Welding robot path planning method fusing global information enhancement and pheromone suppression strategy improved ant colony algorithm

Abstract

The improved ant colony algorithm integrating global information enhancement and pheromone suppression strategies is provided based on the problems existing in a traditional ant colony algorithm and used for solving the path planning problem of the welding robot, the actual contribution of nodes to the path quality is comprehensively considered by introducing the global information enhancement strategy, and the path planning precision of the welding robot is improved. The guiding effect of high-quality nodes on ant path finding is improved, and then the search quality of the algorithm is improved. Meanwhile, the pheromone suppression strategy enables the algorithm to find out inferior fragments from historical information, and interference of inferior nodes on ant path finding is reduced. Simulation results show that the GIEPI-IACO algorithm has a remarkable effect on solving the welding robot path planning problem, and the overall performance of the GIEPI-IACO algorithm is superior to that of a basic ant colony algorithm and an existing improved ant colony algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of robot path planning, and in particular to a welding robot path planning method integrating a global information enhancement and pheromone inhibition strategy improved ant colony algorithm (Global Information Enhancement-based Improved Ant Colony Optimization with Pheromone Inhibition, GIEPI-IACO). Background Art

[0002] Path planning technology plays a vital role in the field of welding robot research. Its main goal is to find an optimal or suboptimal path that efficiently connects all welding nodes without collision in a complex three-dimensional space based on a series of preset standards (e.g., shortest path, highest safety, shortest time, etc.). This technology is not only an important part of the welding automation process, but also a symbolic indicator for measuring the intelligence level of welding robots.

[0003] With the development of technology, the method and efficiency of path planning directly determine the success rate and production efficiency of welding tasks. An excellent path planning algorithm can ensure the efficient operation of welding robots in complex environments. Through precise calculation and sophisticated design, it not only shortens the operation time, but also improves safety, further promoting the advancement of welding technology and the development of the industry. Therefore, the research and optimization of path planning technology not only reflects the continuous pursuit of welding robot capabilities, but also directly contributes to the improvement of the intelligence level of the entire manufacturing industry. Through continuous exploration and practice, researchers and engineers in this field are committed to developing more efficient, intelligent and safe path planning solutions, laying a solid foundation for the development of welding robot technology.

[0004] At present, many experts and scholars at home and abroad are committed to the research of path planning algorithms. Commonly used optimization algorithms include genetic algorithms, simulated annealing algorithms, ant colony optimization algorithms, neural networks, particle swarm algorithms, etc.

[0005] Among them, the Ant Colony Algorithm (ACO) is a population-based probabilistic selection algorithm, which is inspired by the pheromones left by ants when looking for food and their communication behavior. It simulates the process of ants looking for food to continuously search and update the path to find the optimal solution. Compared with other commonly used algorithms in path planning, the ant colony algorithm can be applied to the optimization of various problems and has better parallel computing capabilities and stronger robustness, but the traditional ant colony algorithm still has some defects that need to be improved. First, the distance heuristic function in the traditional ant colony algorithm only considers the distance between the current node and the next node, and does not consider the actual contribution of the next node to the overall path quality from a global perspective. Secondly, the concentration of pheromones released by ants is only related to the length of the entire path, and cannot accurately identify the poor quality segments in the path, which reduces the efficiency of ant pathfinding. Summary of the invention

[0006] The purpose of the present invention is to provide an improved ant colony algorithm (Global Information Enhancement-based Improved Ant Colony Optimization with Pheromone Inhibition, GIEPI-IACO) that integrates global information enhancement and pheromone inhibition strategies for path planning of welding robots. This method can overcome the problem that the distance heuristic function in the traditional ant colony algorithm only considers the benefit of the next step selection, but cannot globally measure the overall contribution of the node to the path, and ants cannot accurately find the defects of inferior fragments through pheromones.

[0007] In the iterative process of the traditional ant colony algorithm, ants mainly use formula 1 to calculate the probability of selecting the next feasible node in the path-finding process, and then substitute the roulette method to make decisions. As can be seen from formula 1, ants can only judge the quality of the path by the pheromone concentration on the path and the distance between the current node and the candidate node, but the distance heuristic function is only related to the distance between the current node and the candidate node, and cannot consider the contribution of the candidate node to the overall path from a global perspective.

[0008]

[0009] in, is the selection probability of the kth ant moving from node i to node j at iteration number t, τ ij (t) is the pheromone concentration heuristic function, η ij (t) is the distance heuristic function, α and β are the distance heuristic factor and the pheromone concentration heuristic factor, respectively, allowed k is a list of feasible nodes.

