A mobile robot path planning method based on GSO particle swarm algorithm
By adopting the grouping and competition update strategy of GSO particle swarm algorithm in mobile robot path planning, the parameter dependence and local minimum value problems of particle swarm optimization algorithm are solved, and more efficient path planning and diversity exploration are achieved, reducing the Local optimal risk.
Patent Information
- Application Number
- CN202410494862.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-23
- Publication Date
- 2024-12-20
- Estimated Expiration
- 2044-04-23
AI Technical Summary
In the path planning of mobile robots, existing particle swarm optimization algorithms have problems such as excessive dependence on parameters of search performance and easy to fall into local minimum values.
Using the GSO particle swarm algorithm, particles are grouped and divided into hierarchical parts, and particle updates are updated using competitive methods and differential evolution strategies. The winner and middle-capable particles are used to guide the learning of losers, so as to enhance the diversity and exploration ability of particles. .
It effectively reduces the path length, enhances the algorithm exploration ability, reduces the risk of falling into local optimality, and improves performance in large-scale optimization problems.
Smart Images
Figure CN118394084B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of mobile robot path planning, and in particular relates to a mobile robot path planning method based on a GSO particle swarm algorithm. Background Art
[0002] With the development of artificial intelligence technology, 5G technology, robotics and other technologies, mobile robots are widely used in deep-sea exploration, automobile manufacturing, medical research 3C electronics, tobacco industry, logistics industry, energy collection, military modernization and other fields, and their application prospects are broad. Among them, the path planning problem of mobile robots is of great significance to the realization of full intelligence of mobile robots.
[0003] Among the algorithms for mobile robot paths, currently used evolutionary algorithms, neural network algorithms, particle swarm optimization algorithms and other methods are widely used in real-time obstacle avoidance and trajectory control. Among them, the particle swarm optimization algorithm has no requirements on the properties of the objective function, is easy to implement and has good optimization performance, and has gradually become a hot topic in the current research of swarm intelligence optimization methods. The mobile robot path planning modeling based on the particle swarm optimization algorithm is simple, and the algorithm calculation difficulty is low, but there are problems such as excessive dependence of the search performance on parameters and easy to fall into local minima.
[0004] In order to find an effective learning method, the inventors looked for inspiration from human society. Among them, in the education process, different students have different potentials, and group cooperative learning strategies are widely used to improve their learning ability. Similarly, in the evolutionary process, particles have different abilities in exploring and developing search spaces. Therefore, they should also be treated differently. In addition, by further studying the particle swarm optimization algorithm, it is found that these optimization algorithms have better performance in dealing with high-dimensional optimization problems compared with the classical particle swarm optimization algorithm. The main reason is that these optimizers mainly use particles in the current group to update particles, rather than using the global optimal particles to guide the learning of particles, so the diversity of particles is enhanced, and these optimizers perform well in dealing with high-dimensional problems.
[0005] Inspired by the above observations and phenomena, this paper proposes a population-based particle swarm learning strategy, which divides different particles into several groups, treats them differently, and uses two dominant particles in the current group and a differential evolution strategy to guide the learning of particles to find the global optimal solution. It can be applied to the path planning problem of mobile robots in large-scale scenarios. Summary of the invention
[0006] The purpose of the present invention is to provide a mobile robot path planning method based on the GSO particle swarm algorithm to solve the problems of the existing particle swarm optimization algorithm proposed in the above background technology, such as the search performance is too dependent on parameters and is easy to fall into local minima.
[0007] To achieve the above object, the present invention adopts the following technical solutions:
[0008] A first aspect of the present invention provides a mobile robot path planning method based on a GSO particle swarm algorithm, comprising the following steps:
[0009] S1. Constructing the objective function: establishing a coordinate system according to the working environment of the mobile robot, using the grid method to model the working environment space, and taking the shortest path function as the objective function;
[0010] S2. Calculate the path points of the shortest path through the GSO particle swarm algorithm; replace the mobile robot with the initialized particles, and use the GSO particle swarm algorithm to optimize and update the particles to obtain the shortest path for the mobile robot path planning;
[0011] The GSO particle swarm algorithm optimizes and updates particles, which specifically includes the following steps:
[0012] S201, grouping particles and classifying particles in the group by using a competitive method;
[0013] S202, according to the level classification results, different update strategies are adopted to update the particles. The winner particles directly enter the next round of iteration, the medium-ability particles are updated in this round by random evolution, and the loser particles are updated in this round by learning from the winner particles and the medium-ability particles;
[0014] S203, iterating in a loop until convergence, the position of the global optimal particle is the optimal path point, and the shortest path is obtained according to the optimal path point;
[0015] S3, smoothing the shortest path; performing cubic spline interpolation adjustment on the shortest path obtained by particle optimization iteration to obtain a smooth shortest path.
