Intelligent robot path planning and optimizing method based on fusion ant colony algorithm

Abstract

The invention relates to an intelligent robot path planning and optimizing method and device based on a fusion ant colony algorithm, equipment and a medium. The method comprises the following steps: acquiring a global grid map and a dominance evaluation function, and initializing an artificial potential field and an ant colony; placing the ant colony at the starting point grid for searching to obtain an ant colony path space after each search; if the ant colony path space does not meet the iteration ending condition, re-placing the ant colony at the starting point grid for searching until the ant colony path space meets the iteration ending condition; and if the ant colony path space meets an iteration ending condition, generating a global optimal path based on an optimal collision-free path in the ant colony path space. By the adoption of the method, global optimal target guidance of the ant colony algorithm can be determined through the advantage evaluation function, and the search efficiency is improved based on the improved heuristic function combining the iteration parameters and the artificial potential field.

Description

Technical Field

[0001] The present application belongs to the field of path planning, and in particular, relates to an intelligent robot path planning and optimization method based on the fusion of ant colony algorithm. Background Art

[0002] With the development of intelligent robot technology, path planning technology has emerged. Path planning technology is an important part of the research field of mobile intelligent robots. It can find an optimal collision-free path from the starting position node to the target position node for the intelligent robot in an environment with obstacles according to a given goal.

[0003] As a probabilistic selection algorithm based on population, the ant colony algorithm has been widely used in the field of path planning because it has strong robustness and better solution search ability compared with other heuristic algorithms, and it is easy to combine with other algorithms to improve its performance.

[0004] However, the traditional ant colony algorithm has problems such as search stagnation, local convergence, slow convergence speed, slow path generation speed, etc. when performing path planning. In addition, the traditional ant colony algorithm is blind in the early stage of search, and the pheromone concentration it produces will be disturbed by redundant paths. The global search ability is not strong, and it is easy to fall into the local optimum, which in turn affects the accuracy of ants' selection. Summary of the invention

[0005] Based on this, it is necessary to provide an intelligent robot path planning and optimization method, device, computer equipment and storage medium based on the fusion ant colony algorithm, which can avoid the stagnation of ant colony algorithm search, accelerate the convergence speed of ant colony algorithm and improve the path planning ability of ant colony algorithm in response to the above technical problems.

[0006] The present application provides an intelligent robot path planning and optimization method based on fusion ant colony algorithm, including:

[0007] Obtain the global grid map and advantage evaluation function, and initialize the artificial potential field and ant colony. The advantage evaluation function is used to characterize the distance advantage, obstacle avoidance advantage and steering advantage of the path corresponding to the ant colony.

[0008] The ant colony is placed at the starting grid for searching, and the ant colony path space after each search is obtained. The ant colony path space is used to represent the path taken by the ant colony during the search process;

[0009] If the ant colony path space does not meet the iteration end condition, the ant colony is re-placed at the starting grid to search until the ant colony path space meets the iteration end condition; the iteration end condition is used to characterize that the ant colony path space includes the optimal collision-free path from the starting grid to the end grid, and the optimal collision-free path is the path whose advantage value meets the global optimal condition;

[0010] If the ant colony path space meets the iteration end condition, a global optimal path is generated based on the optimal collision-free path in the ant colony path space;

[0011] The ant colony is placed at the starting grid for searching, including the steps of:

[0012] The heuristic information is calculated using an improved heuristic function combining iteration parameters and artificial potential fields;

[0013] Obtain pheromone information, and input the pheromone information and heuristic information into the improved state transition probability model of each ant, select the next path grid from the territory grid of the path grid where the ant is located, and update the heuristic information;

[0014] If the single search end condition is met, the ant's path is input into the advantage evaluation function to calculate the advantage value, and the parameters and pheromone information of the ant colony are updated.

[0015] The above-mentioned intelligent robot path planning and optimization method, device, computer equipment and storage medium based on the fusion ant colony algorithm use the advantage evaluation function to characterize the optimal collision-free path from the starting grid to the end grid, so that the search of the ant colony has a clear goal orientation pointing to the global optimal. It avoids the situation of falling into local optimality and search stagnation due to excessive accumulation of pheromones in local areas, enhances the global search ability, helps to jump out of the dilemma of local convergence, and is more likely to find the real global optimal path. In addition, the advantage evaluation function is used to characterize the distance advantage, obstacle avoidance advantage and steering advantage of the path corresponding to the ant colony, and can comprehensively weigh multiple key factors such as the length of the path, whether it can effectively avoid obstacles and the convenience of steering operations to judge the quality of the path. It enhances the practicality and rationality of the ant colony algorithm and avoids the situation of generating paths that are difficult to actually execute by focusing only on the distance.

