Path Optimization of Gluing Robot Based on Improved Genetic Algorithm

Abstract

The steel cord conveyor belt surface is prone to damage in mining. The worn belt surface has acceleration characteristics, so timely and rapid repair is very necessary. To quickly and automatically repair the worn belt surface is a core design objective of the gluing robot (GR). Based on this objective, a new variant Traveling Salesman Problem (TSP) is put forward: after the worn segments are divided according to the worn information and GR’s workspace, path optimization of the gluing robot (POGR) problem is presented at a certain worn segment; then the POGR is simplified into a “double vertices” TSP problem by Hamilton graph, and the mathematical model is built. An improved genetic algorithm (IGA) is proposed to handle the POGR problem, which is called IGA-POGR. The main benefit of the proposed IGA-POGR is the ability to solve POGR of different scales in different ways. The performance of the IGA-POGR is illustrated on four well-known TSP problems. Numerical results show that IGA-POGR does not give any deviation (0%) from the optimal solution. Compared with discrete particle swarm optimization (DPSO), IGA-POGR has better performance in terms of the solving quality and time consumption when solving four idealized POGR problems.