A Mobile Robot Path Planning Method Based on an Improved Sparrow Search Algorithm

By improving the population initialization and update strategy of the sparrow search algorithm, the problems of uneven population initialization, parameter fixation, and insufficient topological constraints are solved, and efficient and safe path planning in complex environments is achieved.

CN123065490APending Publication Date: 2026-11-10COLLEGE OF MOBILE TELECOMM CHONGQING UNIV OF POSTS & TELECOMM
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202611090245.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-22
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing sparrow search algorithms suffer from uneven population initialization, fixed parameters, lack of topological constraints, and susceptibility to local optima in path planning, resulting in low efficiency and insufficient security in path planning.

Method used

A fourth-order Chebyshev chaotic map is used for population initialization. An adaptive step size and a ring topology neighborhood reconstruction mechanism are designed. Combined with a phased multi-strategy weighted cooperation mechanism, the population update strategy is optimized to improve the diversity and security of path planning.

Benefits of technology

It significantly improves the quality and stability of path planning in complex environments, enhances the efficiency and safety of path planning, and reduces the risk of generating invalid paths and getting trapped in local optima.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123065490A_ABST
    Figure CN123065490A_ABST
Patent Text Reader

Abstract

This invention claims protection for a mobile robot path planning method based on an improved sparrow search algorithm, belonging to the field of intelligent optimization and autonomous navigation technology for mobile robots. It includes: constructing a discrete feasible region initialization mechanism using a fourth-order Chebyshev chaotic map, utilizing its uniform distribution characteristics to map the population to a free grid index space; constructing a ring-shaped topology-aware neighborhood reconstruction mechanism based on fitness ranking sequences; designing an adaptive step-size control strategy based on population diversity feedback perception, using normalized population diversity as a driving indicator to dynamically adjust the exploration factor and safety threshold, coordinating global exploration and local development capabilities; and designing a phased, multi-strategy weighted collaborative scout update mechanism, balancing the convergence speed of inferior individuals with the maintenance of diversity of superior individuals, and avoiding infeasible solutions generated by continuous-discrete transformation. This method effectively improves the solution quality, convergence speed, and optimization stability of mobile robot path planning in complex grid environments.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the fields of intelligent swarm optimization algorithms and autonomous navigation and path planning for mobile robots, specifically a path planning method for mobile robots based on an improved sparrow search algorithm. Background Technology

[0002] With the widespread application of mobile robots in warehousing and logistics, service inspection, agricultural operations, emergency rescue, and intelligent manufacturing, their operating environments are becoming increasingly complex and variable, placing higher demands on the efficiency and safety of path planning in autonomous navigation systems. Swarm intelligence optimization algorithms, with their strong global search capabilities and ability to perform optimization calculations without relying on precise environmental gradient information, are well-suited for path solving in complex and unstructured environments and have become one of the mainstream solution methods in the current path planning field. The sparrow search algorithm achieves iterative optimization by simulating the foraging division of labor, predator warning, and cooperative update mechanisms of sparrow populations. It possesses advantages such as fewer parameter settings, faster convergence speed, higher basic optimization accuracy, and clear population division logic, demonstrating good application potential in many optimization problems. However, SSA still has the following technical shortcomings:

[0003] First, the random sampling initialization of the population makes individuals prone to clustering in local areas or falling into obstacle nodes, resulting in uneven distribution and insufficient coverage of the initial population within the feasible region, reducing subsequent optimization efficiency. Second, the discoverer uses a fixed step size update with a constant safety threshold, making it impossible to dynamically adjust the exploration and development intensity based on the real-time convergence status of the population, which can easily lead to insufficient exploration in the early stages or slow development in the later stages. Third, the follower only uses the globally optimal individual as the guiding benchmark, without combining the topological connectivity constraints of the grid map, making it easy for the guiding direction to cross obstacles, generating a large number of invalid paths and wasting iteration resources. Fourth, the scout's update strategy is simplistic and lacks a differentiated processing mechanism for superior and inferior individuals, resulting in excessively rapid decay of population diversity in the later stages of iteration, making it easy to get trapped in local optima and unable to escape the local extreme value region of complex obstacle environments. Fifth, the discrete position update operator is not bound to the free node sequence, and the updated nodes may deviate from the passable area, and there is a lack of path connectivity repair mechanism, so the physical feasibility and execution safety of the output path cannot be guaranteed.

