Research on simulation of motion compensation for omnidirectional platform based on neural network

Abstract

In order to improve the accuracy of omnidirectional platform, this paper makes an analysis of kinematics and builds the dynamic virtual prototyping. The anisotropy of slippage and its influence on translation is simulated by virtual experiment. A method of neural network is proposed for nonlinear motion compensation. Simulation results show that motion error can be reduced and the performance is greatly improved, which indicates that the method is effective and feasible. © 2011 Springer-Verlag.