Mobile robot path planning method and system using direction auxiliary enhancement algorithm

Abstract

The invention discloses a mobile robot path planning method and system using a direction auxiliary enhancement algorithm. The method comprises the following steps: marking a starting point, a target point and a plurality of obstacle points of a mobile robot; constructing a gridding map on the public two-dimensional plane; constructing a two-dimensional coordinate system on the gridding map to determine the target direction of the mobile robot; according to the target direction of the mobile robot, establishing a path connection graph in the gridding map; marking a jump path corresponding to the individual with the optimal fitness in the path connection graph by using an enhanced black-wing algorithm; outputting the jumping path corresponding to the individual with the optimal fitness as an optimal navigation path of the mobile robot; according to the method, the direction-assisted jump path is introduced, and the jump connection relationship between the free grids is presented in a graph structure manner, so that the construction of the long-distance jump path based on the target direction is realized, the limitation of a traditional point-by-point extension method is broken through, and the path planning efficiency of the mobile robot during path planning is realized.

Description

Technical Field

[0001] The present invention relates to robot path planning, specifically to a path planning method and system for a mobile robot using a direction-assisted enhancement algorithm. Background Art

[0002] Existing mobile robot path planning methods, such as the Dijkstra algorithm, Rapidly-Exploring Random Trees (RRT), etc., usually rely on a point-by-point expansion strategy with a single grid as the step size. The patent document with the patent publication number CN118672267A discloses a mobile robot path planning method based on an improved black-winged kite optimization algorithm, which can establish a global path optimization function with the goal of the shortest driving path without colliding with obstacles, and solve the global path optimization function through the improved black-winged kite algorithm, and then plan the optimal collision avoidance path. Although this prior art defines the path planning of the grid map as the path construction of grid-level adjacency relations, it does not achieve direct jump connections between distant free grids, resulting in a large number of intermediate nodes being required in the path construction process, the path being too long, and the navigation efficiency being low. That is to say, in the path construction process, the prior art usually does not abstract the "jumpability" between free grids into a graph structure, that is, does not establish the edge connection relationship between reachable grids within the global scope. Therefore, the jump schemes spanning multiple grids cannot be evaluated as a whole during the path construction process, and only point-by-point advancement can be carried out. Moreover, in the path construction process, there is a lack of a guiding mechanism for the target direction, which easily causes path branches to diverge, the number of turns to increase, the convergence speed to decrease, and affects the path quality and planning efficiency. Summary of the Invention

[0003] Aiming at the deficiencies of the prior art, the present invention provides a path planning method and system for a mobile robot using a direction-assisted enhancement algorithm, and solves the technical problems raised in the background art by establishing jump paths in free grids and using an enhancement algorithm to obtain the optimal jump path.

[0004] To achieve the above objectives, the present invention is realized through the following technical solutions: In the first aspect, the present invention provides a path planning method for a mobile robot using a direction-assisted enhancement algorithm, including: S1. Mark the starting point, target point and several obstacle points of the mobile robot; wherein, the several obstacle points enclose an obstacle area, and both the starting point and the target point are located on the common two-dimensional plane of the obstacle area; S2. Construct a grid map on the common two-dimensional plane; S3. Construct a two-dimensional coordinate system on the grid map to determine the target direction of the mobile robot; S4. Establish a path connection graph within the grid map according to the target direction of the mobile robot; S5. Mark the jump path corresponding to the optimal fitness individual in the path connection graph using the enhanced black-winged kite algorithm; S6. Output the jump path corresponding to the optimal fitness individual as the optimal navigation path of the mobile robot.

[0005] In some specific embodiments, a grid map is constructed on a common two-dimensional plane, including: S2-1. On the common two-dimensional plane, construct the smallest closed area enclosing the starting point, the target point, and all obstacle areas, and use the smallest closed area as the path planning plane of the mobile robot; S2-2. Perform rasterization processing on the path planning plane to generate a number of square grids; among them, the number of square grids includes boundary grids and non-boundary grids; S2-3. Determine the four edge neighborhood grids adjacent to each non-boundary grid and the diagonal neighborhood grids between the edge neighborhood grids to construct a grid map; among them, the center point directions of the neighborhood grids form eight candidate directions of the non-boundary grid.

[0006] In some specific embodiments, a two-dimensional coordinate system is constructed on the grid map to determine the target direction of the mobile robot, including: S3-1. According to the grids where the starting point, the target point, and several obstacle points are located in the grid map, determine the grid state of each grid; the grid state includes the starting grid, the target grid, the free grid, and the obstacle grid; S3-2. Use the center point of the starting grid as the coordinate origin to construct a two-dimensional coordinate system; S3-3. Obtain the coordinates of the center point of the target grid in the two-dimensional coordinate system; S3-4. Calculate the coordinate difference between the center point of the target grid and the center point of the starting grid, and determine the target direction among the eight candidate directions.