[0004] The improvements proposed in application publication number CN116795098A, which introduce nonlinear sine factors, Lévy flight, and local search strategies, provide valuable insights for algorithm optimization. However, their optimizations largely remain at the level of combining and superimposing general operators. In contrast, this invention implements a systematic reconstruction from the underlying architecture of the algorithm: it employs a fourth-order Chebyshev chaotic mapping to achieve uniform initialization of the discrete feasible region, reducing the probability of the initial solution falling into the obstacle region from the source; it designs an adaptive parameter control mechanism driven by normalized population diversity to dynamically coordinate global exploration and local development capabilities; it constructs a ring-shaped topology-aware neighborhood reconstruction mechanism based on fitness ranking sequences to form an ordered information transmission link to achieve differentiated guidance for inferior individuals; and it proposes a phased multi-strategy weighted collaborative scout update mechanism to balance the convergence efficiency of inferior individuals with the preservation of the diversity of superior individuals, effectively suppressing the generation of infeasible solutions during the continuous-discrete space transformation process. This invention achieves targeted breakthroughs in information guidance accuracy, search strategy adaptability, and discrete constraint handling capabilities through the collaborative design of topology optimization, differentiated update strategy, and adaptive step size control, significantly improving the solution quality, convergence speed, and optimization stability of grid path planning in complex environments. Summary of the Invention

[0005] This invention aims to solve the problems of the prior art mentioned above. It proposes a path planning method for mobile robots based on an improved sparrow search algorithm. The technical solution of this invention is as follows:

[0006] A path planning method for a mobile robot based on an improved sparrow search algorithm includes the following steps:

[0007] S1, acquire environmental map information and construct a rasterized search space model;

[0008] S2 uses fourth-order Chebyshev chaotic mapping for discrete initialization to generate the initial path population X;

[0009] S3, calculate the fitness value of each individual in the population and sort them in ascending order, determine the discoverers and followers according to a preset ratio, and randomly select a number of individuals as scouts;

[0010] S4. During the iterative optimization process, the position update strategy is executed as follows: the position of the discoverer is updated using an adaptive step size mechanism based on population diversity, the position of the follower is updated using a ring topology-aware neighborhood reconstruction mechanism, and the position of the scout is updated using a phased multi-strategy weighted collaborative mechanism.

[0011] S5. Determine whether the preset termination condition is met. If it is met, output the optimal collision-free path.

[0012] Furthermore, the discrete initialization of the fourth-order Chebyshev chaotic map in step S2 is as follows: for the first... Dimension, filter all free nodes in this dimension, and construct the first dimension. 3D free node sequence The free node sequence Defined as the first in a raster map Under the corresponding column, it is an ordered set consisting of the coordinates of all grid rows that are unobstructed and safe for mobile robots to pass through. Indicates the sequence number of the first... The free node in the th _ Coordinates in the dimension Indicates the first The number of free nodes;

[0013] Set the initial value of chaos Based on the fourth-order Chebyshev chaotic iterative formula, a length of [length missing] is generated. chaotic sequence N is the population size. The transient discard length is used to eliminate the transient effects of the initial segment of the chaotic sequence. The chaotic sequence is calculated. The formula is as follows:

[0014] ;

[0015] Before discarding Each transient value retains a valid sequence segment. The starting index of this sequence is The terminating index is This ensures that each individual obtains a unique and valid initial chaotic value in the current dimension; subsequently, a discrete index mapping formula is used. Calculate index value, constraints Ensure the index is valid and legal; This represents the number of nodes in the path encoding.

