic factor η is set to 1, the pheromone factor α is set to 1, and the initial pheromone evaporation coefficient ρ is set to 1. 0Set to 0.1, the initial maximum information entropy threshold H max Set to 10, the minimum information entropy threshold H min is set to 0.1, and the entropy decay factor ω is set to 0.05;

[0080] S3. Non-uniform pheromone initialization: For each non-obstacle node (i, j), calculate its Euclidean distance minDist to the nearest obstacle, and initialize the pheromone concentration τ through formula (1) ij ,like Figure 2 As shown;

[0081]

[0082] Among them, maxPheromone represents the maximum pheromone concentration of the entire map, and this hyperparameter is set to 1. obstacleMap(i,j) indicates whether there is an obstacle at the position (i,j) on the map. If there is an obstacle, the pheromone concentration is 0. minDist is the Euclidean distance from the current (i,j) node to its nearest obstacle, and σ is the standard deviation based on the distance.

[0083] S4. Path construction:

[0084] S41. Set the starting point of the robot as the initial node of the ant and add it to the taboo list. All ants move from the current node i to the next unvisited node j at each time step. The ants select the next node according to the path selection probability. The path selection probability is calculated by formula (2) and formula (3):

[0085]

[0086]

[0087] in, represents the probability of ant k moving from node i to node j;

[0088] d ij Represents the Euclidean distance between the current node i and the candidate node j;

[0089] τ ij and η ij are the pheromone concentration and heuristic information on the path respectively;

[0090] α and β are two weight parameters, which measure the influence of pheromone concentration and heuristic information on path selection, respectively, and can be given;

[0091] allowed k represents the set of nodes that ant k has not visited yet;

[0092] s is the node that the ant is about to go to;

[0093] τ is and η is They represent the pheromone concentration and heuristic information on the path from node i to node s respectively;

[0094] S42, using the roulette algorithm to select the next node based on the path selection probability, and adding the current node to the taboo table; repeating S41 and S42 until all ants reach the end point;

[0095] S43, information entropy calculation, calculating the path information entropy of this iteration according to formula (4);

[0096]

[0097] Among them, H(X) is the path information entropy, X represents the set of paths, including all possible paths, and p(xi) represents the specific path x i The probability of being selected, occurrence(x i ) is the path x in the group i The number of times selected is the historical selection frequency of the ant on the path, and n is the number of paths generated in this iteration. This formula is used to measure the uncertainty or information content of the random variable, reflecting the degree of exploration of the path by the entire ant colony in the search space, and quickly locates the optimal solution;

[0098] S44, calculating the entropy guidance factor and entropy difference of this iteration respectively according to formula (5) and formula (9);

[0099] H t =min{L iter (t-1),L iter (t-2),…,L iter (1)}-L best (t) (5)

[0100] ΔH=H x (t)-H x (t-1) (9)

[0101] Among them, H t represents the entropy guiding factor of the tth iteration, L best is the global optimal path so far, L iter (t-1) represents the best path of t-1 iteration, L iter (t-2) represents the best path of t-2 iterations..., L iter (1) represents the best path of the first iteration, H x (t) is the information entropy of this generation, H x (t-1) is the information entropy of the previous generation, ΔH is the difference between the information entropy of this generation and the previous generation, that is, the entropy difference;

[0102] S45. Calculate the pheromone evaporation coefficient according to formula (10). Formula (10) is an adaptive pheromone evaporation method, which can balance the convergence speed and exploration ability of the algorithm, and can achieve an efficient balance of the algorithm, while quickly locating potential excellent solutions, avoiding ignoring other possible better solutions;

[0103] ρ=ρ 0 (1+μΔH) (10)

[0104] Where ρ is the evaporation coefficient of pheromone, ρ 0 is the initialization evaporation coefficient, μ is the evaporation influence factor, and this formula is used to balance the convergence speed and exploration ability of the algorithm to achieve an efficient balance of the algorithm.

[0105] S46, dynamic pheromone update strategy, according to the information entropy H(x) of this iteration and the entropy guidance factor H t The pheromone update method is determined by comparing the current information entropy with the entropy threshold. If the current information entropy is higher than the entropy threshold, the pheromone update amplitude is reduced to accelerate convergence. On the contrary, if the information entropy is lower than the entropy threshold, the update amplitude is increased to enhance the search capability. According to equations (6), (7), and (8), the pheromone update method is dynamically selected and the pheromone concentration of each node is updated, such as Figure 3 shown.