[0007] In some specific embodiments, a path connection graph is established within the grid map according to the target direction of the mobile robot, including: S4-1. Set the starting grid as the initial current departure grid; S4-2. In the eight-neighborhood grids of the current departure grid, mark the neighborhood grids corresponding to the three jump directions closest to the target direction; among them, the three jump directions include the target main direction and its two adjacent included angle directions; S4-3. Starting from the center point of the current departure grid, emit search rays along the three jump directions respectively, and continuously search for free grids until an obstacle grid or a path boundary is encountered; S4-4. Define the continuously recognized free grids in each search ray direction as the jump candidate set; the jump candidate set represents the jump candidate set composed of adjacent grids and long-distance grids reachable from the current starting grid in the target direction. S4-5. Establish jump paths between the current starting grid and each grid in the jump candidate set as the edges of the path connection graph. S4-6. Update each grid in the current round of jump candidate set as the current starting grid in the next round, and repeat S4-2 to S4-5. S4-7. Continuously iterate the current update round until the path connection graph covers all free grids in the grid map or the current starting grid cannot be expanded any further.

[0008] In some specific embodiments, taking the center point of the starting grid as the coordinate origin, a two-dimensional coordinate system is constructed, including: S3-2-1. Anchor the center point of the starting grid and define it as the coordinate origin. S3-2-2. Mark the center points of the four edge neighborhood grids of the starting grid respectively. S3-3-3. Extend the coordinate axes from the coordinate origin to the center points of the four edge neighborhood grids to form the two-dimensional coordinate system.

[0009] In some specific embodiments, using the enhanced black-winged kite algorithm to mark the optimal fitness path in the path connection graph, including: S5-1. Construct an individual space with the edges of the path connection graph as variables; among them, each edge corresponds to a one-dimensional variable of the individual. S5-2. Initialize the population in the individual space to generate several individuals; among them, each individual represents a set of values of the edges, which is used to construct a path connection subgraph. The expression for initializing the population is: ; Among them, represents the j-th dimensional variable of the i-th individual; dim represents the individual dimension, corresponding to the total number of edges in the path connection graph; pop represents the population size. The lower limit of the dimensional variable value, with a value of 0. respectively represent the upper limits of the values of each dimensional variable, with a value of 1. represents a random number uniformly distributed in the interval [0,1].

[0010] S5-3. Perform attack behavior and migration behavior on each individual respectively, update the set of values of its edges, and generate the corresponding path connection subgraph. S5-4. Connect the subgraphs according to the generated paths, use the Dijkstra algorithm to calculate the jumping path from the starting grid to the target grid, and extract the path length and the number of turning points of the jumping path; S5-5. Construct a fitness function that simultaneously considers the path length and the number of turning points, evaluate the fitness of all individuals, and select the individual with the optimal fitness in the current round; S5-6. Repeat the optimization iteration until the preset maximum number of iterations is reached or the convergence condition is satisfied, and finally output the jumping path represented by the individual with the optimal fitness as the optimal navigation path of the mobile robot.

[0011] In some specific embodiments, the attack behavior of the individual is updated based on a variable spiral search strategy, and its position update formula is as follows: ; where the spiral control parameter is defined as: ; where, represents the variable of the i-th individual in the j-th dimension at the t-th iteration; represents the variable after the update of the attack behavior; l represents a random perturbation factor uniformly sampled within the interval; rand1 and rand2 are random numbers between (0, 1); p is a constant used to control the switching of the search method, with a value of 0.9; z represents a search step adjustment factor that changes dynamically with the iteration progress; k represents a spiral growth constant with a value of 5; n represents a contraction coefficient that changes dynamically with the iteration progress; t represents the current number, and T represents the maximum number of iterations.

[0012] In some specific embodiments, the migration behavior of the individual is based on a combination mechanism of golden sine perturbation and Cauchy mutation, and its position update formula is as follows: ; where the auxiliary parameters are defined as follows: ; where, and both represent the triangular perturbation angle terms used to construct the golden sine rhythm, represents the Cauchy distribution random perturbation value, represents the superposition or combination of the perturbation terms, and respectively represent the non-linear edge weight perturbation control parameters based on the golden ratio, represents the position value of the current global optimal individual in the j-th dimension; and respectively represent the fitness of the current individual and the reference individual; a and b represent the boundaries of the search space, with values of and , represents the golden ratio coefficient, with a value of 0.618, used to construct the distribution structure of the jump points; m represents the dynamic jump adjustment factor, used to expand or compress the individual update amplitude.

[0013] In some specific embodiments, the fitness of the individual is calculated based on the following function: ; where Length represents the path length of the jump path from the starting grid to the target grid, C represents the number of turning points in the jump path, represents a non - negative real number, used to adjust the weight relationship between path smoothness and path length.

[0014] The present invention provides a mobile robot path planning method using a direction - assisted enhancement algorithm, which has the following beneficial effects: The present invention establishes a feasible jump path between the current starting grid and the passable free grids in the target direction, constructs all jump connections in a graph structure, constructs a jump path graph with navigation directivity. Moreover, the introduction of long - distance jump paths effectively reduces the intermediate turning points in the path, making the generated path more smooth as a whole, which meets the execution stability requirements of the path in the actual mobile robot navigation. Therefore, by introducing direction - assisted jump paths, the present invention breaks through the limitations of traditional point - by - point expansion methods, realizes the construction of long - distance jump paths based on the target direction, abstracts the jump connection relationship between free grids in a graph structure manner, and improves the path planning efficiency of the mobile robot during path planning.