[0016] The initial path position is assigned based on the mapping index, and the first... The individual in the first The initial position of dimension is , For the first After chaotic mapping and discretization, each individual is used to determine the index value of the initial position of the path node. By traversing all individuals in the population and the path dimension, the initial feasible path population is constructed.

[0017] Furthermore, in step S4, an adaptive step-size mechanism based on population diversity is used to update the discoverer's position; normalized population diversity Δ is used as the driving indicator, specifically including:

[0018] Randomly select from the current population For each individual, each pair contains two distinct individuals, denoted as individuals. and individuals (in ), calculate its average Hamming distance and normalize it to obtain the population diversity measure Δ, as shown in the following formula:

[0019]

[0020]

[0021] in, For indicator functions, For the first For individuals, individuals Path solution vector dimensional coordinates, To indicate the first For individuals, individuals Path solution vector The coordinates of the dimension.

[0022] The exploration factor α and the safety threshold θ are determined by a linear mapping Δ, and the calculation formula is as follows:

[0023] α = clip(0.2+1.6Δ, 0.2, 1.8)

[0024] θ = clip(0.3+0.6Δ, 0.3, 0.9)

[0025] Where clip(v,lo,hi)=max(lo,min(hi,v)); α directly scales the step size of all discrete move operators, and θ determines the probability of selecting the discoverer update mode.

[0026] Furthermore, the ring topology-aware neighborhood reconstruction mechanism in step S4 updates the follower positions, specifically including:

[0027] Based on the ascending fitness sort sequence σ=[σ(1),σ(2),…,σ(N)], a single-node circular chain neighborhood topology is constructed. For a follower individual at sort position j, its circular neighbor index is defined as:

[0028]

[0029] Where σ(N) represents the individual with the worst fitness. Number of discoverers; Let N be the proportion of discoverers and N be the population size. Using the population median H = ⌊N / 2⌋ as the dividing threshold, a differentiated update strategy is implemented for follower individuals at different ranking positions. The specific update method is as follows:

[0030] When the follower's ranking position j>H, a combined repulsion-attraction movement is performed: the individual's position is updated by combining the repulsion force from moving away from the worst solution in the current population and the attraction force from moving closer to the leader in the circular topological neighborhood, in order to avoid local clustering and maintain the directed search trend. The update formula is:

[0031]

[0032] in, and The updated and current individuals are respectively in Position number in the middle; and The worst-fitting individual and their circular neighbors are respectively. exist Position number in the middle; For sign function; repulsion strength With attraction strength The calculation formula is:

[0033] R(t)=max(1, round(3(1−t / T)))

[0034] A(t) = max(1, round(2t / T))

[0035] in, This represents the current iteration number. This represents the maximum number of iterations. The individual with the worst fitness in the population in the 1st month 3D free node sequence Position number in the middle; The circular neighbor individuals corresponding to the current individual In the 3D free node sequence Position number in the middle;

[0036] When the follower individuals are ranked At that time, the strategy of moving towards the guidance center is executed, and the specific steps are as follows:

[0037] Calculate the discretized guiding center: ;

[0038] Adaptive step size calculation formula:

[0039] Update individual position: Drive the current individual along the first... 3D free node sequence Towards the center move One location;

[0040] in, This represents the position vector of the current follower individual; It is the position vector of its corresponding neighbor within its circular topological neighborhood; This is a discretized rounding function; This represents the number of dimensions in the search space.

[0041] Furthermore, the phased multi-strategy weighted collaborative mechanism for updating the scout's location in step S4 specifically includes:

[0042] Randomly select n = max(2, round(0.1N)) individuals as scouts, where N is the population size;

[0043] If the historical best fitness of a scout is worse than the global best fitness value f, then the scout moves in a direction to the global best individual position x, with a step size s=max(1,round(3α)), where α is the exploration factor described in claim 3;

[0044] Otherwise, perform a diversity walk: randomly select a reference individual from the current population, and move randomly along the direction of the positional difference between the scout and the reference individual, with a step size of max(1,round(2α)).

[0045] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a mobile robot path planning method based on an improved sparrow search algorithm as described in any one of the claims.