[0016] Furthermore, by combining the improved heuristic function of iterative parameters and artificial potential fields, the ant colony is provided with prior guidance information, which no longer relies too much on pheromone concentration for path selection, so that the ants are no longer blind in the early stage of the search, reduce the exploration of redundant paths, and thus accelerate the convergence speed and improve the efficiency of path generation. The combination of the ant colony algorithm and the artificial potential field fully utilizes the advantages of the ant colony algorithm itself in terms of strong robustness and good solution search ability in terms of solution ability, while taking advantage of the advantages of the artificial potential field in guiding the search direction and avoiding blind exploration, thereby improving the overall performance of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A schematic diagram of an implementation environment of an intelligent robot path planning and optimization method based on an ant colony algorithm provided in one embodiment of the present application;

[0019] Figure 2 A schematic diagram of search results of a traditional ant colony algorithm in the prior art;

[0020] Figure 3 A flowchart of an intelligent robot path planning and optimization method based on fusion ant colony algorithm provided in one embodiment of the present application;

[0021] Figure 4 A flowchart of a method for updating pheromone information provided by an embodiment of the present application;

[0022] Figure 5 A flowchart of a method for updating parameters and pheromone information of an ant colony provided by an embodiment of the present application;

[0023] Figure 6 A flowchart of a method for obtaining an advantage evaluation function provided by an embodiment of the present application;

[0024] Figure 7 A schematic diagram of search results of an intelligent robot path planning and optimization method based on fusion ant colony algorithm provided by one embodiment of the present application;

[0025] Figure 8 A schematic diagram of search results of another intelligent robot path planning and optimization method based on fusion ant colony algorithm provided by an embodiment of the present application;

[0026] Fig. 9 A schematic diagram of the structure of an intelligent robot path planning and optimization device based on the fusion of ant colony algorithm provided in one embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0028] The intelligent robot path planning and optimization method based on the fusion ant colony algorithm provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the intelligent robot execution subject 102 communicates with the computing control platform 101 through a communication channel. Among them, the intelligent robot execution subject 102 can be but not limited to various industrial automation robots, cleaning robots, nursing robots, logistics robots, agricultural robots and private domain robots, and private domain robots can be but not limited to security monitoring robots, entertainment companion robots and personal assistant robots. The computing control platform 101 can be but not limited to embedded control systems, smart terminals, servers, edge computing platforms and cloud computing platforms. Smart terminals can be but not limited to various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc.

[0029] For example, the computing control platform 101 can determine the moving target of the intelligent robot execution body 102 and construct an environment map of the intelligent robot execution body 102, and then generate the following map by using the ant colony algorithm: Figure 2 The search results shown.

[0030] Further, the computing control platform 101 can be based on Figure 2 The search results shown continuously adjust the control strategy according to the environmental feedback to control the intelligent robot execution body 102 to reach the moving target.

[0031] In an exemplary embodiment, Figure 3 As shown in the figure, a method for intelligent robot path planning and optimization based on ant colony algorithm is provided. Figure 1 The computing control platform 101 in FIG. 1 is used as an example to illustrate the invention, including:

[0032] Step S3010, obtain the global grid map and advantage evaluation function, and initialize the artificial potential field and ant colony.

[0033] Specifically, the computing control platform 101 can obtain a global grid map representing the working environment and mobile target of the intelligent robot execution subject and an advantage evaluation function representing the distance advantage, obstacle avoidance advantage and steering advantage of the path corresponding to the ant colony. The artificial potential field is initialized by setting the attraction coefficient, repulsion coefficient and distance threshold of the repulsion effect; and the ant colony is initialized by setting the number of ants, the maximum number of iterations and the initial value of pheromone.

[0034] Exemplarily, a feasible expression of the advantage evaluation function is as follows:

[0035] L(S k,t )=λ S ×L S (S k，t )+λ Z ×L Z (S k,t )+λ P ×L P (S k，t )

[0036] In the formula, S k,t is the path of the kth ant during the tth search, L(S k,t ) is the advantage evaluation function, L S (S k,t ) is the distance advantage term, L Z (S k,t ) is the obstacle advantage term, L P (S k,t ) is the steering advantage term, λ S is the distance advantage coefficient, λ Z is the obstacle advantage coefficient, λ P is the steering advantage coefficient.

[0037] Furthermore, the feasible expressions of the distance advantage term, obstacle advantage term and turning advantage term are as follows:

[0038]

[0039]

[0040]

[0041] In the formula, is the number of grids that the k-th ant passes through during the t-th search, Λ i is the i-th grid in the path of the k-th ant during the t-th search, Λ i,Z is the obstacle grid closest to the i-th grid in the path of the k-th ant during the t-th search, and is grid Λ i and grid Λ i,ZThe distance between is the vector from the i-th grid in the path of the k-th ant during the t-th search to the i+1-th grid in the path, cos -1 is the inverse cosine function.

[0042] Step S3020, placing the ant colony at the starting grid to search, and obtaining the ant colony path space after each search.

[0043] Specifically, the computing control platform 101 can place the ant colony at the starting grid for searching, and obtain the ant colony path space after each search. The ant colony path space is used to characterize the path taken by the ant colony during the search process. The ant colony path space can include the path obtained from each search and the path obtained in the past search, as well as the advantage value of the path obtained based on the advantage evaluation function.

[0044] Step S3030: If the ant colony path space does not meet the iteration end condition, the ant colony is re-placed at the starting grid to search until the ant colony path space meets the iteration end condition.

