Evalaution and Control of Mirror Sliding Friction of Unmanned Ground Vehicle using Radial Basis Neural Network

Main Article Content

Xiaohui Yang
Jian Zhao

Abstract

In order to effectively analyse the mirror sliding friction(MSF) degree of unmanned ground vehicle(UGV) and improve its anti-disturbance performance, a simulation method for MSF degree of UGV based on RBF neural network is proposed. A single-input and double-output RBF neural network is adopted to estimate the uncertain dynamic parameters of the MSF model. The obtained parameters are used to describe the MSF control law based on RBF neural network. An adaptive law based on slow time-varying disturbance characteristics is designed to estimate the total friction disturbance term in the MSF model online. The simulation results show that the proposed method can analyse the MSF degree of unmanned ground vehicle at different speeds and gradients. The influence of gradient on the decline rate of friction degree is greater than that of vehicle speed. The mean error of friction disturbance term calculated by the method is only about 0.9% which has the advantage of low error of friction degree estimation when compared to conventional methods.

Article Details

Section
Articles