Body-in-white welding path planning method based on nonlinear particle swarm optimization

Abstract

The invention discloses a body-in-white welding path planning method based on a nonlinear particle swarm algorithm. The body-in-white welding path planning method comprises the following steps: constructing a mathematical model for path planning of a welding robot; initializing a population of a particle swarm algorithm; the fitness value of each particle is calculated, and a global optimal solution gbest and a local optimal solution Pbest are initialized; updating the positions and speeds of the particles; whether the limitation of the distance between every two continuous welding spots is met or not is judged, if yes, the fitness value of the current particle is calculated, or if not, the fitness value is increased, and the global optimal solution and the local optimal solution are updated; whether the number of iterations is exceeded or not is judged, and if yes, the optimal welding planning path is output; if not, updating the inertia weight, and returning to the step S4 to continue execution. According to the method, the welding time can be effectively shortened, the welding efficiency is improved, the welding deformation is reduced, and a stable, efficient and engineering-feasible robot welding path planning solution is provided for complex welding scenes such as a body in white.

Description

Technical Field

[0001] The present invention belongs to the technical field of automobile body-in-white welding, and in particular relates to a body-in-white welding path planning method based on a nonlinear particle swarm algorithm. Background Art

[0002] The body-in-white (BIW) refers to the body of a vehicle after welding but before painting. BIW welding primarily involves various welding methods, including resistance spot welding, MIG / MAG arc welding, welding, and stud welding. With the rapid development of artificial intelligence and robotics, many automakers are adopting robots for BIW welding to improve welding quality and efficiency. Currently, BIW welding suffers from defects in welding path planning methods, such as local optima and insufficient search accuracy, leading to low welding efficiency and deformation. Summary of the Invention

[0003] In order to solve the above technical problems existing in the prior art, the present invention provides a body-in-white welding path planning method based on nonlinear particle swarm algorithm to improve production efficiency and welding quality.

[0004] To solve the above technical problems, the present invention adopts the following technical solution: a body-in-white welding path planning method based on nonlinear particle swarm optimization algorithm, comprising the following steps:

[0005] S1. Construct a mathematical model for welding robot path planning;

[0006] S2, using the distance d between each two consecutive welding points = {c i ,c j}≥D limit, initialize the population of the particle swarm algorithm;

[0007] S3, calculate the fitness value fitness(i) of each particle and initialize the global optimal solution g best and the local optimal solution P best ;

[0008] S4, update the position and velocity of the particle;

[0009] S5, determine the distance between each two consecutive welding points d = {c i ,c j If the constraint of}≥D is satisfied, the fitness value of the current particle is calculated; if not, the fitness value is increased and the global optimal solution g is updated. best and the local optimal solution P best ;

[0010] S6. Determine whether the number of iterations is exceeded. If exceeded, output the optimal welding planning path; if not exceeded, update the inertia weight ω and return to step S4 to continue execution.

[0011] Furthermore, the specific process of step S1 is as follows:

[0012] For the N welding point tasks C={c1,c2,…,c n}, the distance between each two consecutive welding points is d={c i ,c j}≥D, where c i ,c j ∈C(1≤i,j≤N), D is the threshold of the distance between two consecutive welding points, and the total path for the robot to complete the welding task is:

[0013]

[0014] In formula (1), L is the shortest welding path length, d(c i ,c i+1 ) represents the distance between the i-th and i+1-th welding points, d(c n ,c1) represents the distance between the last solder point and the initial solder point, specifically expressed as:

[0015]

[0016] Design a new fitness penalty function as follows:

[0017]

[0018] In formula (4), dist(c i ,c i+1 ) represents the distance between two consecutive welding points, π is the welding sequence, dis(c i ,π i ) is the distance between solder point i and other solder points;

[0019] The fitness function is designed as:

[0020]

[0021] Where L(π) represents the welding path length, f(Σdis(c i ,c i+1 )) is the penalty function value.

[0022] Furthermore, the population of the particle swarm algorithm initialized in step S2 includes the population size M, the particle swarm dimension W, the maximum number of iterations T, the learning factor C1, the learning factor C2, the maximum position Xmax, the maximum position -Xmax, the minimum speed Vmax, the minimum speed -Vmax, the position x of each particle i and the velocity v of each particle i .