[0015] In the second aspect, the present invention provides a mobile robot path planning system using a direction - assisted enhancement algorithm, and the planning system includes: A point marking module, used to mark the starting point, target point and several obstacle points of the mobile robot; among them, several obstacle points enclose an obstacle area, and the starting point and the target point are both located on the common two - dimensional plane of the obstacle area; A grid - map construction module, used to construct a grid - map on the common two - dimensional plane; A target - direction determination module, used to construct a two - dimensional coordinate system on the grid - map to determine the target direction of the mobile robot; A path - connection graph construction module, used to establish a path - connection graph in the grid - map according to the target direction of the mobile robot; An optimal - fitness individual marking module, used to mark the jump path corresponding to the optimal - fitness individual in the path - connection graph using an enhanced black - winged kite algorithm; The optimal navigation path output module is used to output the jump path corresponding to the optimal fitness individual as the optimal navigation path of the mobile robot.

[0016] Compared with the prior art, the beneficial effects of the mobile robot path planning system using the direction-assisted enhancement algorithm of the present invention are the same as those of the mobile robot path planning method using the direction-assisted enhancement algorithm described above, so they will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic flowchart of the mobile robot path planning method using the direction-assisted enhancement algorithm of the present invention; Figure 2 is an iterative flowchart of the path connection diagram described in the present invention; Figure 3 is a structural block diagram of the mobile robot path planning system using the direction-assisted enhancement algorithm of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0019] First, the prior art and related concepts involved in the embodiments of the present invention are described: The Black Kite Algorithm (BKA) is a new metaheuristic optimization method inspired by the attack and migration behaviors of black kites proposed by Wang et al. (2024). The Enhanced Black Kite Algorithm forms the Enhanced Black Kite Algorithm (EBKA) by integrating the variable helix search strategy (Mirjalili and Lewis, 2016) and the golden sine strategy (Tanyildizi and Demir, 2017). EBKA performs path planning by searching for all feasible points on a grid map.

[0020] Embodiment 1: Please refer to Figure 1 , Figure 1 discloses a schematic flowchart of the mobile robot path planning method using the direction-assisted enhancement algorithm. The process includes: S1. Mark the starting point, target point, and several obstacle points of the mobile robot; among them, several obstacle points enclose an obstacle area, and the starting point and the target point are both located on the common two-dimensional plane of the obstacle area; S2. Construct a grid map on the common two-dimensional plane; Exemplarily, the specific implementation steps of S2 are: S2-1. On the common two-dimensional plane, construct the smallest closed region that encloses the starting point, the target point, and all obstacle regions, and use this smallest closed region as the path planning plane for the mobile robot. Specifically, although the starting point, the target point, and the obstacle regions are all on the same common two-dimensional plane, for the purpose of computational efficiency and path search limitation, the path planning plane is usually a closed sub-region on this common two-dimensional plane, and its boundary is jointly determined by task-related elements (starting point, ending point, obstacles).

[0021] S2-2. Perform rasterization processing on the path planning plane to generate a number of square grids; among them, the number of square grids includes boundary grids and non-boundary grids. S2-3. Determine the four edge neighborhood grids adjacent to each non-boundary grid and the diagonal neighborhood grids between the edge neighborhood grids to construct a grid map; among them, the central point directions of the neighborhood grids form eight candidate directions for the non-boundary grid.

[0022] In this embodiment, by defining the smallest closed region and constructing a grid map containing eight-neighborhood direction information, the path search range and the feasible jump directions are strictly limited to the relevant regions of the target grid, significantly reducing the number of traversals of invalid grids, and enabling the subsequent jump paths to only expand in the relevant directions of the target grid, thereby reducing the computational complexity and improving the search convergence speed.

[0023] S3. Construct a two-dimensional coordinate system on the grid map to determine the target direction of the mobile robot. Exemplarily, the specific implementation steps of S3 are as follows: S3-1. According to the grids where the starting point, the target point, and several obstacle points are located in the grid map, determine the grid state of each grid; the grid state includes the starting grid, the target grid, the free grid, and the obstacle grid. S3-2. Take the center point of the starting grid as the coordinate origin to construct a two-dimensional coordinate system. S3-3. Obtain the coordinates of the center point of the target grid in the two-dimensional coordinate system. S3-4. Calculate the coordinate difference between the center point of the target grid and the center point of the starting grid, and determine the target direction among the eight candidate directions.

