An improved genetic algorithm based robot path planning method without collision in confined workspace

Abstract

The manufacturing process of ships and marine engineering equipment involves a large number of welding processes of complex components. It is difficult for industrial robots to plan the welding path because of the complex distribution of the weld seams and environmental obstacles as well as the specific technological requirements of arc welding process. In this paper, the typical welding components are numerically modelled. Considering the joint collision-free constraint in the moving process of a robot and the welding technological constraints (e.g., welding direction) of some special welding seams, an improved Re-GA algorithm for planning a robot welding operation path in a complex workspace is proposed, and an optimal welding path can be obtained. The simulation results show that the Re-GA has reliable optimisation ability in the case of a small population size.