[0023] Furthermore, the specific process of step S4 is as follows:

[0024] The rules for updating particle positions and velocities are:

[0025]

[0026] In formula (6), ω is the inertia weight, r1 and r2 are random numbers between [0,1], and p i is the individual optimal value, p g is the global optimal value, j=1,2,…,W, t is the current iteration number;

[0027] The larger the inertia weight value, the stronger the global search ability; conversely, the search ability is weakened and it is easy to converge to the optimal solution. The traditional particle swarm algorithm has poor search performance due to the fixed inertia weight, and is prone to falling into the local optimum and premature convergence. Based on the weight strategy, a nonlinear weight dynamic adjustment mechanism is proposed as follows:

[0028]

[0029] Where: ω z is the middle value of the inertia weight, ω min is the minimum value of the inertia weight, ω max is the maximum value of the inertia weight; t is the number of iterations t z is the middle value of the number of iterations, t min is the minimum number of iterations, t max is the maximum number of iterations.

[0030] By adopting the above technical solution, compared with the existing technology, the present invention has the following beneficial effects:

[0031] In order to improve the calculation speed of the algorithm, the welding path planning is transformed into a three-dimensional traveling salesman problem (TSP). Based on the traditional particle swarm optimization algorithm, a nonlinear weight dynamic adjustment mechanism is established to effectively enhance the local search capability and suppress the algorithm oscillation, which solves the defects of the traditional method that is prone to falling into local optimality and insufficient search accuracy. A weld spot spacing constraint model is constructed, and a new fitness penalty function is designed to dynamically eliminate particles whose adjacent weld spot spacing is less than a safety threshold, ensuring that the path planning meets the actual welding parameter requirements of the robot.

[0032] This paper addresses the need for spot welding path planning for complex three-dimensional body-in-white (BIW) welds by proposing a BIW welding path planning method based on a nonlinear particle swarm algorithm. This method's core innovation lies in the establishment of a nonlinear weight dynamic adjustment mechanism and the design of a novel fitness penalty function. This method effectively addresses the shortcomings of traditional particle swarm algorithms, such as poor local search capabilities, low search accuracy, prone to local optimality, and prone to oscillation in the late stages of the algorithm. This method can effectively shorten welding time, improve welding efficiency, and reduce welding deformation, providing a stable, efficient, and engineering-feasible robotic welding path planning solution for complex welding scenarios such as BIW. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0034] The embodiments of the present invention are described in further detail below with reference to the accompanying drawings and examples.

[0035] like Figure 1 As shown, the body-in-white welding path planning method based on nonlinear particle swarm optimization of the present invention includes the following steps:

[0036] S1. Construct a mathematical model for welding robot path planning;

[0037] S2, using the distance d between each two consecutive welding points = {c i ,c j}≥D limit, initialize the population of the particle swarm algorithm;

[0038] S3, calculate the fitness value fitness(i) of each particle and initialize the global optimal solution g best and the local optimal solution P best ;

[0039] S4, update the position and velocity of the particle;

[0040] S5, determine the distance between each two consecutive welding points d = {c i ,c j If the constraint of}≥D is satisfied, the fitness value of the current particle is calculated; if not, the fitness value is increased and the global optimal solution g is updated. best and the local optimal solution P best ;

[0041] S6. Determine whether the number of iterations is exceeded. If exceeded, output the optimal welding planning path; if not exceeded, update the inertia weight ω and return to step S4 to continue execution.

[0042] Furthermore, the specific process of step S1 is as follows:

[0043] For the N welding point tasks C={c1,c2,…,c n}, the distance between each two consecutive welding points is d={c i ,c j}≥D, where c i ,c j ∈C(1≤i,j≤N), D is the threshold of the distance between two consecutive welding points, and the total path for the robot to complete the welding task is:

[0044]

[0045] In formula (1), L is the shortest welding path length, d(c i ,c i+1 ) represents the distance between the i-th and i+1-th welding points, d(c n ,c1) represents the distance between the last solder point and the initial solder point, specifically expressed as:

[0046]

[0047] Design a new fitness penalty function as follows:

[0048]

[0049] In formula (4), dist(c i ,c i+1 ) represents the distance between two consecutive welding points, π is the welding sequence, dis(c i ,π i ) is the distance between solder point i and other solder points;

[0050] The fitness function is designed as:

[0051]

[0052] Where L(π) represents the welding path length, f(∑dis(c i ,c i+1 )) is the penalty function value.

[0053] Furthermore, the population of the particle swarm algorithm initialized in step S2 includes the population size M, the particle swarm dimension W, the maximum number of iterations T, the learning factor C1, the learning factor C2, the maximum position Xmax, the maximum position -Xmax, the minimum speed Vmax, the minimum speed -Vmax, the position x of each particle i and the velocity v of each particle i .