[0024] In this embodiment, by constructing a unified two-dimensional coordinate system with the center point of the starting grid as the origin, the position of the target grid relative to the starting point can be quantitatively described by the coordinate difference. Based on the differences between the target point and the starting point in the horizontal and vertical axes directions, the relative direction where the target point is located is determined, and it is matched among the preset eight candidate directions, so as to determine the target direction of the jumping path. In the actual implementation of this process, the direction angle of the line connecting the target point and the starting point can be calculated through the arctan2 function. As a two-variable arctangent function, the arctan2 function can accurately identify the quadrant to which the target direction belongs by integrating the horizontal and vertical coordinate differences. Through this method, the direction for guiding the path to jump can be quickly locked, and the search attempts in irrelevant directions can be significantly reduced. Among them, the arctan2 function calculates the direction angle of the target point relative to the starting point, and the value range of this direction angle is from -π to π, which can cover the four quadrants of the entire two-dimensional plane. Subsequently, this direction angle is matched with the preset eight discrete direction angle intervals. For example, from -π / 8 to π / 8 is the due east (right) direction, from π / 8 to 3π / 8 is the northeast (upper right) direction, and so on. The entire angle range is divided into eight equal parts, corresponding to eight candidate jumping directions.

[0025] Further, step S3-2 of this embodiment further includes: S3-2-1, anchor the center point of the starting grid and define it as the coordinate origin; S3-2-2, respectively mark the center points of the grid neighborhoods on the four sides of the starting grid; S3-3-3, extend the coordinate axes from the coordinate origin to the center points of the grid neighborhoods on the four sides to form the two-dimensional coordinate system.

[0026] In this embodiment, by anchoring the center point of the starting grid as the coordinate origin and using the relative positions of the center points of the grid neighborhoods on its four sides to extend and form the coordinate axes in the horizontal and vertical directions respectively, the local construction of the coordinate system and the spatial definition of the direction reference are realized.

[0027] S4. Establish a path connection graph in the grid map according to the target direction of the mobile robot; Please refer to Figure 2 , Figure 2 which discloses a schematic diagram of the iterative process of the path connection graph. Exemplarily, this process includes: S4-1, set the starting grid as the initial current departure grid; S4-2, among the eight-neighborhood grids of the current departure grid, mark the neighborhood grids corresponding to the three jumping directions closest to the target direction; among them, the three jumping directions include the target main direction and its two adjacent included angle directions; S4-3. Starting from the center point of the current starting grid, emit search rays along the three jumping directions respectively, and continuously search for free grids until an obstacle grid or path boundary is encountered; S4-4. Define the free grids continuously identified in each search ray direction as the jumping candidate set; the jumping candidate set represents the jumping candidate set composed of adjacent grids and long-distance grids reachable by the current starting grid along the target direction; S4-5. Establish jumping paths between the current starting grid and each grid in the jumping candidate set as the edges of the path connection graph; S4-6. Update each grid in the current round of jumping candidate set as the current starting grid for the next round in turn, and repeat S4-2 to S4-5; That is to say, for each free point, there is a jumping candidate set. After traversing N free points, there will be N jumping candidate sets, and all possible paths of the N jumping candidate sets form the path connection graph.

[0028] S4-7. Continuously iterate the current update round until the path connection graph covers all free grids in the grid map or the current starting grid cannot be expanded continuously.

[0029] In this embodiment, first, the starting grid is used as the initial starting grid for path expansion, and three directions with the smallest included angle with the target direction are selected from the eight neighborhoods of this grid as the preferred jumping directions, thus introducing clear direction constraints to avoid the divergence of path expansion to areas irrelevant to the target direction. Subsequently, search rays are emitted along the three directions respectively to search for continuous free grids in a jumping manner until an obstacle or boundary is encountered, so as to directly jump from a starting grid to a reachable position far away and skip the redundant expansion process of multiple intermediate grids. In this embodiment, all jumpable grids identified by the search rays are defined as the jumping candidate set, and a directed edge is established between the current starting grid and each jumping grid in the set to construct the edges in the path connection graph. By restricting the generation of edges only based on real and feasible jumping paths, the construction of invalid edges is avoided, and the edge density of the graph is reduced. In each round of iteration, the newly identified jumping grids are used as the starting points again, and continue to perform jumping expansion along the target-related directions to form a hierarchical expansion structure of the graph. This process continues until the graph covers all free grids or cannot be expanded continuously, thus completing the construction process of a connection graph with directionality, sparsity, and jumping ability. This embodiment significantly reduces the number of redundant edges in the graph and reduces the complexity of the search graph.

[0030] S5. Use the enhanced black-winged kite algorithm to mark the jumping path corresponding to the individual with the optimal fitness in the path connection graph; S6. Output the jumping path corresponding to the individual with the optimal fitness as the optimal navigation path of the mobile robot.

[0031] In this embodiment, by introducing target direction assistance and jump path modeling, the construction of a long-distance jump connection graph is realized on the basis of a traditional grid map, and the enhanced black-winged kite algorithm is combined to globally optimize the path, effectively reducing unnecessary turns in path search and improving the calculation efficiency of the navigation path.

[0032] Embodiment 2: The technical solution of this Embodiment 2 is different from that of Embodiment 1 in that the specific application steps of the enhanced black-winged kite algorithm in Embodiment 1 are disclosed. The specific application steps include: S5-1. Taking the edges of the path connection graph as variables, construct an individual space; wherein, each edge corresponds to a one-dimensional variable of the individual; S5-2. Initialize a population in the individual space to generate a number of individuals; wherein, each individual represents a set of values of the edges, which is used to construct a path connection sub-graph; The expression for initializing the population is: ; wherein, represents the j-th dimensional variable of the i-th individual; dim represents the individual dimension, corresponding to the total number of edges in the path connection graph; pop represents the population size; The lower limit of the value of the dimensional variable, with a value of 0; respectively represent the upper limits of the values of each dimensional variable, with a value of 1; represents a random number uniformly distributed in the interval [0, 1].

