Implementation of Optimal Weights Initialization Technology in Robot Learning

Abstract

An intelligent obstacle avoidance model based on BP neural network is established.Also a novel optimal weights initialization technology is proposed so that the sample sets and initial weights can match perfectly.Consequently,the convergence speed increases evidently.In order to improve the real-time performance,hybrid programming using C and assemble language is adopted.Computer simulation and real test show that the system has a strong ability of learning and good performance of human computer interaction.