Adaptive Genetic Algorithm Based Particle Swarm Optimization for Industrial Robotic Arm Obstacle Avoidance Trajectory Optimization

Abstract

In this paper, we propose an obstacle avoidance algorithm, which selects a point of the obstacle avoidance path as the chromosome, constructs the fitness function together with the path length, joint angle increment, and movement time as evaluation indexes, and performs scale transformation on the fitness to improve the competitiveness of the population. The algorithm cycles through the process of optimizing the velocity term in the chromosome in the first step with a particle swarm algorithm; selection in the second step; and crossover and mutation operations on individuals in the third step, in order to avoid the population falling into premature maturity, where the crossover and mutation probabilities vary adaptively with the results of the previous generation. The final smooth and continuous obstacle avoidance trajectory is obtained.