[0010] likeFigure 1 As shown in the figure, suppose the ant starts from point A and first chooses point B. Then the ant will choose one of the remaining unselected solder joints as the next node. Since the distance heuristic function in the traditional ant colony algorithm is only related to the distance between the current node and the node to be selected, and the closest node to point B is point C, the ant will be more inclined to choose point C. However, from a global perspective, the ant chooses point E as the next node when it is at point B, which is as follows Figure 2 The length of the path shown is shorter. Based on the above problem, the present invention proposes a global information enhancement strategy to improve the algorithm's use of global information.

[0011] After each round of iteration of the algorithm, the paths obtained by the ants are analyzed and the influence of each node is scored. The influence of each node is related to the actual path length and the frequency of its optimal solution. Nodes with higher scores have a higher probability of being selected.

[0012] The new probability selection formula is shown in Formula 2.

[0013]

[0014] Where γ is the influence weight, E ij (t) is the influence score from node i to node j, and the calculation formula is shown in Formula 3.

[0015]

[0016] where d ij is the Euclidean distance between node i and node j; l ij is the shortest length of the path formed by the edge connecting node i and node j. Its initial value is the same as d ij The same, the update formula is shown in Formula 4, PathLength(i, j) is the length set of the path containing edge (i, j), and the min function takes the minimum value among them; the update formula is shown in Formula 5, e ij is the contribution index, and its initial value is 0.

[0017] l ij =min(PathLength(i,)) (4)

[0018]

[0019] In addition, in the traditional ant colony algorithm, each ant will leave pheromones on the path it has walked, but the ant can only release pheromones of different concentrations according to the overall length of the path it has walked, and cannot distinguish between high-quality and low-quality areas in the path, which causes the ant's path-finding process to be disturbed by low-quality paths. In view of the above defects, the present invention proposes a pheromone suppression strategy to enhance the ant's ability to distinguish the quality of the path.

[0020] Figure 3 The analysis process of inferior paths is intuitively presented. The historical optimal solution and the contemporary optimal solution memorize the previous and subsequent iteration information respectively. Figure 3 In the figure, the edge represented by the red dashed line represents the difference path between the historical optimal path and the contemporary optimal path. Further converted into a sequence form, the blue block (or dark block) represents the difference path sequence, and the white block represents the common path sequence. Figure 3 In the case shown, it is obvious that the difference path length in the historical optimal path sequence is larger, so the path segment is a low-quality segment. The algorithm will suppress the pheromone on the low-quality segment to prevent the ants from being disturbed by the low-quality segments during the pathfinding process. The new pheromone update formula is shown in Formula 6 and Formula 7.

[0021] τ ij (t+1)=(1-ρ)·τ ij (t)+Δτ ij (t)+Δ-τ ij (t) (6)

[0022]

[0023] Among them, Δτ ij (t) is the pheromone increment left by all ants on the path from node i to node j in the tth iteration; Δ - τ ij (t) is the increment of inhibitory pheromone on the inferior fragment; ρ is the volatility coefficient; Q is a numerical constant used to control the pheromone concentration left by each ant; L k Represents the length of the solution found by the kth ant.

[0024] The technical solution adopted by the present invention to solve its technical problem is:

[0025] A welding robot path planning method based on Global Information Enhancement-based Improved Ant Colony Optimization with Pheromone Inhibition (GIEPI-IACO) includes the following steps:

[0026] S1. Create a robot working environment map;

[0027] S2, initialize the node influence score matrix E, distance heuristic factor α, γ influence weight, number of ants num, number of iterations T, volatility coefficient ρ and other algorithm parameters;

[0028] S3, algorithm iteration begins:

[0029] S3.1, put num ants to the starting point to start path finding, the ants search for paths according to the node probability selection formula shown in Formula 2, record all path finding results, and save path node information and path length information;

[0030]

[0031] Where γ is the influence weight, E ij (t) is the influence score from node i to node j, and the calculation formula is shown in Formula 3;

[0032]

[0033] where d ij is the Euclidean distance between node i and node j, l ij is the shortest length of the path formed by the edge connecting node i and node j. Its initial value is the same as d ij The same, update formula is shown in formula 4,

[0034] l ij =min(PathLength(i, j)) (4)

[0035] Where PathLength(i, j) is the set of lengths of the paths containing edge (i, j), the min function takes the minimum value among them, and the update formula is shown in Formula 5;

[0036]

[0037] where e ij is the contribution index, and its initial value is 0;

[0038] S3.2. Update the node influence score matrix E according to formulas 3, 4, and 5;

[0039] S3.3, update the pheromone concentration matrix according to formulas 5 and 6;

[0040] τ ij (t+1)=(1-ρ)·τ ij (t)+Δτ ij (t)+Δ - τ ij (t) (6)

[0041] S3.4, if the current number of iterations is less than the maximum number of iterations, execute S3.1; otherwise, execute S4;

[0042] S4: End the loop after T generations and output the global optimal solution.

