Research On Welding Path Planning Of Automobile Door Panel Based On Improved Particle Swarm Optimization

The automobile manufacturing industry is an important pillar industry of China's economy,and auto parts are an important part of the automobile industry.As the trend of domestic consumption upgrades rises,the speed of car product replacement will continue to accelerate,and China's auto parts manufacturing industry will face greater challenges.At present,China's auto parts industry,especially the field of welding of internal and external decorative parts,is still produced in a simple manual workshop style,and a small number of special machines or robots are used.Path planning is a key step for welding robots to complete production tasks.Therefore,studying the path planning of welding robots is of great significance for improving the production quality of auto parts.First of all,the three key factors in the welding process of automotive interior door panel are introduced: welding object,ultrasonic welding technology and welding robot.Meanwhile the overall structure and specific work process of the automotive interior door panel welding production line are introduced.Secondly,the particle swarm optimization algorithm is selected and improved as the path planning algorithm of the welding robot.According to the task characteristics of the automobile interior door welding production line,a variety of path planning algorithms are compared and analyzed,and the particle swarm optimization algorithm is found to meet the requirements of the production line.Next,aiming at the problems that the classical particle swarm optimization algorithm is easy to fall into local optimization and has the disadvantage of low solution accuracy,multiple optimization strategies such as hybrid population initialization strategy,elite particle strategy,comprehensive operator are proposed,and redefine the calculation model of "position-velocity",thus getting an modified discrete particle swarm optimization algorithm(MDPSO).Based on typical test examples in TSPLIB,MDPSO is compared with other algorithms to verify the effectiveness and superiority of MDPSO.Finally,the welding path planning problem of single and multiple robots collaborative are studied.The kinematics of the robot is briefly analyzed,and the kinematic equations of therectangular coordinate welding robot are established.Then,the front and rear door panels of the car interior are used as the welding objects,combined with the path planning requirements of a single welding robot,a mathematical model of the path planning problem is established and simulation experiments are carried out.The problem of multi-robot collaborative welding path planning is analyzed and introduced,and the mathematical models of welding point assignment and anti-interference between robots are established.Welding point allocation strategy based on dynamic adjustment k-means algorithm and anti-interference strategy based on safety distance are proposed,therefore the overall flow of multi-robot collaborative welding path planning algorithm is obtained.Taking four robots as examples,the simulation verifies the feasibility and effectiveness of the algorithm.