[0046] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements a mobile robot path planning method based on an improved sparrow search algorithm as described in any one of the claims.

[0047] The advantages and beneficial effects of this invention are as follows:

[0048] This invention provides a mobile robot path planning method based on an improved sparrow search algorithm, aiming to solve the problems of low population initialization quality, lack of topological constraints, parameter rigidity, and susceptibility to local optima in traditional algorithms. Specific improvement strategies include: For the population initialization step in claim 2, a fourth-order Chebyshev chaotic mapping combined with free node index constraints is used to filter invalid obstacle nodes, eliminating population aggregation blind spots through the synergy of mathematical ergonomics and physical feasibility; For the parameter control step in claim 3, an adaptive threshold and exploration factor synergistic control mechanism based on population diversity is designed to dynamically balance global exploration and local development capabilities, solving the problems of parameter rigidity and susceptibility to local optima; For the neighborhood reconstruction step in claim 4, a ring-shaped topological neighborhood reconstruction mechanism is proposed, and modulo operations are used to match neighboring high-quality individuals as guiding benchmarks, reducing invalid cross-obstacle updates without the need for distance thresholds and connectivity detection, highly adaptable to grid obstacle avoidance scenarios; For the scout update step in claim 5, a phased weighted collaborative scout update strategy is constructed, balancing the convergence speed of inferior individuals with the maintenance of the diversity of high-quality individuals, significantly improving the optimization stability in complex environments. Attached Figure Description

[0049] Figure 1 This is a flowchart of a preferred embodiment of the present invention for a mobile robot path planning method based on an improved sparrow search algorithm;

[0050] Figure 2 This is a comparison chart of the average path length convergence curves of the algorithm provided in this embodiment of the invention and other algorithms in the mobile robot path planning task;

[0051] Figure 3 This is a comparison chart of the path planning results of the algorithm provided in this embodiment of the invention and other algorithms in the same grid obstacle environment. Detailed Implementation

[0052] The technical solutions of the present invention will be clearly and thoroughly described below with reference to the accompanying drawings. The described embodiments are merely some embodiments of the present invention.

[0053] The technical solution of the present invention to solve the above-mentioned technical problems is:

[0054] like Figure 1 As shown in the figure, this embodiment of the invention provides a mobile robot path planning method based on an improved sparrow search algorithm, the steps of which are as follows:

[0055] S1. Obtain environmental map information, construct a gridded search space model, and obtain gridded map information of the mobile robot's operating environment, including:

[0056] Starting coordinates End point coordinates Including the distribution of obstacles, a resolution of [resolution value missing] is constructed. Binary raster map (Where 0 represents a free area and 1 represents an obstacle area). In this embodiment, the map resolution is set to [resolution value missing]. ,Right now .

[0057] For the search space Dimension, filter all free nodes in this dimension, and construct the first dimension. 3D free node sequence ,in, Indicates the sequence number of the first... The free node in the th _ Coordinates in the dimension Indicates the first The number of free nodes.

[0058] S2, Population initialization based on fourth-order Chebyshev chaotic mapping: In order to solve the problems of strong randomness and uneven distribution in traditional population initialization, this embodiment uses fourth-order Chebyshev chaotic mapping to complete population initialization, thereby improving the ergodicity and diversity of the initial population.

[0059] Set the initial value of chaos Based on the fourth-order Chebyshev chaotic iterative formula, a length of [length missing] is generated. chaotic sequence Among them, the preset population size Transient discard length This is used to eliminate the transient effects of the initial segment of a chaotic sequence, and the calculation formula is as follows:

[0060]

[0061] Before discarding Each transient value retains a valid sequence segment. The starting index of this sequence is The terminating index is This ensures that each individual obtains a unique and valid initial chaotic value in the current dimension. The index value is then calculated using the discrete index mapping formula:

[0062]

[0063] and constraints To ensure the index is valid and effective. Assign the initial path position based on the mapped index, and then... The individual in the first The initial position of dimension is:

[0064]

[0065] Traverse all path dimensions With population individuals Complete the construction of all initial feasible path populations.