[0043] The flowchart of the improved GIEPI-IACO algorithm proposed in the present invention is as follows Figure 4 shown.

[0044] The beneficial effect of the present invention is that by introducing a global information enhancement strategy, the actual contribution of nodes to path quality is comprehensively considered, the guiding role of high-quality nodes in ant pathfinding is improved, and the search quality of the algorithm is improved. At the same time, the pheromone suppression strategy enables the algorithm to find low-quality fragments from historical information and reduce the interference of low-quality nodes on ant pathfinding. The simulation results show that the GIEPI-IACO algorithm of the present invention is effective in solving the path planning problem of welding robots, and its overall performance is better than the basic ant colony algorithm and the existing improved ant colony algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The present invention is further described below in conjunction with the accompanying drawings and embodiments:

[0046] Figure 1 Schematic diagram of suboptimal paths;

[0047] Figure 2 Optimal path diagram;

[0048] Figure 3 The inferior path analysis process includes the historical optimal length, contemporary optimal length, historical optimal path, contemporary optimal path and inferior path;

[0049] Figure 4 Flow chart of the GIEPI-IACO algorithm of the present invention;

[0050] Figure 5 Traditional ant colony algorithm simulation results;

[0051] Figure 6 Algorithm simulation results after introducing the global information enhancement strategy;

[0052] Figure 7 The simulation results of the GIEPI-IACO algorithm of the present invention;

[0053] Figure 8 Simulation results of the GIEPI-IACO algorithm of the present invention. DETAILED DESCRIPTION

[0054] The purpose of the present invention is to provide a global information enhancement and pheromone inhibition strategy improved ant colony algorithm (Global Information Enhancement-based Improved Ant Colony Optimization with Pheromone Inhibition, GIEPI-IACO) to solve the path planning problem of welding robots. This method can overcome the problem that the distance heuristic function in the traditional ant colony algorithm cannot truly reflect the node quality and the ants cannot accurately find the inferior fragments through pheromone, thereby improving the efficiency of the algorithm in path planning.

[0055] The present invention provides an improved ant colony algorithm integrating global information enhancement and pheromone suppression strategy, comprising the following steps:

[0056] S1. Create a robot working environment map;

[0057] S2, initialize the node influence score matrix E, distance heuristic factor α, γ influence weight, number of ants num, number of iterations T, volatility coefficient ρ and other algorithm parameters;

[0058] S3, algorithm iteration begins:

[0059] S3.1, put num ants to the starting point to start path finding, the ants search for paths according to the node probability selection formula shown in Formula 2, record all path finding results, and save path node information and path length information;

[0060]

[0061] Where γ is the influence weight, E ij (t) is the influence score from node i to node j, and the calculation formula is shown in Formula 3;

[0062]

[0063] where d ij is the Euclidean distance between node i and node j, l ij is the shortest length of the path formed by the edge connecting node i and node j. Its initial value is the same as d ij The same, update formula is shown in formula 4,

[0064] l ij =min(PathLength(i,j)) (4)

[0065] Where PathLength(i, j) is the set of lengths of the paths containing edge (i, j), the min function takes the minimum value among them, and the update formula is shown in Formula 5;

[0066]

[0067] where e ij is the contribution index, and its initial value is 0;

[0068] S3.2. Update the node influence score matrix E according to formulas 3, 4, and 5;

[0069] S3.3, update the pheromone concentration matrix according to formulas 5 and 6;

[0070] τ ij (t+1)=(1-ρ)·τ ij (t)+Δτ ij (t)Δ - τ ij (t) (6)

[0071] S3.4, if the current number of iterations is less than the maximum number of iterations, execute S3.1; otherwise, execute S4;

[0072] S4: End the loop after T generations and output the global optimal solution.

[0073] In order to verify the feasibility of the global information enhancement strategy and pheromone suppression strategy proposed in this invention, Python language programming is used to simulate the actual welding task data of a welding enterprise as the test set. The simulation parameters are shown in Table 1, and the welding point coordinate data are shown in Table 2. The simulation results are shown in Table 1. Figure 5 , Figure 6 , Figure 7 And as shown in Table 3.