[0045] Specifically, if the computing control platform 101 determines that the ant colony path space does not meet the iteration end condition, the computing control platform 101 can re-place the ant colony at the starting grid for searching until it is determined that the ant colony path space meets the iteration end condition. The iteration end condition is used to characterize that the ant colony path space includes the optimal collision-free path from the starting grid to the end grid, and the optimal collision-free path is a path whose advantage value meets the global optimal condition.

[0046] Illustratively, the iteration end condition may be that the maximum value of the advantage value of the path in the ant colony path space meets the iteration end boundary threshold condition or the number of iterations of the path in the ant colony path space reaches a maximum number of iterations.

[0047] Furthermore, if the distance value of the path is positively correlated with the distance advantage term, the global optimal condition may be the minimum value in the ant colony path space, and the iteration end boundary threshold condition may be that the minimum value of the path advantage value is less than the iteration end boundary threshold. If the distance value of the path is negatively correlated with the distance advantage term, the global optimal condition may be the maximum value in the ant colony path space, and the iteration end boundary threshold condition may be that the maximum value of the path advantage value is greater than the iteration end boundary threshold.

[0048] Step S3040: if the ant colony path space meets the iteration end condition, a global optimal path is generated based on the optimal collision-free path in the ant colony path space.

[0049] Specifically, if the computing control platform 101 determines that the ant colony path space meets the iteration end condition, the computing control platform 101 can optimize the optimal collision-free path in the ant colony path space to generate a global optimal path. The optimization process can be, but is not limited to, path smoothing, path simplification, and dynamic obstacle adaptive optimization.

[0050] Exemplarily, the computing control platform 101 can perform path smoothing processing on the optimal collision-free path in the ant colony path space based on the uniform B-spline curve to generate a global optimal path. The feasible formula of the N-fold uniform B-spline curve is as follows:

[0051]

[0052]

[0053] In the formula, is the motion control point of the Bezier curve, P i is the location node, is the B-spline basis function.

[0054] In step S3020, the ant colony is placed at the starting grid for searching, including:

[0055] Step S3021, calculating heuristic information using an improved heuristic function combining iteration parameters and artificial potential field.

[0056] Specifically, the computing control platform 101 may use an improved heuristic function combining iterative parameters and an artificial potential field to calculate the heuristic information. A feasible expression of the improved heuristic function is as follows:

[0057]

[0058]

[0059]

[0060] In the formula, To improve the heuristic function, is the artificial potential field term, E2(N iter ) is the iteration parameter item, is the distance between the path grid with horizontal coordinate i and vertical coordinate j and the end grid, A is the distance coefficient, is the resultant force received by the intelligent robot in the artificial potential field when it is located on the path grid with horizontal coordinate i and vertical coordinate j, is the vector from the path grid with abscissa i and ordinate j to the destination grid, γ is the coefficient of the artificial potential field term, N iter is the current iteration number, N max is the maximum number of iterations.

[0061] Furthermore, the heuristic function can be optimized based on the distance between the starting grid and the end grid. The feasible expression of the optimized improved heuristic function is as follows:

[0062]

[0063] Where A' is the distance coefficient of the optimized improved heuristic function, is the distance between the starting grid and the ending grid, is the distance between the starting grid and the path grid with horizontal coordinate i and vertical coordinate j.

[0064] Illustratively, the computing control platform 101 may calculate the heuristic information based on the improved heuristic function of path advantage evaluation entropy enhancement. The feasible expression of the improved heuristic function based on path advantage evaluation entropy enhancement is as follows:

[0065]

[0066]

[0067]

[0068] In the formula, E3[H(S t-1 )] is the path advantage evaluation entropy term, C S is the entropy coefficient of path advantage evaluation, H(S t ) is the path advantage evaluation entropy of the path of the tth iteration, L(S k,t ) is the advantage evaluation function.

[0069] Step S3022, obtaining pheromone information, and inputting the pheromone information and the heuristic information into the improved state transition probability model of each ant, selecting the next path grid from the territory grid of the path grid where the ant is located, and updating the heuristic information.

[0070] Schematically, the feasible expression of the improved state transition probability model is as follows:

[0071]

[0072]

[0073] In the formula, in the formula, Λ n+1 (t) is the next path grid of the ant selected by the improved state transition probability model at the tth iteration, ARG MAX is the maximum point function, Λ∈allowed represents the set of grids that the ant is allowed to visit next, α is the pheromone factor, β is the heuristic information factor, q is a random number, q0 is the roulette threshold, Λ rouletteis the path grid selected by roulette, is the probability of selecting the path grid by roulette at the tth iteration.

[0074] Step S3023: If the single search end condition is met, the ant path is input into the advantage evaluation function to calculate the advantage value, and the pheromone information and the parameters of the ant colony are updated.

[0075] In the above-mentioned intelligent robot path planning and optimization method based on the fusion ant colony algorithm, the advantage evaluation function can provide a clear and globally optimal goal orientation for the search of the ant colony. By weighing the length of the path, whether it can effectively avoid obstacles, and the convenience of steering operation, the quality of the path is judged, which avoids the situation where the traditional algorithm only focuses on the distance and generates a path that is difficult to actually execute, and enhances the practicality and rationality of the ant colony algorithm in practical applications.

