Multi-objective-based configuration generation and optimization for reconfigurable modular robot

Abstract

A new method for reconfigurable modular robot on configuration auto-generation, optimization and evaluation was proposed in this paper. On the basis of assign and design of the module, a new descriptive method of configuration-CCM was proposed, and it can describe the configuration directly and completely which needed to satisfy the assembly requirement. Genetic-simulated annealing algorithm has been used for multi-objective optimization of configuration in the accessibility and the optimality of joint variables. Through the 3D dynamic simulation of configuration by Robotics Toolbox, the overall optimality was proved.