[0074] Table 1 Algorithm simulation parameters

[0075]

[0076] Table 2 Solder point coordinate data

[0077]

[0078] Table 3 Comparison of simulation data of two algorithms

[0079]

[0080]

[0081] Through a detailed analysis of the simulation results, it can be seen that although the traditional ant colony algorithm can find a near-optimal solution, namely 1973.77mm, this process requires 79 iterations. In contrast, the ant colony algorithm that introduces the global information enhancement strategy successfully locates the suboptimal solution of 1951.68mm in only 20 iterations. After adding the pheromone suppression strategy on this basis, the algorithm only needs 18 iterations to find the optimal solution of 1942.72mm. Although the integration of these two strategies leads to a slight increase in the execution steps of the algorithm, thereby slightly improving the running time, the algorithm shows significant advantages over the traditional ant colony algorithm in terms of speed and stability in finding the optimal solution. Based on this, we can conclude that the improved strategy proposed in this paper is effective in improving the performance of the algorithm.

[0082] In order to further verify the effectiveness of the improved algorithm proposed in the present invention, the present invention is compared with another improved IACO algorithm. The other improved IACO algorithm is the improved IACO algorithm recorded in the 2021 master's thesis "Research on Collision-free Path and Trajectory Optimization of Welding Robots Based on Virtual Simulation". In the first section of the third chapter of the article, the method of the present invention is used to simulate the weld point data set of the simulation experiment of the IACO in the article, and the experimental results are compared with the literature data, such as Figure 8 And as shown in Table 4.

[0083] Table 4 Comparison of simulation data of three algorithms

[0084]

[0085] By analyzing the simulation result data, it can be seen that the optimal path result obtained by the GIEPI-IACO algorithm of the present invention is 8127.3381mm, which is better than the result of 9456.1193mm obtained by the traditional ant colony algorithm and the result of 8234.0100mm obtained by the IACO algorithm in the literature. Although the IACO algorithm in the literature first obtained the optimal solution in 100 generations, which is significantly better than the traditional ACO algorithm, the GIEPI-IACO algorithm of the present invention only needs 10 iterations to find a better solution. Therefore, the improved ant colony algorithm of the present invention is better than the traditional ACO algorithm and the IACO algorithm in the literature in terms of optimization effect and search speed, which shows that the GIEPI-IACO algorithm proposed in the present invention has certain feasibility and practicality in the path planning of welding robots.

[0086] The above is only a preferred embodiment of the present invention, and does not limit the present invention in any form; any technician familiar with the art can make many possible changes and modifications to the technical solution of the present invention by using the above disclosed methods and technical contents without departing from the scope of the technical solution of the present invention, or modify it into an equivalent embodiment of equivalent changes. Therefore, any simple modification, equivalent replacement, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of protection of the technical solution of the present invention.

Claims

1. A welding robot path planning method that integrates global information enhancement and pheromone suppression strategy to improve ant colony algorithm, characterized in that The following steps are involved: S1. Create a robot working environment map; S2, initialize the node influence score matrix E, distance heuristic factor α, γ influence weight, number of ants num, number of iterations T, volatility coefficient ρ and other algorithm parameters; S3, algorithm iteration begins: S3.1, put num ants to the starting point to start path finding, the ants search for paths according to the node probability selection formula shown in Formula 2, record all path finding results, and save path node information and path length information; Where γ is the influence weight, E ij (t) is the influence score from node i to node j, and the calculation formula is shown in Formula 3; where d ij is the Euclidean distance between node i and node j, l ij is the shortest length of the path formed by the edge connecting node i and node j. Its initial value is the same as d ij The same, update formula is shown in formula 4, l ij =min(PathLength(i,j)) (4) Where PathLength(i, j) is the set of lengths of the paths containing edge (i, j), the min function takes the minimum value among them, and the update formula is shown in Formula 5; where e ij is the contribution index, and its initial value is 0; S3.

2. Update the node influence score matrix E according to formulas 3, 4, and 5; S3.3, update the pheromone concentration matrix according to formulas 5 and 6; t ij (t+1)=(1-ρ)·τ ij (t)+Δτ ij (t)+Δ-τ ij (t) (6) S3.4, if the current number of iterations is less than the maximum number of iterations, execute S3.1; Otherwise, execute S4; S4: End the loop after T generations and output the global optimal solution.

Images