[0076] Furthermore, in the above-mentioned intelligent robot path planning and optimization method based on the fusion ant colony algorithm, by using the improved heuristic function combining iterative parameters and artificial potential field to calculate the heuristic information, it can provide a priori guidance information for the ant colony, thereby effectively accelerating the convergence speed and improving the path generation efficiency. By integrating the ant colony algorithm and the artificial potential field, it is effective to avoid falling into the local optimum and the search stagnation caused by the excessive accumulation of pheromones in the local area, and greatly enhance the global search capability.

[0077] In an alternative embodiment, please refer to Figure 4 , updated pheromone information, including:

[0078] Step S401, based on the proportion of path grids whose repulsive force values ​​of the artificial potential field in the neighborhood grid are lower than the obstacle risk threshold, a penalty mechanism of the dynamic weighted pheromone sputtering global update strategy with a penalty mechanism is set.

[0079] Step S402, setting the dynamic weight of the dynamic weight pheromone sputtering global update strategy with a penalty mechanism based on the path advantage evaluation entropy.

[0080] Specifically, the dynamic weights include pheromone volatilization dynamic weights, penalty dynamic weights, and pheromone splashing dynamic weights.

[0081] Step S403, globally updating the pheromone information based on the dynamic weight pheromone sputtering global update strategy with a penalty mechanism.

[0082] Specifically, the expression of the dynamic weight pheromone splash update strategy based on the penalty mechanism is as follows:

[0083]

[0084]

[0085]

[0086]

[0087] In the formula, For the updated pheromone information, This is the pheromone information before the update. is the pheromone information increment after splashing, is the pheromone information increment before splashing, Λ i,j is a grid with horizontal coordinate i and vertical coordinate j, ρ[H(S t )] is the dynamic weight of pheromone volatilization, H(S t ) is the path advantage evaluation entropy of the t-th search, S is the path of the ant, is the penalty dynamic weight, N acc is the number of path grids in the neighborhood grid whose repulsive force value of the artificial potential field is lower than the obstacle risk threshold, N nbh is the number of path grids in the neighborhood grid, N iter is the number of iterations, is the pheromone information increment before the rth neighborhood path grid is splashed, σ[H(S t )] is the dynamic weight of pheromone splash, is the second norm of the vector from the grid with horizontal coordinate i and vertical coordinate j to its rth neighboring path grid, Q is the pheromone increment coefficient, m is the number of ants, L(S k,t ) is the advantage evaluation item of the k-th ant at the tth iteration.

[0088] Furthermore, feasible expressions of pheromone volatilization dynamic weight and penalty dynamic weight are as follows:

[0089]

[0090]

[0091] In the formula, ρ min is the minimum value of the dynamic weight of pheromone volatilization, R is the initial coefficient of the dynamic weight of pheromone volatilization, W is the path advantage evaluation entropy coefficient of the dynamic weight of pheromone volatilization, λ c is the initial coefficient of the penalty dynamic weight, K c Evaluate the entropy coefficient for path advantage in penalizing dynamic weights.

[0092] For example, in an ant colony system, in addition to global pheromone updates, local pheromone updates are also required. When an ant selects a new node, the feasible expression for local pheromone updates is as follows:

[0093]

[0094] In the formula, is the pheromone information at the tth iteration after the grid with horizontal coordinate i and vertical coordinate j is updated. is the pheromone information of the grid with abscissa i and ordinate j at the tth iteration before the local update, ζ is the local pheromone volatility coefficient, is the initial value of the pheromone of the grid with horizontal coordinate i and vertical coordinate j.

[0095] In the above-mentioned intelligent robot path planning and optimization method based on the fusion ant colony algorithm, the penalty mechanism and the dynamic weight pheromone sputtering global update strategy set based on the path advantage evaluation entropy can improve the safety and rationality of path planning, improve the efficiency of searching for the global optimal path, and accelerate the convergence speed, thereby improving the overall performance of the ant colony algorithm in complex path planning scenarios.

[0096] In one embodiment, please refer to Figure 5 , update the parameters and pheromone information of the ant colony, including:

[0097] Step S501, obtaining the local best ant, the local worst ant and the global best ant based on the advantage values ​​of the ant paths.

[0098] Step S502: mutate the parameters of the improved state transition probability model of the local optimal ant and update them into the improved state transition probability model of the local worst ant.

[0099] Specifically, the parameters include heuristic information factor, pheromone factor and roulette threshold.

[0100] Step S503, after the path advantage evaluation entropy meets the path geometry pruning condition, a path geometry pruning operation is performed on the path of the global optimal ant to generate a global optimization path.

[0101] Specifically, the path geometry pruning operation is used to connect path nodes in the same row, column, or diagonal line.

[0102] Step S504: Generate pheromone enhancement information on the global optimization path, and update the pheromone information increment before sputtering based on the pheromone enhancement information.