[0033] In this embodiment, to construct the initial solution space required for path optimization, a standard random initialization mechanism of the enhanced black-winged kite algorithm is used to generate an optimization population. Each individual represents a combination method of a group of edges (jump paths), and its coding dimension is equal to the number of all feasible edges in the path connection graph. Each dimension corresponds to a variable of an edge. Uniform random values are generated within the upper and lower limits defined by the variable of each edge, thereby constructing a set of initial individual sets that cover different combination methods of the jump path space.

[0034] In the direction-assisted jump graph proposed by the present invention, each edge represents the actual path possibility of jumping from one free grid to another far-distance free grid. The population initialized in the above manner can comprehensively cover the possible jump connection combinations in the graph, providing sufficient search diversity for the subsequent iterative process.

[0035] S5-3. Execute the attack behavior and migration behavior on each individual respectively, update the set of values of its edges, and generate the corresponding path connection sub-graph; S5-4. Connect the subgraphs according to the generated paths, use the Dijkstra algorithm to calculate the jumping path from the starting grid to the target grid, and extract the path length and the number of turning points of the jumping path; S5-5. Construct a fitness function that simultaneously considers the path length and the number of turning points, evaluate the fitness of all individuals, and select the individual with the optimal fitness in the current round; S5-6. Repeat the optimization iteration until the preset maximum number of iterations is reached or the convergence condition is met, and finally output the jumping path represented by the individual with the optimal fitness as the optimal navigation path of the mobile robot.

[0036] In this embodiment, an individual corresponds to a path connection subgraph, the dimension of each individual is the number of edges in the original path connection graph, and when an individual changes once, the path connection graph changes.

[0037] In this embodiment, the jumping path in the path connection graph is abstracted as a multi-dimensional variable of the individual, and a continuous optimization search space for the combination of jumping paths is constructed. Each optimized individual corresponds to a set of activation states of the edges, and through the attack behavior and migration behavior in the enhanced black-winged kite algorithm for dynamic iterative update, the search and evolution of the path configuration method are realized globally. By generating the path connection subgraph in each round of update and calling the Dijkstra algorithm to evaluate the path validity, the system can dynamically obtain two key navigation indicators, namely the path length and the number of turning points, and construct a multi-objective fitness function based on this to ensure that the path solution takes into account both the shortest distance and the smoothness of the path.

[0038] Furthermore, the Dijkstra algorithm is a classic algorithm for calculating the shortest path of units in a graph and is applicable to all graph structures with non-negative edge weights. Its goal is to find the shortest path from the starting node to any other node in a graph composed of nodes and edges.

[0039] In this embodiment, the system constructs a path connection graph with free grids as nodes and jumping paths as edges, and forms a weighted graph with the jumping paths. The weight of each edge can be defined as the Euclidean distance of the jumping path (i.e., the geometric distance between the center points of two grids). The system takes the starting grid as the starting point and the target grid as the end point, and uses the Dijkstra algorithm to search for the shortest jumping path from the starting point to the end point in the graph.

[0040] Furthermore, the attack behavior of the individual in this embodiment updates the position based on the variable spiral search strategy, and its position update formula is as follows: ; where the spiral control parameter is defined as: ; where, represents the variable of the tth iteration of the i-th individual on the j-th dimension (i.e., the jump path); represents the updated variable of the attack behavior; l represents the The random perturbation factor uniformly sampled within the interval; rand1 and rand2 are random numbers between (0,1); p is a constant used to control the switching of search modes, with a value of 0.9; z represents the search step adjustment factor, which changes dynamically with the progress of the iteration; k represents the spiral growth constant, with a value of 5; n represents a contraction coefficient for the dynamic change of the iteration progress; t represents the current number, and T represents the maximum number of iterations.

[0041] Specifically, the contraction coefficient n gradually decreases with the increase of the number of iterations t; therefore, the spiral control parameter z constructs a search trend of "wide exploration in the early stage and contraction and refinement in the later stage".

[0042] In the early stage of the algorithm, t / T is close to 0, the exponential term approaches 1, and n is large → the search radius is large, which is conducive to jumping; in the later stage of the algorithm, t / T is close to 1, which causes n to decrease, causing the search radius to shrink, which is conducive to fine convergence; therefore, the shrinkage coefficient n is used to calculate the perturbation factor to further control the overall search radius.

[0043] Furthermore, since l represents The random perturbation factor uniformly sampled within the interval, so after multiplying l by 2π, A periodic disturbance term is formed, oscillating in the range [-1,1]; therefore, It simulates the directional rotation change of the black-winged kite circling, that is, the direction of the jump has a "rotation amplitude".

[0044] Furthermore, in In the example, rand represents a random number between 0 and 1. ≈[0,0.84], after adding 1, the range is [1,1.84]; thus, is a moderate growth disturbance factor to avoid the disturbance amplitude being too small. Finally, a controllable spiral amplitude adjustment mechanism is realized.