[0016] Preferably, the S1 is as follows:
[0017] Taking the shortest path as the optimization goal, a plane rectangular coordinate system is established with the starting point as the origin and the target point as the end point of the X-axis, and the grid method is used to divide the working environment space for modeling. The size of the grid is determined according to the size of the mobile robot itself. The total number of grids is D, and there are D-1 grids from the starting point to the target point; establish the objective function.
[0018] Preferably, the objective function is established as follows:
[0019]
[0020] Among them, L is the path length and also the objective function ;( y G , 0) is the target point coordinate, D is the total number of grids, x d and x d-1 For the The X-axis coordinates corresponding to the grid and the -1 The X-axis coordinate corresponding to the grid.
[0021] Preferably, the S201 is specifically as follows:
[0022] All particles are randomly divided into groups of three and the objective function corresponding to each particle in the group is calculated. The values are classified into different levels according to the objective function. The values are divided into winners, medium-ability and losers from large to small; the set of all winners is W , the group of people with medium ability is L1 , the loser set is L2 ,use w , l1 and l2 Represent any element in the winner set, the medium-ability set, and the loser set respectively.
[0023] Preferably, the medium-capability particles in S202 are updated in this round by random evolution, as follows:
[0024]
[0025] in, and yes l1 The result after particle crossover mutation, factor F is the positive control parameter as weight, Cr represents the crossover rate, and determines The number of mutant individuals; is the objective function of optimization. rand(0,1) generates a random number between 0 and 1, and controls the probability of random evolution by comparing it with the crossover rate Cr.
[0026] Preferably, the loser particles in S202 are updated in this round by learning from the winner particles and the particles with medium capabilities, as follows:
[0027]
[0028] in, , , Representing three levels of particles in each group of particles The location at the moment, Indicates loser The speed of time ψ For control l1 Influencing capacity parameters; r 1. r 2. r 3 is a control parameter, which takes a random value in [0,1].
[0029] Preferably, the use of a grid method to divide and model the working environment space also includes: performing an expansion process on the obstacles, and the expansion spacing is determined by the size of the mobile robot itself.
[0030] Preferably, the shortest path is obtained according to the optimal path point in S203, specifically as follows:
[0031] The detection of whether there is an intersection between the shortest path and the expanded obstacle is used as the criterion for judging collision; if a collision is detected, the path point will be reselected and planned; if no collision occurs, the shortest path will be confirmed and output.
[0032] The second aspect of the invention provides a mobile robot, including a memory, a processor, and an executable program stored in the memory and running on the processor, wherein the processor implements the above-mentioned mobile robot path planning method based on the GSO particle swarm algorithm when executing the program.
[0033] Compared with the prior art, the present invention has the following beneficial effects:
[0034] (1) In order to reduce the length of the mobile robot's travel path, the present invention proposes a navigation path planning method based on the GSO particle swarm algorithm. A path planning model is established with the goal of reducing the path length; in order to prevent the robot from colliding, an obstacle expansion processing method is given. The particles are randomly grouped into groups of three, and the particles in the group are graded through competition, and different update strategies are adopted for particles of different grades. Finally, the optimal path points are obtained according to the particle optimization iteration, and the path points are interpolated three times to update the path curve to a smoother path. The present invention increases the diversity of particles, thereby enhancing the exploration ability of the algorithm.