[0103] In the above-mentioned intelligent robot path planning and optimization method based on the fusion ant colony algorithm, the overall search efficiency and quality of the population can be improved by mutating the parameters of the improved state transition probability model of the local optimal ant and updating it to the improved state transition probability model of the local worst ant. The search efficiency of the ant colony algorithm can be improved by performing path geometry pruning operations on the path of the global optimal ant.

[0104] In one embodiment, the pheromone information and the heuristic information are input into the improved state transition probability model of each ant, and the next path grid is selected from the territory grid of the path grid where the ant is located, including the steps of:

[0105] The pheromone factor and heuristic information factor in the state transition probability model are adjusted and improved based on the path advantage evaluation entropy.

[0106] The roulette threshold in the improved state transition probability model is adjusted based on the number of iterations.

[0107] The expression of the adjusted improved state transition probability model is as follows:

[0108]

[0109]

[0110]

[0111] In the formula, The next path grid of the kth ant at the tth iteration is selected by the improved state transition probability model. ARG MAX is the maximum point function, Λ∈allowed represents the set of grids that the ant is allowed to visit next, and α k [H(S t-1 )] is the pheromone factor of the kth ant, β k [H(S t-1 )] is the heuristic information factor of the kth ant, N iter is the number of iterations, q is a random number, is the roulette threshold of the kth ant, Λ roulette is the path grid selected by roulette, is the probability that the kth ant selects the path grid through roulette at the tth iteration, m is the number of ants, L(S k,t ) is the advantage evaluation item of the k-th ant at the t-th iteration.

[0112] Furthermore, the feasible expressions of the roulette threshold, pheromone factor, and heuristic information factor of the kth ant are as follows:

[0113]

[0114]

[0115]

[0116] Where U is the final coefficient of the roulette threshold, and V is the starting coefficient of the roulette threshold. is the maximum value of the pheromone factor of the kth ant, is the initial coefficient of the pheromone factor of the kth ant, is the path advantage evaluation entropy coefficient of the pheromone factor of the kth ant, is the minimum value of the heuristic information factor of the kth ant, is the starting coefficient of the heuristic information factor of the kth ant, is the path advantage evaluation entropy coefficient of the heuristic information factor of the kth ant.

[0117] In one embodiment, please refer to Figure 6 and Figure 7 , obtain the advantage evaluation function, including:

[0118] Step S601: input the ratio of the length of the ant's path to the distance between the starting grid and the end grid into an exponential path evaluation formula to construct a distance advantage term of the advantage evaluation function.

[0119] Specifically, the distance advantage term is used to characterize the distance advantage of the path corresponding to the ant colony. The feasible expression of the distance advantage term is as follows:

[0120]

[0121] Where n is the grid coefficient in the path, Λ i is the grid that the path of the kth ant passes through, where the zeroth grid is the starting grid and the nth grid is the ending grid; For path S k The distance from the i-th grid to the i+1-th grid is the path S k The vector from the i-th grid to the i+1-th grid.

[0122] Step S602 : generating an obstacle sputtering coefficient for a path grid in a neighborhood grid of the obstacle grid based on an obstacle two-dimensional Gaussian model.

[0123] Specifically, the obstacle splash coefficient is used to characterize the impact of the obstacle grid on the intelligent robot passing through the path grid.

[0124] Step S603: constructing an obstacle avoidance advantage term of the advantage evaluation function based on the obstacle sputtering coefficient.

[0125] Specifically, the obstacle avoidance advantage term is used to characterize the obstacle avoidance advantage of the path corresponding to the ant colony. The feasible expression of the obstacle avoidance advantage term is as follows:

[0126]

[0127] In the formula, Ψ i is the obstacle grid closest to the i-th path grid, and is the grid Λ i Pointing grid Ψi Vector.

[0128] Step S604: input the steering angle of the path into the second-order trigonometric Fourier expansion formula of the proportional function, obtain the smoothing coefficient of the steering angle, and construct the steering advantage term of the advantage evaluation function based on the smoothing coefficient.

[0129] Specifically, the turning advantage term is used to characterize the turning advantage of the path corresponding to the ant colony. A feasible expression of the turning advantage term is as follows:

[0130]

[0131] In the formula, A θ is the initial coefficient of the steering advantage term, is the grid Λ i Pointing grid Λ i+1 The vector of is the grid Λ i+1 Pointing grid Λ i+2 The vector of is the steering angle.

[0132] Furthermore, the computing control platform can express the first-order expansion term and the second-order expansion term in the second-order trigonometric Fourier expansion formula based on the sine and cosine functions of the angle through the trigonometric function identity transformation formula. Another feasible expression of the steering advantage term after expression is as follows:

[0133]

[0134] Wherein, is the expression of the first-order expansion term and the second-order expansion term in the second-order trigonometric Fourier expansion formula based on the sine function and cosine function of the steering angle.

[0135] In one embodiment, if the ant colony path space satisfies the iteration termination condition, generating a global optimal path based on the optimal collision-free path in the ant colony path space includes the steps of:

[0136] The non-adjacent path grids whose connecting lines in the optimal collision-free path avoid obstacle grids and whose obstacle splashing coefficients of the path grids passed through meet the collision avoidance conditions are marked as grids to be optimized, and the path grids in the path in the middle of the grids to be optimized are deleted to generate a smooth optimized initial path. The collision avoidance conditions are used to characterize the boundaries of the intelligent robot where there is a collision risk.