[0045] In summary, The combination of:

[0046] Determines the rotation direction to form a spiral radial structure; n controls z so that the disturbance amplitude decays with iteration; The disturbance amplitude is slightly extensible to enhance the randomness of the disturbance. Finally, the black-winged kite's aerial "dynamic zoom + directional circling" search trajectory is simulated. , to generate perturbations in both positive and negative directions; thereby guiding the individual to make tentative jumps in an uncertain direction; in cooperation with Provide perturbation direction + amplitude control, and finally simulate the "random wandering search" of the black-winged kite when the target is not locked. That is, when the individual has not judged the direction advantage, a more randomized and two-way diffusion perturbation is used to jump out of the current area, enhancing the global search ability and avoiding falling into the local optimum prematurely.

[0047] Therefore, the attack behavior strategy is actually composed of two update modes: when p < rand, it is the rotation approximation update mode: This branch controls the jump direction and perturbation size through sine perturbation and exponential scaling to generate a progressive search path around the target direction, so as to simulate the hovering approximation process of the black-winged kite when locking the prey.

[0048] Otherwise, when the jump direction is uncertain or the perturbation is large, use the factor to generate two-way perturbations, and realize the positive and negative direction diffusion search of the edge variables, which is beneficial to jump out of the local optimum solution area.

[0049] Furthermore, in this embodiment, the migration behavior of the individual is based on the combined mechanism of golden sine perturbation and Cauchy mutation, and its position update formula is as follows: ; Among them, the auxiliary parameters are defined as follows: ; Among them, and both represent the triangular perturbation angle term, which is used to construct the golden sine rhythm, represents the Cauchy distribution random perturbation value, represents the superposition or combination of the perturbation terms, and respectively represent the non-linear edge weight perturbation control parameters based on the golden ratio, represents the position value of the current global optimal individual in the j-th dimension; and respectively represent the fitness of the current individual and the reference individual; a and b represent the boundaries of the search space, and the values are and , represents the golden ratio coefficient, with a value of 0.618, which is used to construct the distribution structure of the jump points; m represents the dynamic jump adjustment factor, which is used to expand or compress the individual update amplitude.

[0050] In this embodiment, to further improve the jumping breadth and search flexibility of an individual in the path jump combination space, the migration behavior strategy in the Enhanced Black-winged Kite Algorithm (EBKA) is adopted, which integrates the golden sine perturbation mechanism and the Cauchy mutation mechanism to update the variables of the individual. This strategy simulates the jumping and mutation behaviors of the black-winged kite during its aerial migration, and is particularly suitable for the requirements of "jumping out of local optima" and "crossing obstacle areas" in path search. The current position of the i-th individual in the j-th edge dimension of the individual is denoted as .

[0051] Specifically, the migration behavior strategy also consists of two perturbation modes. When the current individual is better than the random control individual, it mainly takes the variables of the current path as the benchmark and fine-tunes along the direction deviating from the global optimal solution to maintain the stability of the current individual's path configuration and achieve positive jump perturbation (when the fitness is better); if the current individual is inferior to the reference individual, it takes the global optimal solution as the benchmark and simultaneously introduces m terms for enhanced perturbation, forming a periodic and asymmetric edge adjustment through the golden ratio and the sine function to enhance the long-distance search ability and achieve reverse jump perturbation (when the fitness is not good).

[0052] The golden sine control term makes the individual jump update show a structural change of "period + ratio", while the Cauchy perturbation term introduces large-perturbation tail jumps to avoid the path falling into the local structure optimum. After the two are combined, it can significantly improve the jump diversity, jumping-out ability and global solution exploration quality of path optimization in a complex connection graph.

[0053] Among them, the fitness of the individual is calculated based on the following function: ; where Length represents the path length of the jump path from the starting grid to the target grid, C represents the number of turning points in the jump path, represents a non-negative real number used to adjust the weight relationship between path smoothness and path length.

[0054] When the path optimization iteration process reaches the set maximum number of generations, or the fitness value of the current optimal path individual fluctuates below the preset threshold in consecutive multiple rounds, the iteration terminates, and finally the individual path with the minimum fitness is selected as the navigation path output.

[0055] This fitness function takes the path length as the basic navigation cost, encourages the generation of a globally shorter path; at the same time, it considers the number of turning points in the path, restricts the path tortuosity, and encourages a smoother and more directionally consistent jump path; by adjusting the coefficient , the "path efficiency (shortest)" and "path executability (smooth)" can be flexibly weighted and balanced according to the actual task requirements. During the optimization process, after each path individual completes the attack and migration behaviors and generates a path subgraph, it will calculate its actual jump path and evaluate the total jump length of the current path and the number of turns C experienced, and calculate the fitness; as the optimization iteration progresses, the individual with the smallest fitness value will be continuously selected as the optimal solution for the current round until one of the following termination conditions is met: reaching the preset maximum number of iterations; the fitness of the optimal solution does not change significantly in several rounds (meeting the convergence threshold); finally, the jump path corresponding to the optimal individual is selected as the output result of the robot navigation path.