[0106]

[0107]

[0108] Among them, L best is the global optimal path length, L iter is the optimal path length of the current iteration, and are the pheromone increments released by the ants that find the global optimal path and the optimal path of the current iteration in this iteration, respectively, ij (t+1) represents the updated pheromone concentration. edge(i,j) is usually used to indicate whether a certain edge is used in the process of ants building solutions. Q is usually a positive constant used to adjust the absolute amount of pheromone released on the path. H(x) is the information entropy of this iteration. H t is the entropy-guiding factor of this iteration, τ ij and η ij are the pheromone concentration and heuristic information on the path respectively.

[0109] S47, iterative termination mechanism. When the information entropy of the algorithm is lower than a certain threshold, it indicates that a sufficiently good solution has been found or a better solution cannot be found, and the algorithm iteration can be stopped intelligently. The entropy threshold is calculated using equations (11) and (12), and it is checked whether the information entropy of the algorithm in this iteration is greater than the entropy threshold. If the condition is not met, it returns to S41 and starts the iteration again from the starting point. If the condition is met, the optimal solution found is output, that is, the shortest path currently found, and the algorithm ends.

[0110]

[0111]

[0112] Among them, H threshod represents the entropy threshold of this iteration, H max is the initial maximum threshold, H min is the minimum acceptable information entropy, ω is the entropy decay factor, which is used to balance the termination timing and search range, t is the current number of iterations, T max is the maximum number of iterations. As the number of iterations increases, the information entropy threshold decreases linearly to H min .

[0113] In a specific embodiment of the present invention, Figure 4 As shown, the path planning algorithm of the present invention is based on the dynamic information entropy ant colony optimization algorithm, that is, the DIEACO algorithm process steps are as follows:

[0114] Use grid method for environmental modeling;

[0115] Initialize the grid map and set the initial parameters of the ant colony algorithm;

[0116] Non-uniform initialization pheromone;

[0117] Start path construction, for each ant k:

[0118] Determine whether k ≤ M, where M represents the total number of ants exploring;

[0119] If k≤M, find all non-obstruction grid points and calculate the transfer probability of each grid. Use the roulette strategy to determine the next feasible grid for the ant, update the taboo table, and determine whether the ant is in a deadlock. If so, return to update the taboo table. If not, determine whether ant k has reached the end point. If ant k has reached the end point, record the path and length of the ant. When k=k+1, return to determine whether k≤M. If ant k has not reached the end point, continue to find all non-obstruction grid points and calculate the transfer probability of each grid.

[0120] If k>M, it means that all ants in this round have completed the path construction. Then calculate the information entropy of this iteration, calculate the pheromone evaporation coefficient of this iteration, and judge H(X)>H t If yes, then use iterative optimal update pheromone, if not, then use global optimal update pheromone, judge H x (t)≤H threshold (t), if not, end the calculation, if yes, return to ant k = 1 and proceed to the next path construction.

[0121] like Figure 5-10 As shown in the figure, the DIEACO algorithm is compared with other traditional algorithms on grid maps of different sizes (20×20, 30×30, 50×50), and the optimal path and convergence curve are recorded to evaluate the performance of the DIEACO algorithm in terms of global optimal solution search capability and path planning efficiency.

[0122] like Figure 5 As shown in the figure, in the 20×20 map, the obstacles are relatively sparsely distributed. The path generated by the DIEACO algorithm shows high flexibility and can effectively avoid the obstacle area in the middle and directly connect the starting point and the end point. In contrast, the green dashed AS algorithm path is relatively lengthy when selected and is easily affected by the distribution of obstacles, resulting in a tortuous path. In addition, the paths generated by other algorithms (such as EAS, RAS, ACS, etc.) are relatively scattered, while the path of the DIEACO algorithm is closer to the shortest path while avoiding obstacles, showing its efficiency in finding a better path on a small-scale grid map.

[0123] like Figure 6 As shown in Table 1, on the 20×20 grid map, the DIEACO algorithm of the present invention found the optimal value of 34.485, and achieved the optimal solution each time in 10 experiments with a standard deviation of 0, which is significantly better than the 7 optimal solutions of AS and 1 of RAS and EAS, and its stability is much higher than that of RAS (standard deviation 1.1446).

[0124] Table 1 Statistical results of different algorithms for 20*20 grid maps

[0125]

[0126] Note: AS represents the optimal path generated by the AS algorithm, EAS represents the optimal path generated by the EAS algorithm, RAS represents the optimal path generated by the RAS algorithm, ACS represents the optimal path generated by the ACS algorithm, MAACO represents the optimal path generated by the MAACO algorithm, and DIEACO represents the optimal path generated by the DIEACO algorithm.