[0035] (2) The GSO particle swarm algorithm used in the present invention first increases the information exchange between particles by grouping, and at the same time increases the diversity of particles by adopting a competitive hierarchical learning scheme, wherein different update strategies are adopted to update particles according to the hierarchical division results, and the winner particles maintain their advantages and directly enter the next round of iterations, and the medium-capability particles adopt differential evolution to increase the diversity of particles, and the loser particles learn from the two advantage particles instead of the global optimal particles, further increasing the diversity of particles. While maintaining the mining ability of the algorithm, the present invention greatly increases the exploration ability of the algorithm, and effectively reduces the risk of falling into the local optimum when solving large-scale optimization problems. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Figure 1 is a flow chart of a mobile robot path planning method based on GSO particle swarm algorithm in the present invention;
[0037] Figure 2 A schematic diagram of the environmental space modeling in Embodiment 1 of the present invention;
[0038] Figure 3 Schematic diagram of the particle update strategy of the GSO particle swarm algorithm in the present invention. DETAILED DESCRIPTION
[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0040] Embodiment 1:
[0041] See also Figure 1 , a mobile robot path planning method based on GSO particle swarm algorithm, including the following steps:
[0042] Step 1: Determine the coordinate system according to the starting position and the environment space, and use the grid method to model the environment space;
[0043] Taking the shortest path as the optimization goal, establish the environmental coordinate system, take the starting point as the origin and the target point as the end point of the X-axis to establish a plane rectangular coordinate system, and use the grid method to divide the space. The size of the grid is determined according to the size of the robot itself. The total number of grids is D. There are D-1 grids from the starting point to the target point, such as Figure 2 shown.
[0044] The objective function is:
[0045]
[0046] Among them, L is the path length and also the objective function ;( y G , 0) is the target point coordinate, D is the total number of grids, x d and x d-1 For the The X-axis coordinates corresponding to the grid and the -1 The X-axis coordinate corresponding to the grid.
[0047] Furthermore, in order to treat the robot as a particle for path planning, the obstacles are expanded, and the expansion spacing is determined by the size of the robot itself to avoid collisions.
[0048] Step 2: Calculate the path points of the shortest path through the GSO particle swarm algorithm;
[0049] The initialized particles are used to replace the mobile robot, and the particle swarm optimization algorithm is used to optimize and update the particles to obtain the optimal mobile robot planning path map.
[0050] Specifically, all particles are first randomly divided into groups of three, and the objective function corresponding to the particles in the group is calculated. Then the medium-capable particles and losers are evolved according to the position update formula. The winners directly enter the next round of iterations as winners. The iterations are repeated until convergence. The position of the global optimal particle is the shortest path.
[0051] The position update formula is as follows:
[0052] like Figure 3 As shown, express The collection of all particles at a given moment. At this moment, after all groups are divided into different levels, all winners are gathered as follows: W , the group of people with medium ability is L1 , the loser set is L2 .use w , l1 and l2 Represent any element in the winner set, the medium-ability set, and the loser set respectively.
[0053] For losers l2 For the player, it needs to learn from the winner and the medium-capable players, so its position and velocity update formula is:
[0054]
[0055] In the above formula , , Representing three levels of particles in each group of particles The location at the moment, Indicates loser The speed of time ψ For control l1 Influencing capacity parameters; r 1. r 2. r 3 is the control parameter, which takes a random value in [0,1];
[0056] For those with medium ability l1 , random evolution method is used to increase the diversity of particles, and the update formula is:
[0057]
[0058] in, and yes l1 The result after particle crossover mutation, factor F is the positive control parameter as weight, Cr represents the crossover rate, and determines The number of mutant individuals. It is an optimization objective function problem. rand(0,1) generates a random number between 0 and 1, and controls the probability of random evolution by comparing it with the crossover rate Cr.
[0059] In order to maintain the advantage of the algorithm's mining ability, the winner directly enters the next round of iteration with the current position until convergence. The corresponding function The particle position with the minimum value is the path point for the shortest path search.
[0060] The collision is determined by the intersection between the path and the enlarged obstacle. If a collision is detected, the system will reselect the path points and plan; if no collision occurs, the optimal path will be confirmed and output.
[0061] Step 3: Make the robot path smoother by using a cubic spline interpolation function on the path points; this is to ensure that the robot can move to the target position with a smooth and efficient trajectory while avoiding obstacles.
[0062] The above description is only used to help understand the method of the present invention and its core essence, but the protection scope of the present invention is not limited thereto. For those skilled in the art in the art, equivalent replacement or change according to the technical solution and inventive concept of the present invention within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. In summary, the content of this specification should not be understood as limiting the present invention.