[0137] The distance from the obstacle grid closest to the corner of the smooth optimized initial path to the corner is set as the corner radius, and the corner is replaced with a fillet with a radius equal to the corner radius to generate a global optimal path.

[0138] Schematically, through the above intelligent robot path planning and optimization method based on fusion ant colony algorithm, the computing control platform can obtain the following Figure 8 The global optimal path search result is shown.

[0139] In one embodiment, the intelligent robot path planning and optimization method based on the ant colony algorithm further includes the steps of:

[0140] When the intelligent robot passes through a rounded corner, the steering angle of the intelligent robot is regulated based on the corner radius, the driving speed of the intelligent robot and the control parameters of the intelligent robot.

[0141] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0142] Based on the same inventive concept, the embodiment of the present application also provides an intelligent robot path planning and optimization device based on a fusion ant colony algorithm for implementing the intelligent robot path planning and optimization method based on a fusion ant colony algorithm. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the intelligent robot path planning and optimization device based on a fusion ant colony algorithm provided below can refer to the limitations of the intelligent robot path planning and optimization method based on a fusion ant colony algorithm above, and will not be repeated here.

[0143] In an exemplary embodiment, Fig. 9 As shown, an intelligent robot path planning and optimization device 9000 based on fusion ant colony algorithm is provided, comprising:

[0144] Initialization module 9010 can be used to obtain a global grid map and an advantage evaluation function, and initialize an artificial potential field and an ant colony. The advantage evaluation function is used to characterize the distance advantage, obstacle avoidance advantage, and steering advantage of the path corresponding to the ant colony.

[0145] The iterative search module 9020 can be used to place the ant colony at the starting grid for searching, and obtain the ant colony path space after each search, and the ant colony path space is used to represent the path taken by the ant colony during the search process;

[0146] The iteration determination module 9030 can be used to re-place the ant colony at the starting grid for searching if the ant colony path space does not meet the iteration end condition, until the ant colony path space meets the iteration end condition; the iteration end condition is used to characterize that the ant colony path space includes the optimal collision-free path from the starting grid to the end grid, and the optimal collision-free path is a path whose advantage value meets the global optimal condition;

[0147] The result output module 9040 may be used to generate a global optimal path based on the optimal collision-free path in the ant colony path space if the ant colony path space meets the iteration end condition.

[0148] Among them, the iterative search module includes a heuristic information calculation component, a path grid stepping component and a data calculation and update component.

[0149] The heuristic information calculation component 9021 can be used to calculate the heuristic information using an improved heuristic function combining iteration parameters and artificial potential fields.

[0150] The path grid stepping component 9022 can be used to obtain pheromone information, and input the pheromone information and heuristic information into the improved state transition probability model of each ant, select the next path grid from the territory grid of the path grid where the ant is located, and update the heuristic information.

[0151] The data calculation and update component 9023 can be used to input the ant's path into the advantage evaluation function to calculate the advantage value if the single search end condition is met, and update the pheromone information and the parameters of the ant colony.

[0152] In an optional embodiment, the data calculation and update component 9023 can also be used to set the penalty mechanism of the dynamic weight pheromone sputtering global update strategy with a penalty mechanism based on the proportion of path grids in which the repulsive force value of the artificial potential field in the neighborhood grid is lower than the obstacle risk threshold; set the dynamic weight of the dynamic weight pheromone sputtering global update strategy with a penalty mechanism based on the path advantage evaluation entropy; and globally update pheromone information based on the dynamic weight pheromone sputtering global update strategy with a penalty mechanism.

[0153] In an optional embodiment, the data calculation and update component 9023 can also be used to obtain the local optimal ant, the local worst ant and the global optimal ant based on the advantage value of the ant's path; update the parameters of the improved state transition probability model of the local optimal ant to the improved state transition probability model of the local worst ant after mutation; after the path advantage evaluation entropy meets the path geometry pruning conditions, perform path geometry pruning operations on the path of the global optimal ant to generate a global optimization path; generate pheromone enhancement information on the global optimization path, and update the pheromone information increment before splashing based on the pheromone enhancement information.

[0154] In an optional embodiment, the path grid stepping component 9022 can also be used to adjust the pheromone factor and the heuristic information factor in the improved state transition probability model based on the path advantage evaluation entropy; and adjust the roulette threshold in the improved state transition probability model based on the number of iterations.

[0155] In an optional embodiment, the initialization module 9010 can also be used to input the ratio of the length of the ant's path and the distance between the starting grid and the end grid into the exponential path evaluation formula to construct the distance advantage term of the advantage evaluation function; for the path grid in the neighborhood grid of the obstacle grid, generate an obstacle splash coefficient based on the obstacle two-dimensional Gaussian model; construct the obstacle avoidance advantage term of the advantage evaluation function based on the obstacle splash coefficient; input the steering angle of the path into the second-order trigonometric Fourier expansion formula of the proportional function to obtain the smoothing coefficient of the steering angle, and construct the steering advantage term of the advantage evaluation function based on the smoothing coefficient.