[0127] like Figure 7As shown in the figure, on a 30×30 map, the distribution of obstacles presents a complex layout that is staggered vertically and horizontally, which increases the difficulty of path planning. The advantage of the DIEACO algorithm here is clearly reflected in its dynamic adjustment of pheromones and adaptive evaporation methods, which can maintain global search capabilities in complex paths. In contrast, the paths of the EAS and ACS algorithms have more detours, showing the local optimal problem of path selection, while DIEACO can find an efficient and smoother path more quickly through a global update strategy, avoiding too many detours and complex intersections. This shows that the search performance of the DIEACO algorithm is more outstanding in medium-sized grid maps.

[0128] like Figure 8 As shown in Table 2, on a 30×30 grid map, the DIEACO of the present invention achieved an optimal value of 51.456, and also maintained a stable performance of 10 optimal values. In contrast, AS, EAS, and ACS only found 1-2 optimal solutions. In addition, the standard deviation of DIEACO is still 0, which is much lower than other algorithms, such as ACS's 4.0217. Although the convergence time of DIEACO is 35 rounds, which is slightly higher than some algorithms, it terminates early at 44 iterations, reducing the computational cost.

[0129] Table 2 Statistical results of different algorithms for 30*30 grid maps

[0130]

[0131] like Fig. 9 As shown in the figure, in a large-scale 50×50 grid map, obstacles are densely distributed and randomly distributed, and path planning is the most difficult. The DIEACO algorithm performs well in this environment. Through the dynamic adjustment mechanism of information entropy, it effectively maintains the diversity of the population, making it less likely to fall into local optimality on complex maps. As can be seen in the figure, the red path generated by DIEACO can effectively find a path that is as direct as possible and avoids obstacles among dense obstacles, while other algorithms (such as AS and RAS) find it difficult to achieve the same effect due to long paths or multiple detours. DIEACO's dynamic pheromone update and evaporation mechanism performs particularly well in large-scale maps, giving it stronger adaptability and global search capabilities in complex maps. Fig.10 As shown in Table 3, on a 50×50 grid map, both DIEACO and MAACO found the optimal value of 73.983, but DIEACO stably reached the optimal value in 10 experiments, while MAACO only reached the optimal value in 8 experiments. In terms of standard deviation, DIEACO is still 0, while other algorithms are generally between 0.8-1.1. Although the convergence time is 36 rounds, DIEACO ends early with 58 iterations, which is better than MAACO's 44 rounds of convergence time and 100 iterations.

[0132] Table 3 Statistical results of different algorithms for 50*50 grid maps

[0133]

[0134] The above data show that DIEACO has significant advantages in global search capability, stability, and computational efficiency on maps of all sizes, especially on larger 30×30 and 50×50 maps.

[0135] The mobile robot path planning method based on the dynamic information entropy ant colony optimization algorithm of the present invention is a method for non-uniform initialization of pheromone distribution based on obstacle distance. By applying pheromones of different concentrations on the grid, ants are guided to preferentially choose grids away from obstacles, thereby reducing ant mortality, improving the diversity of initial path search, and optimizing the search direction. At the same time, the uncertainty and randomness of information entropy are used to dynamically adjust the pheromone update strategy and the adaptive pheromone evaporation method to flexibly control the search process, ensure that a higher quality global path is obtained, and improve the algorithm efficiency and global search capability. In addition, in order to solve the problem that the optimal solution cannot be found with a fixed number of iterations or the running time is too long due to excessive iterations, the present invention introduces an entropy threshold as an intelligent stopping criterion to ensure that the algorithm continues to iterate when the optimal solution is not reached, ensure that the optimal solution is found, and terminate in time when a satisfactory solution is reached, thereby avoiding unnecessary computational overhead.

[0136] It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0137] The series of detailed descriptions listed above are only specific descriptions of feasible embodiments of the present invention. They are not intended to limit the scope of protection of the present invention. All equivalent embodiments or changes that do not deviate from the technical spirit of the present invention should be included in the scope of protection of the present invention.

Claims

1. A mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm, characterized in that: The following steps are involved: S1. Environment modeling: setting the robot’s starting point, end point, and obstacle locations; S2. Set the initial parameters of the ant algorithm: set the maximum number of iterations T max , the number of ants k, the heuristic factor η, the pheromone factor α, the initial pheromone evaporation coefficient ρ0, the initial maximum information entropy threshold H max , minimum information entropy threshold H min and entropy decay factor ω; S3, non-uniform pheromone initialization: According to the obstacle distance, the pheromone is initialized non-uniformly. For each non-obstacle node (i, j), the Euclidean distance minDist to the nearest obstacle is calculated, and the pheromone concentration τ is calculated and initialized. ij ; S4. Path construction: S41, setting the starting point of the robot as the initial node of the ant, adding it to the taboo list, and calculating the path selection probability between the current node i and the candidate node j; S42, using the roulette algorithm to select the next node based on the path selection probability, and adding the current node to the taboo list, repeating S41 and S42 until all ants reach the end point; S43, calculating the information entropy of this iteration; S44. Calculate the entropy guidance factor and entropy difference of this iteration: S45, adaptive adjustment of pheromone evaporation coefficient; S46, dynamic pheromone update: determine the pheromone update method based on the comparison between the information entropy of this iteration and the entropy guidance factor; S47, iteration termination mechanism: calculate the entropy threshold, check whether the information entropy of this iteration of the algorithm is greater than the entropy threshold. If the condition is not met, return to S41 and start the iteration again from the starting point. If the condition is met, output the optimal solution found and the algorithm ends.