[0066] S3, calculate the fitness of all individuals and sort them in ascending order, divide them into discoverers and followers according to a preset ratio, and randomly select individuals as scouts.

[0067] Let decision variables be defined. This is a sequence of ordinates for intermediate nodes along the path, with the x-coordinates assigned at fixed equal intervals. First, apply boundary truncation constraints to the decoded nodes: Ensure that the node falls on the grid map. Within the valid range. After splicing the start and end points, a complete path row coordinate sequence is obtained. .

[0068] Count the number of intermediate nodes (excluding the start and end points) that fall on obstacle grids along the path. :

[0069] ;

[0070] Where I(·) is an indicator function, which takes the value of 1 when the condition in parentheses is true, and takes the value of 0 otherwise.

[0071] fitness value Calculated according to the following tiered strategy:

[0072] (1) Feasibility path evaluation ): Determine if there are no collisions on the path, and then execute the continuous path generation function sequentially on the original discrete nodes. With double smoothing function To obtain a smooth continuous path The fitness value is calculated using the sum of the Euclidean distances of the smoothed paths, as shown in the following formula:

[0073]

[0074] Where represents the original discrete path coordinate sequence obtained by decoding the decision variables; represents the continuous path generation function, used to connect discrete grid points into geometrically continuous line segments and perform transition processing; This represents a double smoothing function used to perform curve fitting on continuous paths to eliminate sharp corners; This represents the final executable trajectory obtained after continuumization and smoothing. Represents a smooth path The total number of discrete sampling points included; , These represent the first and second parts of the smooth path, respectively. The x and y coordinates of each sampling point.

[0075] (2) Penalty for infeasible paths A dynamic penalty strategy is employed to ensure that the fitness value of any infeasible path is strictly greater than the maximum possible length of all feasible paths. The penalty formula is:

[0076]

[0077] in, This is an enlarged upper bound on the theoretically maximum possible path length. The design aims to prioritize the elimination of illegal paths during iteration, while avoiding excessively large penalty values ​​that could lead to numerical overflow. , These represent the number of rows and columns in the raster map, respectively.

[0078] The fitness values ​​are calculated using the fitness function described above and sorted in ascending order. The top 20% of individuals with the best fitness values ​​are designated as discoverers, responsible for global exploration; the middle 70% of individuals are designated as followers, relying on high-quality individuals for local optimization; and 10% of individuals are randomly selected from the entire population as scouts.

[0079] S4 employs three update strategies during the iteration process: an adaptive step-size mechanism based on population diversity to update the discoverer's position, a ring-shaped topological neighborhood reconstruction mechanism to update followers, and a phased multi-strategy weighted collaborative mechanism to update scouts. The specific process is as follows:

[0080] For individual discoverers, an adaptive step-size mechanism based on population diversity is used to update their positions. Random samples are randomly drawn from the current population. For individuals, the normalized population diversity Δ is obtained by calculating the average Hamming distance and then normalizing it. The formulas for calculating Δ are as follows:

[0081]

[0082]

[0083] in Here, d represents the characteristic function, d represents the dimension of the search space, and the exploration factor represents the characteristic function. With safety threshold Depend on The linear mapping is determined by the following formula: α = clip(0.2+1.6Δ, 0.2, 1.8)

[0084] θ = clip(0.3+0.6Δ, 0.3, 0.9)

[0085] The exploration factor α is used to uniformly scale the iteration step size of all discrete movement operators, and the safety threshold θ is used to control the selection probability of the discoverer's update mode, realizing adaptive parameter coordination based on the real-time state of the population. For individual followers, a ring topology-aware neighborhood reconstruction mechanism is used to complete the position update. A single-node ring chain neighborhood topology structure is constructed based on the ascending order of the population fitness sequence σ=[σ(1),σ(2),…,σ(N)], and the ring neighbor index of the j-th follower is:

[0086]

[0087] With the population median As a threshold, differentiated position updates are performed on followers: When a follower's ranking position j > H, a combined repulsion-attraction movement is performed. The individual position is updated by a combination of the repulsion force moving away from the worst solution and the attraction force moving closer to the neighboring leader.