[0056] By calculating the fitness, the phenomenon of "the shortest distance but the path is too tortuous" can be effectively avoided, especially applicable to task scenarios with high requirements for "path smoothness" and "navigation controllability" in complex obstacle environments, such as path planning tasks of mobile platforms such as service robots and inspection robots.

[0057] Example 3, please refer to Figure 3 , Figure 3 The structural block diagram of a mobile robot path planning system using a direction-assisted enhancement algorithm is disclosed. The technical solution of this Example 3 is different from that of Example 1 and Example 2 in that an execution planning system for the planning methods described in Example 1 and Example 2 is disclosed. The system includes: A point marking module for marking the starting point, target point, and several obstacle points of the mobile robot; among them, several obstacle points enclose an obstacle area, and the starting point and the target point are both located on the common two-dimensional plane of the obstacle area; A grid map construction module for constructing a grid map on the common two-dimensional plane; A target direction determination module for constructing a two-dimensional coordinate system on the grid map to determine the target direction of the mobile robot; A path connection graph construction module for establishing a path connection graph within the grid map according to the target direction of the mobile robot; An optimal fitness individual marking module for marking the jump path corresponding to the optimal fitness individual in the path connection graph using the enhanced black-winged kite algorithm; An optimal navigation path output module for outputting the jump path corresponding to the optimal fitness individual as the optimal navigation path of the mobile robot.

[0058] By introducing target direction assistance and jump path modeling, this system realizes the construction of a long-distance jump connection graph on the basis of a traditional grid map, and combines the enhanced black-winged kite algorithm to globally optimize the path, effectively reducing unnecessary turns in path search and improving the calculation efficiency of the navigation path.

[0059] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as infrared, wireless, microwave, etc.).

[0060] The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that contains one or more collections of available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVD ), or semiconductor media. The semiconductor media can be a solid-state drive.

[0061] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of devices or units can be in electrical, mechanical, or other forms.

[0062] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered by the protection scope of the present application.

Claims

1. A mobile robot path planning method using a direction-assisted enhancement algorithm, characterized in that: include: S1. Marking a starting point, a target point, and several obstacle points of the mobile robot; wherein the several obstacle points together form an obstacle area, and the starting point and the target point are both located on a common two-dimensional plane of the obstacle area; S2, constructing a grid map on a common two-dimensional plane; S3, constructing a two-dimensional coordinate system on the grid map to determine the target direction of the mobile robot; S4, establishing a path connection diagram in the grid map according to the target direction of the mobile robot; S5, using the enhanced black-winged kite algorithm to mark the jump path corresponding to the individual with the best fitness in the path connection graph; S6. Output the jumping path corresponding to the individual with the best fitness as the optimal navigation path of the mobile robot.

2. The mobile robot path planning method using the direction-assisted enhancement algorithm according to claim 1 is characterized in that: Construct a gridded map on a common 2D plane, including: S2-1, constructing a minimum closed area enveloping the starting point, the target point and all obstacle areas on the common two-dimensional plane, and using the minimum closed area as the path planning plane of the mobile robot; S2-2, rasterizing the path planning plane to generate a plurality of square grids; wherein the plurality of square grids include boundary grids and non-boundary grids; S2-3. For each non-boundary grid, determine the diagonal neighbor grids between the four adjacent edge neighbor grids and the edge neighbor grids to construct a gridded map; wherein the center point directions of the neighbor grids constitute eight candidate directions of the non-boundary grids.

3. The mobile robot path planning method using the direction-assisted enhancement algorithm according to claim 1 is characterized in that: Construct a two-dimensional coordinate system on the grid map to determine the target direction of the mobile robot, including: S3-1, determining the grid state of each grid according to the grids where the starting point, the target point and the plurality of obstacle points are located in the grid map; the grid state includes the starting grid, the target grid, the free grid and the obstacle grid; S3-2, construct a two-dimensional coordinate system with the center point of the starting grid as the coordinate origin; S3-3, obtaining the coordinates of the center point of the target grid in the two-dimensional coordinate system; S3-4. Calculate the coordinate difference between the center point of the target grid and the center point of the starting grid, and determine the target direction from the eight candidate directions.

4. The mobile robot path planning method using the direction-assisted enhancement algorithm according to claim 1 is characterized in that: According to the target direction of the mobile robot, a path connection diagram is established in the grid map, including: S4-1, setting the starting grid as the initial current starting grid; S4-2, in the eight neighborhood grids of the current starting grid, marking the neighborhood grids corresponding to the three jump directions closest to the target direction; wherein the three jump directions include the target main direction and its two adjacent angle directions; S4-3, starting from the center point of the current starting grid, emitting search rays along the three jumping directions respectively, and continuously searching for free grids until an obstacle grid or a path boundary is encountered; S4-4, defining the free grids continuously identified in the direction of each search ray as a jump candidate set; the jump candidate set represents the jump candidate set consisting of adjacent grids and distant grids reachable from the current starting grid along the target direction; S4-5, establishing a jump path between the current departure grid and each grid in the jump candidate set as an edge of the path connection graph; S4-6, updating each grid in the current round of jump candidate set to the current starting grid of the next round in sequence, and repeating S4-2 to S4-5; S4-7. Continue to iterate the current update round until the path connection diagram covers all free grids in the grid map, or the current starting grid cannot be expanded further.

