A New Rapid Neural Network Algorithm and Simulation Research for 6-DOF Parallel Platform

Abstract

The 6-DOF parallel platform has a strong carrying capacity, good structural stability. But it also has problems such as high cost, complexity of controlling and so on. The traditional process of forward solutions for 6-DOF parallel platform is complex and there is no fixed formula for solving this kind of nonlinear problems. The commonly used method is numerical analysis method which is solving slowly and has a large amount of calculation. Although neural networks and fuzzy control can be very good at nonlinear mapping, its calculating speed will affect real-time control and simulation results. Therefore, this paper based on existing neural network calculation method, proposing a new method for calculating the best number of neural network’s hidden layer’s nodes and the bubbling contraction algorithms for the 6-DOF platform forward solutions. From two perspectives can this paper improved control performance on the 6-DOF platform, and finally this paper gives its Matlab and Simulink simulation model design for further research.