New variants of nature inspired metaheuristics for trajectory optimization of industrial robots

Abstract

Multi-objective trajectory optimization of industrial robots is an emerging research area. Trajectory optimization problems are non-parametric hard optimization problems. All trajectory optimization problems are tough to solve. No algorithm gives 100% best solution to a real world trajectory optimization problem. So, the efficiency of present algorithms has to be improved. Lot of effort is taken by many researchers in improving the existing algorithms or introducing new variants of present algorithms. Majority of researchers used nature-inspired phenomena (evolutionary and swarm intelligence) based techniques for trajectory optimization of robots. Major limitations of present industrial robots trajectory optimization research works are a. Only few researchers done experimental validation b. Efficiency and performance of multi objective techniques (swarm intelligence and evolutionary techniques) have to be increased c. Metrics used for comparing multi objective optimization algorithms are to be improved d. Best trajectory defining function is to be usedTo overcome above limitations, this research work aims to improve the efficiency of few nature-based metaheuristic algorithms namely DE and PSO. Five new variants of these algorithms are developed and used. They are tested in real industrial robots applications. They have been used for multi-objective trajectory optimization of industrial robots. The important contributions of this research work newline