[0054] Furthermore, the specific process of step S4 is as follows:

[0055] The rules for updating the position and velocity of particles are:

[0056]

[0057] In formula (6), ω is the inertia weight, r1 and r2 are random numbers between [0,1], and p i is the individual optimal value, p g is the global optimal value, j=1,2,…,W, t is the current iteration number;

[0058] The larger the inertia weight value, the stronger the global search ability; conversely, the search ability is weakened and it is easy to converge to the optimal solution. The traditional particle swarm algorithm has poor search performance due to the fixed inertia weight, and is prone to falling into the local optimum and premature convergence. Based on the weight strategy, a nonlinear weight dynamic adjustment mechanism is proposed as follows:

[0059]

[0060] Where: ω z is the middle value of the inertia weight, ω min is the minimum value of the inertia weight, ω max is the maximum value of the inertia weight; t is the number of iterations t z is the middle value of the number of iterations, t min is the minimum number of iterations, t max is the maximum number of iterations.

[0061] The above embodiments illustrate the basic principles and features of the present invention, but the above only illustrates preferred embodiments of the present invention and is not limited to the embodiments. Under the guidance of this patent, those skilled in the art can make many variations and improvements without departing from the scope of the present invention and the scope of protection of the claims, all of which fall within the scope of protection of the present invention. Therefore, the patent and protection scope of the present invention shall be based on the appended claims.

Claims

1. A body-in-white welding path planning method based on nonlinear particle swarm optimization is characterized by: The following steps are involved: S1. Construct a mathematical model for welding robot path planning; S2, using the distance d between each two consecutive welding points = {c i ,c j }≥D limit, initialize the population of the particle swarm algorithm; S3, calculate the fitness value fitness(i) of each particle and initialize the global optimal solution g best and the local optimal solution P best ; S4, update the position and velocity of the particle; S5, determine the distance between each two consecutive welding points d = {c i ,c j If the constraint of}≥D is satisfied, the fitness value of the current particle is calculated; if not, the fitness value is increased and the global optimal solution g is updated. best and the local optimal solution P best ; S6. Determine whether the number of iterations is exceeded, and if so, output the optimal welding planning path; If it does not exceed, the inertia weight ω is updated and the process returns to step S4 to continue.

2. The method for body-in-white welding path planning based on nonlinear particle swarm optimization according to claim 1, characterized in that: The specific process of step S1 is: For the N welding point tasks C={c1,c2,…,c n }, the distance between each two consecutive welding points is d={c i ,c j }≥D, where c i ,c j ∈C(1≤i,j≤N), D is the threshold of the distance between two consecutive welding points, and the total path for the robot to complete the welding task is: In formula (1), L is the shortest welding path length, d(c i ,c i+1 ) represents the distance between the i-th and i+1-th welding points, d(c n ,c1) represents the distance between the last solder point and the initial solder point, specifically expressed as: Design a new fitness penalty function as follows: In formula (4), dist(c i ,c i+1 ) represents the distance between two consecutive welding points, π is the welding sequence, dis(c i ,π i ) is the distance between solder point i and other solder points; The fitness function is designed as: Where L(π) represents the welding path length, f(∑dis(c i ,c i+1 )) is the penalty function value.

3. The method for body-in-white welding path planning based on nonlinear particle swarm optimization according to claim 2, characterized in that: The population of the particle swarm algorithm initialized in step S2 includes the population size M, the particle swarm dimension W, the maximum number of iterations T, the learning factor C1, the learning factor C2, the maximum position Xmax, the maximum position -Xmax, the minimum speed Vmax, the minimum speed -Vmax, the position x of each particle i and the velocity v of each particle i .

4. The method for body-in-white welding path planning based on nonlinear particle swarm optimization according to claim 3, characterized in that: The specific process of step S4 is: The rules for updating the position and velocity of particles are: In formula (6), ω is the inertia weight, r1 and r2 are random numbers between [0,1], and p i is the individual optimal value, p g is the global optimal value, j=1,2,…,W, t is the current iteration number; The larger the inertia weight value, the stronger the global search ability; conversely, the search ability is weakened and it is easy to converge to the optimal solution. The traditional particle swarm algorithm has poor search performance due to the fixed inertia weight, and is prone to falling into the local optimum and premature convergence. Based on the weight strategy, a nonlinear weight dynamic adjustment mechanism is proposed as follows: Where: ω z is the middle value of the inertia weight, ω min is the minimum value of the inertia weight, ω max is the maximum value of the inertia weight; t is the number of iterations, t z is the middle value of the number of iterations, t min is the minimum number of iterations, t max is the maximum number of iterations.

Images