[0088]

[0089] in, and The updated and current individuals are respectively in Position number in the middle; and The worst-fitting individual and their circular neighbors are respectively. exist Position number in the middle; For sign function. Repulsion strength. With attraction strength Adaptive adjustment with iteration:

[0090]

[0091]

[0092] The repulsive effect is weakened by decreasing the value with each iteration. Iterative increments enhance neighborhood attraction, achieving a dynamic balance between exploration and development.

[0093] When the follower's sorting position At that time, a strategy of moving towards the guidance center is executed to achieve local fine-grained optimization.

[0094] The first step is to calculate the discretized guidance center:

[0095]

[0096] The second step is to calculate the adaptive step size:

[0097]

[0098] The third step is to update the individual's position: drive the current individual along the d-th dimension of the free node sequence. Move s positions toward the guide center c to complete the position iteration update.

[0099] in, This represents the position vector of the current follower individual; It is the position vector of its corresponding neighbor within its circular topological neighborhood; This is a discretized rounding function; This represents the number of dimensions in the search space.

[0100] For individual scouts, a phased, multi-strategy weighted collaborative mechanism is adopted to complete the location update. First, n = max(2, round(0.1N)) individuals are randomly selected from the population as scouts, where N is the population size, to ensure that the number of scouts adaptively matches the population size and stably maintains the population monitoring and diversity regulation capabilities.

[0101] Based on the relationship between the individual scout's historical best fitness and the population's global best fitness, differentiated position updates are performed in two modes:

[0102] When the historical best fitness of a scout is worse than the global best fitness value f of the population, it is determined that the search quality of the individual is poor and it needs to converge quickly to a high-quality search area. The scout is then controlled to move in a directional direction towards the position of the global best individual. The iteration step size is calculated as s=max(1,round(3α)), where α is the aforementioned adaptive update exploration factor, so as to realize the step size dynamically adapts to the iteration state of the population.

[0103] When the scout's historical best fitness is greater than or equal to the population's global best fitness value f, the individual is deemed to have high search potential, and a diversity walk strategy is initiated. A reference individual is randomly selected from the current population, and random steps are taken along the direction of the positional difference between the scout and the reference individual, with a step size upper limit of max(1,round(2α)). This effectively expands the algorithm's search space, continuously maintains population diversity, and avoids the problem of the algorithm getting stuck in local optima in the later stages of iteration.

[0104] S5 determines whether the maximum number of iterations has been reached. If so, the globally optimal collision-free path is output. The maximum number of iterations, T=100, is set as the algorithm's iteration termination condition. Termination is determined after each update of population roles, fitness calculation, and adaptive parameter adjustment.

[0105] If the current iteration count reaches the preset maximum iteration count, the algorithm terminates the iteration and outputs the globally optimal collision-free smooth path; if the termination condition is not met, it returns to the adaptive parameter update stage to continue iterative optimization, and executes in a loop until the termination condition is met.

[0106] To verify the effectiveness and superiority of the proposed ISSA algorithm, a simulation experimental environment was built based on the MATLAB platform. The experiment used a 30×30 resolution binary raster map with multiple irregular obstacles. The starting coordinates were (1,1) and the ending coordinates were (30,30). The algorithm's parameters were fixed: population size N=50 and maximum number of iterations T=100. To evaluate the statistical stability of the algorithm, all algorithms were run independently and repeatedly 10 times. Genetic Algorithm (GA), Standard Sparrow Search (SSA), Particle Swarm Optimization (PSO), Differential Evolution (DE), and Grey Wolf Optimization (GWO) were selected as benchmarks for comparison. All comparison algorithms were executed under the same map environment, parameter configuration, and hardware conditions.