2. The mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm according to claim 1 is characterized in that: The pheromone concentration τ in S3 ij Calculated by formula (1): Among them, maxPheromone represents the maximum pheromone concentration of the entire map, which is set to 1; obstacleMap(i,j) indicates whether there is an obstacle at the position (i,j) on the map. If there is an obstacle, the pheromone concentration is 0; minDist is the Euclidean distance from the current (i, j) node to its nearest obstacle; σ is the standard deviation based on the distance.

3. The mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm according to claim 2 is characterized in that: The S41 path selection probability is based on equation (2) and equation (3): Among them, p ij k (t) represents the probability of ant k moving from node i to node j; d ij Represents the Euclidean distance between the current node i and the candidate node j; τ ij and η ij are the pheromone concentration and heuristic information on the path respectively; α and β are two weight parameters, which respectively measure the influence of pheromone concentration and heuristic information on path selection; allowed k represents the set of nodes that ant k has not visited yet; s is the node that the ant is about to go to; τ is and η is They represent the pheromone concentration and heuristic information on the path from node i to node s respectively.

4. The mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm according to claim 1 is characterized in that: The step S43 calculates the information entropy of this iteration according to formula (4): Among them, H(X) is the path information entropy, X represents the set of paths, including all possible paths, and p(xi) represents the specific path x i The probability of being selected, occurrence(x i ) is the path x in the group i The number of times selected is the historical selection frequency of the ant on the path, and n is the number of paths generated in this iteration.

5. The mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm according to claim 1 is characterized in that: The step S44 calculates the entropy guidance factor of this iteration according to formula (5): H t =min{L iter (t-1),L iter (t-2),…,L iter (1)}-L best (t) (5) Among them, H t represents the entropy guiding factor of the tth iteration, L best is the global optimal path so far, L iter (t-1) represents the best path of t-1 iteration, L iter (t-2) represents the best path of t-2 iterations..., L iter (1) represents the best path of the first iteration.

6. The mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm according to claim 1 is characterized in that: The step S44 calculates the entropy difference of this iteration according to formula (9): ΔH=H x (t)-H x (t-1) (9) H x (t) is the information entropy of this generation, H x (t-1) is the information entropy of the previous generation, and ΔH is the difference between the information entropy of this generation and the previous generation, that is, the entropy difference.

7. The mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm according to claim 6 is characterized in that: The step S45 calculates the pheromone evaporation coefficient according to formula (10): ρ=ρ0(1+μΔH) (10) Among them, ρ is the evaporation coefficient of pheromone, ρ0 is the initialization evaporation coefficient, and μ is the evaporation influence factor.

8. The mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm according to claim 7 is characterized in that: In step S46, the pheromone is updated according to equations (6), (7), and (8): Among them, L best is the global optimal path length, L iter is the optimal path length of the current iteration, Δτ ij gbest and Δτ ij iter are the pheromone increments released by the ants that find the global optimal path and the optimal path of the current iteration in this iteration, respectively, ij (t+1) represents the updated pheromone concentration, edge(i,j) represents whether a certain edge is used in the process of ants building solutions, Q is a positive constant used to adjust the absolute amount of pheromone released on the path, H(x) is the information entropy of this iteration, H t is the entropy-guiding factor of this iteration, τ ij and η ij are the pheromone concentration and heuristic information on the path respectively.

9. The mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm according to claim 1 is characterized in that: The step S47 calculates the entropy threshold according to formula (11): Among them, H threshod represents the entropy threshold of the current iteration, H max is the initial maximum entropy threshold, H min is the minimum entropy threshold, ω is the entropy decay factor, which is used to balance the termination timing and search range, t is the current number of iterations, T max is the maximum number of iterations.

10. The mobile robot path planning method based on dynamic information entropy ant colony optimization algorithm according to claim 1 is characterized in that: The iteration condition in step S47 is: H x (t) is the information entropy of this generation, H threshod It is a dynamic threshold based on information entropy, which is used to control the termination condition of the ant colony algorithm.

Images