Claims
1. A mobile robot path planning method based on GSO particle swarm algorithm, characterized in that: The steps include: S1. Constructing the objective function: establishing a coordinate system according to the working environment of the mobile robot, using the grid method to model the working environment space, and taking the shortest path function as the objective function; The objective function of the construction is as follows: Where L is the path length and also the objective function f(x); (y G ,0) is the target point coordinate, D is the total number of grids, x d and x d-1 is the X-axis coordinate corresponding to the d-th grid and the X-axis coordinate corresponding to the d-1-th grid; S2. Calculate the path points of the shortest path through the GSO particle swarm algorithm; replace the mobile robot with the initialized particles, and use the GSO particle swarm algorithm to optimize and update the particles to obtain the shortest path for the mobile robot path planning; The GSO particle swarm algorithm optimizes and updates particles, which specifically includes the following steps: S201, grouping particles and classifying particles in the group by using a competitive method; All particles are randomly divided into groups of three, and the objective function f(x) value corresponding to each particle in the group is calculated, and the levels are divided; the levels are divided into winners, medium-ability and losers in descending order according to the objective function f(x) value; the set of all winners is W, the set of medium-ability is L1, and the set of losers is L2. w, l1 and l2 are used to represent any element in the winner set, the medium-ability set and the loser set respectively; S202, according to the level classification results, different update strategies are adopted to update the particles. The winner particles directly enter the next round of iteration, the medium-ability particles are updated in this round by random evolution, and the loser particles are updated in this round by learning from the winner particles and the medium-ability particles; The particles of medium ability are updated in this round by random evolution, as follows: M l1 =X w (t)+F*(X l1 (t)-X l2 (t)), Among them, M l1 and C l1 is the result of the crossover mutation of l1 particles. The factor F is a positive control parameter, which is the weight. Cr represents the crossover rate. l1 f(x) is the objective function of optimization, rand(0,1) generates a random number between 0 and 1, and controls the probability of random evolution by comparing it with the crossover rate Cr; X w (t), X l2 (t), X l1 (t) represents the position of particles of the three levels of each group of particles at time t; The loser particles learn from the winner particles and the medium-capability particles to perform this round of updates, as follows: V l2 (t)=r1V w (t)+r2(X w (t)-X l2 (t))+ψr3(X l1 (t)-X l2 (t)), X l2 (t+1)=X l2 (t)+V l2 (t) Among them, X w (t), X l2 (t), X l1 (t) represents the position of the three levels of particles in each group of particles at time t, V l2 (t) represents the loser’s speed at time t. ψ is the influencing capacity parameter for controlling l1; r1, r2, r3 are control parameters, which are randomly selected in [0,1]; S203, iterating in a loop until convergence, the position of the global optimal particle is the optimal path point, and the shortest path is obtained according to the optimal path point; S3, smoothing the shortest path; performing cubic spline interpolation adjustment on the shortest path obtained by particle optimization iteration to obtain a smooth shortest path.
2. A mobile robot path planning method based on GSO particle swarm algorithm according to claim 1, characterized in that: The S1 is specifically as follows: Taking the shortest path as the optimization goal, a plane rectangular coordinate system is established with the starting point as the origin and the target point as the end point of the X-axis, and the grid method is used to divide the working environment space for modeling. The size of the grid is determined according to the size of the mobile robot itself. The total number of grids is D, and there are D-1 grids from the starting point to the target point; establish the objective function.
3. A mobile robot path planning method based on GSO particle swarm algorithm according to claim 2, characterized in that: The method of dividing and modeling the working environment space by using the grid method also includes: performing expansion processing on obstacles, and the expansion spacing is determined by the size of the mobile robot itself.
4. A mobile robot path planning method based on GSO particle swarm algorithm according to claim 3, characterized in that: In S203, the shortest path is obtained according to the optimal path point, as follows: The detection of whether there is an intersection between the shortest path and the expanded obstacle is used as the criterion for judging collision; if a collision is detected, the path point will be reselected and planned; if no collision occurs, the shortest path will be confirmed and output.
5. A mobile robot, characterized in that: It comprises a memory, a processor and an executable program stored in the memory and running on the processor, wherein when the processor executes the program, a mobile robot path planning method based on a GSO particle swarm algorithm as described in any one of claims 1 to 4 is implemented.