[0107] The following table shows the path length statistics for 10 independent runs of the six algorithms in the aforementioned grid map environment: Table 1. Statistical results of the path length of 10 independent runs of each algorithm

[0108]

[0109] As shown in Table 1, the average path length of the ISSA algorithm is 42.011 grid cells, which is 8.70% shorter than the best DE algorithm (46.012 grid cells) and 11.64% shorter than the standard SSA algorithm (47.547 grid cells). In terms of operational stability, the standard deviation of ISSA's path length is only 0.188, far lower than SSA's 4.487 and other comparative algorithms (all with standard deviations greater than 2.4), indicating that the improved strategy proposed in this invention can effectively suppress solution fluctuations caused by random initialization and local search, demonstrating significant robustness. Furthermore, the longest path length of ISSA (42.545 grid cells) is still better than the shortest path length of all comparative algorithms, proving that it can still stably output high-quality paths even in the worst-case scenario. This verifies the effectiveness and engineering practical value of the technical solution of this invention in complex obstacle environments.

[0110] Combination Figure 2 The convergence curves show that the ISSA algorithm decreases at a significantly faster rate than the comparison algorithms such as GA and SSA in the early stages of iteration, and does not experience any plateaus in the middle and late stages. It eventually converges stably to the minimum average path length, which verifies the improvement of the strategy's ability to balance global exploration and local development as well as its global optimization ability.

[0111] Combination Figure 3The path planning results show that, in the same grid obstacle environment, the path generated by the ISSA algorithm maintains a safe obstacle avoidance distance, has smooth turns, and has no redundant detours throughout the entire process, demonstrating the optimization effect of this invention in terms of path quality and executability.

[0112] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions.

[0113] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0114] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0115] The above embodiments should be understood as illustrative only and not as limiting the scope of protection of the present invention. After reading the description of the present invention, those skilled in the art can make various alterations or modifications to the present invention, and these equivalent changes and modifications also fall within the scope defined by the claims of the present invention.

Claims

1. A mobile robot path planning method based on an improved sparrow search algorithm, characterized in that, Includes the following steps: S1, acquire environmental map information and construct a rasterized search space model; S2 uses fourth-order Chebyshev chaotic mapping for discrete initialization to generate the initial path population X; S3, calculate the fitness value of each individual in the population and sort them in ascending order, determine the discoverers and followers according to a preset ratio, and randomly select a number of individuals as scouts; S4. During the iterative optimization process, the position update strategy is executed as follows: the position of the discoverer is updated using an adaptive step size mechanism based on population diversity, the position of the follower is updated using a ring topology-aware neighborhood reconstruction mechanism, and the position of the scout is updated using a phased multi-strategy weighted collaborative mechanism. S5. Determine if the maximum number of iterations has been reached. If so, output the optimal collision-free path.

2. The mobile robot path planning method based on the improved sparrow search algorithm according to claim 1, characterized in that, The discrete initialization of the fourth-order Chebyshev chaotic map in step S2 is as follows: for the first... Dimension, filter all free nodes in this dimension, and construct the first dimension. 3D free node sequence The free node sequence Defined as the first in a raster map Under the corresponding column, it is an ordered set consisting of the coordinates of all grid rows that are unobstructed and allow safe passage for mobile robots; where, Indicates the sequence number of the first... The free node in the th _ Coordinates in the dimension Indicates the first The number of free nodes; Set the initial value of chaos Based on the fourth-order Chebyshev chaotic iterative formula, a length of [length missing] is generated. chaotic sequence N is the population size. The transient discard length is used to eliminate the transient effects of the initial segment of the chaotic sequence. The chaotic sequence is calculated. The formula is as follows: ; Before discarding Each transient value retains a valid sequence segment. The starting index of this sequence is The terminating index is This ensures that each individual obtains a unique and valid initial chaotic value in the current dimension; subsequently, a discrete index mapping formula is used. Calculate index value, constraints Ensure the index is valid and legal; This represents the number of nodes in the path encoding; The initial path position is assigned based on the mapping index, and the first... The individual in the first The initial position of dimension is , For the first After chaotic mapping and discretization, each individual is used to determine the index value of the initial position of the path node. By traversing all individuals in the population and the path dimension, the initial feasible path population is constructed.