[0156] In an optional embodiment, the result output module 9040 can also be used to mark non-adjacent path grids in the optimal collision-free path in which the connecting lines avoid obstacle grids and pass through path grids whose obstacle splash coefficients meet the collision avoidance conditions as grids to be optimized, and delete the path grids in the path in the middle of the grids to be optimized to generate a smooth optimized initial path; set the distance from the obstacle grid closest to the corner of the smooth optimized initial path to the corner as the corner radius, and replace the corner with a fillet with a radius equal to the corner radius to generate a global optimal path.

[0157] In an optional embodiment, the intelligent robot path planning and optimization device 9000 based on the fusion ant colony algorithm can also be used to control the steering angle of the intelligent robot based on the corner radius, the driving speed of the intelligent robot and the control parameters of the intelligent robot when the intelligent robot passes through a rounded corner.

[0158] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of a power supply safety management method as described above when executing the computer program.

[0159] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0160] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can refer to the partial description of the method embodiments. The device embodiments described above are only schematic, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without paying any creative work.

[0161] The above-mentioned embodiments only express several implementation methods of the embodiments of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the embodiments of the present application. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the embodiments of the present application, and these all belong to the protection scope of the embodiments of the present application.

Claims

1. An intelligent robot path planning and optimization method based on ant colony algorithm, characterized in that: The method comprises: Obtaining a global grid map and an advantage evaluation function, and initializing an artificial potential field and an ant colony, wherein the advantage evaluation function is used to characterize the distance advantage, obstacle avoidance advantage, and steering advantage of the path corresponding to the ant colony; The ant colony is placed at the starting grid for searching, and the ant colony path space after each search is obtained, wherein the ant colony path space is used to represent the path taken by the ant colony during the search process; If the ant colony path space does not meet the iteration end condition, the ant colony is re-placed at the starting grid to search until the ant colony path space meets the iteration end condition; the iteration end condition is used to indicate that the ant colony path space includes an optimal collision-free path from the starting grid to the end grid, and the optimal collision-free path is a path whose advantage value meets the global optimal condition; If the ant colony path space satisfies the iteration end condition, generating a global optimal path based on the optimal collision-free path in the ant colony path space; The step of placing the ant colony at the starting grid for searching comprises the following steps: The heuristic information is calculated using an improved heuristic function combining iteration parameters and artificial potential fields; Acquire pheromone information, and input the pheromone information and the heuristic information into an improved state transition probability model of each ant, select a next path grid from the domain grid of the path grid where the ant is located, and update the heuristic information; If the single search end condition is met, the path of the ant is input into the advantage evaluation function to calculate the advantage value, and the parameters of the ant colony and the pheromone information are updated.

2. The method according to claim 1, characterized in that The updating of the pheromone information comprises: Based on the proportion of the path grids in which the repulsive force value of the artificial potential field in the neighborhood grid is lower than the obstacle risk threshold, a penalty mechanism of the dynamic weighted pheromone sputtering global update strategy with a penalty mechanism is set; The dynamic weight of the dynamic weight pheromone sputtering global update strategy with a penalty mechanism is set based on the path advantage evaluation entropy, wherein the dynamic weight includes the pheromone volatilization dynamic weight, the penalty dynamic weight and the pheromone sputtering dynamic weight; Globally updating the pheromone information based on the dynamic weight pheromone sputtering global update strategy with penalty mechanism; The expression of the dynamic weight pheromone splash update strategy based on the penalty mechanism is as follows: In the formula, For the updated pheromone information, This is the pheromone information before the update. is the pheromone information increment after splashing, is the pheromone information increment before splashing, Λ i,j is a grid with horizontal coordinate i and vertical coordinate j, ρ[H(S t )] is the dynamic weight of pheromone volatilization, H(S t ) is the path advantage evaluation entropy of the t-th search, S is the path of the ant, is the penalty dynamic weight, N acc is the number of path grids in the neighborhood grid whose repulsive force value of the artificial potential field is lower than the obstacle risk threshold, N nbh is the number of path grids in the neighborhood grid, N iter is the number of iterations, is the pheromone information increment before the rth neighborhood path grid is splashed, σ[H(S t )] is the dynamic weight of pheromone splash, is the second norm of the vector from the grid with horizontal coordinate i and vertical coordinate j to its rth neighboring path grid, Q is the pheromone increment coefficient, m is the number of ants, L(S k,t ) is the advantage evaluation item of the k-th ant at the tth iteration.

3. The method according to claim 2, characterized in that The updating of the parameters of the ant colony and the pheromone information includes: Obtaining a local optimal ant, a local worst ant and a global optimal ant based on the advantage value of the path of the ant; mutating the parameters of the improved state transition probability model of the local optimal ant and updating them to the improved state transition probability model of the local worst ant, wherein the parameters include a heuristic information factor, a pheromone factor and a roulette threshold; After the path advantage evaluation entropy meets the path geometry pruning condition, a path geometry pruning operation is performed on the path of the global optimal ant to generate a global optimization path, wherein the path geometry pruning operation is used to connect path nodes in the same row, column or diagonal line; Generate pheromone enhancement information on the global optimization path, and update the pheromone information increment before sputtering based on the pheromone enhancement information.