5. The mobile robot path planning method using the direction-assisted enhancement algorithm according to claim 3 is characterized in that: Take the center point of the starting grid as the coordinate origin and construct the two-dimensional coordinates, including: S3-2-1. Anchor the center point of the starting grid and define it as the coordinate origin; S3-2-2, mark the center points of the four neighboring grids of the starting grid respectively; S3-3-3. Extend the coordinate axes from the coordinate origin toward the center points of the four side neighborhood grids to form the two-dimensional coordinate system.

6. The mobile robot path planning method using the direction-assisted enhancement algorithm according to claim 1 is characterized in that: Use the enhanced black kite algorithm to mark the best fitness path in the path connection graph, including: S5-1. Use the edges of the path connection graph as variables to construct an individual space, where each edge corresponds to a one-dimensional variable of an individual; S5-2. Initialize the population in the individual space and generate a number of individuals; each individual represents a set of edge values, which are used to construct a path-connected subgraph; The expression of the initialization population is: ； in, represents the j-th dimension variable of the i-th individual; dim represents the individual dimension, corresponding to the total number of edges in the path connection graph; pop represents the population size; The lower limit of the dimension variable value is 0; Respectively represent the upper limit of the value of each dimension variable, and the value is 1; Represents a random number uniformly distributed in the interval [0,1]; S5-3, execute attack behavior and migration behavior for each individual, update the value set of its edge, and generate the corresponding path connection subgraph; S5-4, according to the generated path connection subgraph, using the Dijkstra algorithm to calculate the jump path from the starting grid to the target grid, and extracting the path length and the number of turning points of the jump path; S5-5. Construct a fitness function that takes both the path length and the number of turning points into consideration, evaluate the fitness of all individuals, and select the individual with the best fitness in the current round; S5-6. Repeat the optimization iteration until the preset maximum number of iterations is reached or the convergence condition is met, and finally output the jumping path represented by the individual with the best fitness as the optimal navigation path of the mobile robot.

7. The mobile robot path planning method using the direction-assisted enhancement algorithm according to claim 1 is characterized in that: The attack behavior of the individual is based on the variable spiral search strategy to update the position, and the position update formula is as follows: ； Among them, the spiral control parameter is defined as: ； in, represents the variable of the tth iteration of the i-th individual on the j-th dimension; represents the variable after the attack behavior is updated; l represents A random perturbation factor uniformly sampled within the interval; rand1 and rand2 are random numbers between (0,1); p is a constant used to control the switching of search modes, with a value of 0.9; z represents the search step adjustment factor, which changes dynamically with the progress of the iteration; k represents the spiral growth constant, with a value of 5; n represents a contraction coefficient for the dynamic change of the iteration progress; t represents the current number, and T represents the maximum number of iterations.

8. The mobile robot path planning method using the direction-assisted enhancement algorithm according to claim 1 is characterized in that: The migration behavior of the individual is based on the combined mechanism of golden sine perturbation and Cauchy mutation, and its position update formula is as follows: ； Among them, the auxiliary parameters are defined as follows: ； in, and Both represent triangular perturbation angle terms, which are used to construct the golden sine rhythm. represents the Cauchy distributed random perturbation value, represents the superposition or combination of disturbance terms, and They represent the nonlinear edge weight perturbation control parameters based on the golden ratio, Represents the position value of the current global optimal individual in the jth dimension; and Respectively represent the fitness of the current individual and the reference individual; a and b represent the boundaries of the search space, and their values ​​are and , represents the golden ratio coefficient, which is 0.618 and is used to construct the distribution structure of jump points; m represents the dynamic jump adjustment factor, which is used to expand or compress the individual update amplitude.

9. The mobile robot path planning method using the direction-assisted enhancement algorithm according to claim 1, characterized in that: The fitness of the individual is calculated based on the following function: ； Among them, Length represents the path length of the jump path from the starting grid to the target grid, C represents the number of turning points in the jump path, It is expressed as a non-negative real number and is used to adjust the weight relationship between path smoothness and path length.

10. A mobile robot path planning system using a direction-assisted enhancement algorithm, characterized in that: The system comprises: A point marking module is used to mark the starting point, the target point and several obstacle points of the mobile robot; wherein the several obstacle points together form an obstacle area, and the starting point and the target point are both located on a common two-dimensional plane of the obstacle area; A grid map construction module, used to construct a grid map on a common two-dimensional plane; A target direction determination module is used to construct a two-dimensional coordinate system on the grid map to determine the target direction of the mobile robot; A path connection graph building module is used to build a path connection graph in a grid map according to the target direction of the mobile robot; The optimal fitness individual marking module is used to mark the jump path corresponding to the optimal fitness individual in the path connection graph using the enhanced black kite algorithm; The optimal navigation path output module is used to output the jumping path corresponding to the optimal fitness individual as the optimal navigation path of the mobile robot.

Images