3. The mobile robot path planning method based on the improved sparrow search algorithm according to claim 1, characterized in that, In step S4, normalized population diversity is used. To adjust the adaptive step size factor to drive the indicator, and to update the discoverer position based on the adaptive step size factor, the following is specifically included: Randomly select from the current population For each individual, each pair contains two distinct individuals, denoted as individuals. and individuals (in ), calculate its average Hamming distance and normalize it to obtain the population diversity measure Δ, as shown in the following formula: ; ; in, For indicator functions, For the first For individuals, individuals Path solution vector dimensional coordinates, To indicate the first For individuals, individuals Path solution vector Dimensional coordinates; The exploration factor α and the safety threshold θ are determined by a linear mapping Δ, and the calculation formula is as follows: α = clip(0.2+1.6Δ, 0.2, 1.8); θ = clip(0.3+0.6Δ, 0.3, 0.9); Where clip(v,lo,hi)=max(lo,min(hi,v)); α directly scales the step size of all discrete move operators, and θ determines the probability of selecting the discoverer update mode.

4. The mobile robot path planning method based on the improved sparrow search algorithm according to claim 1, characterized in that, The ring topology-aware neighborhood reconstruction mechanism in step S4 updates the follower positions, specifically including: A single-node circular chain neighborhood topology is constructed based on the fitness ascending sort sequence σ=[σ(1),σ(2),…,σ(N)]. For a follower individual at sort position j, its circular neighbor index is defined as: ; Where σ(N) represents the individual with the worst fitness. For the number of discoverers; The population size is N, representing the proportion of discoverers. Using the population median H = ⌊N / 2⌋ as the dividing threshold, a differentiated update strategy is implemented for follower individuals at different ranking positions. The specific update method is as follows: When the follower's ranking position j>H, a combined repulsion-attraction movement is performed: the individual's position is updated by combining the repulsion force from moving away from the worst solution in the current population and the attraction force from moving closer to the leader in the circular topological neighborhood, in order to avoid local clustering and maintain the directed search trend. The update formula is: ; in, and The updated and current individuals are respectively in Position number in the middle; and The worst-fitting individual and their circular neighbors are respectively. exist Position number in the middle; For sign function; repulsion strength With attraction strength The calculation formula is: R(t)=max(1, round(3(1−t / T))) A(t) = max(1, round(2t / T)) in, This represents the current iteration number. This represents the maximum number of iterations. The individual with the worst fitness in the population in the 1st month 3D free node sequence Position number in the middle; The circular neighbor individuals corresponding to the current individual In the 3D free node sequence Position number in the middle; When the follower individuals are ranked When this happens, the strategy of moving towards the guidance center is executed, and the specific steps are as follows: Calculate the discretized guiding center: ; Adaptive step size calculation formula: ; Update individual position: Drive the current individual along the first... 3D free node sequence Towards the center move One location; in, This represents the position vector of the current follower individual; It is the position vector of its corresponding neighbor within its circular topological neighborhood; This is a discretized rounding function; This represents the number of dimensions in the search space.

5. The mobile robot path planning method based on the improved sparrow search algorithm according to claim 1, characterized in that, The phased multi-strategy weighted collaborative mechanism for updating the scout's location in step S4 specifically includes: Randomly select n = max(2, round(0.1N)) individuals as scouts, where N is the population size; If the historical best fitness of a scout is worse than the global best fitness value f, then the scout moves in a direction to the global best individual position x, with a step size s=max(1,round(3α)), where α is the exploration factor described in claim 3; Otherwise, perform a diversity walk: randomly select a reference individual from the current population, and move randomly along the direction of the positional difference between the scout and the reference individual, with a step size of max(1,round(2α)).

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the mobile robot path planning method based on the improved sparrow search algorithm as described in any one of claims 1 to 5.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the mobile robot path planning method based on the improved sparrow search algorithm as described in any one of claims 1 to 5.

Citation Information

Patent Citations

  • Spherical amphibious robot path planning method based on improved sparrow search algorithm

    CN116795098A