4. The method according to claim 1, characterized in that The step of inputting the pheromone information and the heuristic information into the improved state transition probability model of each ant, and selecting the next path grid from the domain grid of the path grid where the ant is located, comprises: Adjusting the pheromone factor and the heuristic information factor in the improved state transition probability model based on the path advantage evaluation entropy; adjusting a roulette wheel threshold in the improved state transition probability model based on the number of iterations; The expression of the improved state transition probability model after adjustment is as follows: In the formula, is the next path grid of the kth ant at the tth iteration selected by the improved state transition probability model, ARG MAX is the maximum point function, Λ∈allowed represents the set of grids that the ant is allowed to visit next, α k [H(S t-1 )] is the pheromone factor of the kth ant, β k [H(S t-1 )] is the heuristic information factor of the kth ant, N iter is the number of iterations, q is a random number, is the roulette threshold of the kth ant, Λ roulette is the path grid selected by roulette, is the probability that the kth ant selects the path grid through roulette at the tth iteration, m is the number of ants, L(S k,t ) is the advantage evaluation item of the k-th ant at the t-th iteration.

5. The method according to any one of claims 1 to 4, characterized in that: The obtaining of the advantage evaluation function comprises: The ratio of the length of the path of the ant to the distance between the starting grid and the end grid is input into the exponential path evaluation formula to construct the distance advantage term of the advantage evaluation function, and the distance advantage term is used to characterize the distance advantage of the path corresponding to the ant colony; For a path grid in a neighborhood grid of the obstacle grid, an obstacle splash coefficient is generated based on an obstacle two-dimensional Gaussian model, wherein the obstacle splash coefficient is used to characterize the influence of the obstacle grid on the intelligent robot passing through the path grid; constructing an obstacle avoidance advantage term of the advantage evaluation function based on the obstacle splash coefficient, wherein the obstacle avoidance advantage term is used to characterize the obstacle avoidance advantage of the path corresponding to the ant colony; The steering angle of the path is input into the second-order trigonometric Fourier expansion formula of the proportional function to obtain the smoothing coefficient of the steering angle, and the steering advantage term of the advantage evaluation function is constructed based on the smoothing coefficient. The steering advantage term is used to characterize the steering advantage of the path corresponding to the ant colony.

6. The method according to claim 5, characterized in that If the ant colony path space satisfies the iteration end condition, generating a global optimal path based on the optimal collision-free path in the ant colony path space includes: The non-adjacent path grids whose obstacle splash coefficients of the path grids that the connecting line in the optimal collision-free path avoids and passes through meet the collision avoidance condition are marked as grids to be optimized, and the path grids in the path in the middle of the grids to be optimized are deleted to generate a smooth optimized initial path, and the collision avoidance condition is used to characterize the boundary of the intelligent robot where there is a collision risk; The distance between the obstacle grid closest to the corner of the smooth optimized initial path and the corner is set as the corner radius, and the corner is replaced with a rounded corner with a radius equal to the corner radius to generate the global optimal path.

7. The method according to claim 6, characterized in that The method further comprises: When the intelligent robot passes through the rounded corner, the steering angle of the intelligent robot is regulated based on the corner radius, the driving speed of the intelligent robot and the control parameters of the intelligent robot.

8. An intelligent robot path planning and optimization device based on ant colony algorithm, characterized in that: The device comprises: An initialization module is used to obtain a global grid map and an advantage evaluation function, and initialize an artificial potential field and an ant colony, wherein the advantage evaluation function is used to characterize the distance advantage, obstacle avoidance advantage, and steering advantage of the path corresponding to the ant colony; An iterative search module, used for placing the ant colony at a starting grid for searching, and obtaining an ant colony path space after each search, wherein the ant colony path space is used to represent the path taken by the ant colony during the search process; an iterative determination module, for re-placing the ant colony at the starting grid for searching if the ant colony path space does not meet the iterative end condition, until the ant colony path space meets the iterative end condition; the iterative end condition is used to indicate that the ant colony path space includes an optimal collision-free path from the starting grid to the end grid, and the optimal collision-free path is a path whose advantage value meets the global optimal condition; A result output module, configured to generate a global optimal path based on the optimal collision-free path in the ant colony path space if the ant colony path space satisfies the iteration end condition; Wherein, the iterative search module includes a heuristic information calculation component, a path grid stepping component and a data calculation and update component; The heuristic information calculation component is used to calculate the heuristic information using an improved heuristic function combining an iteration parameter and an artificial potential field; The path grid stepping component is used to obtain pheromone information, and input the pheromone information and the heuristic information into the improved state transition probability model of each ant, select the next path grid from the territory grid of the path grid where the ant is located, and update the heuristic information; The data calculation and update component is used to input the path of the ant into the advantage evaluation function to calculate the advantage value if the single search end condition is met, and to update the pheromone information and the parameters of the ant colony.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images