Research And Realization Of Robot Obstacle Avoidance System Based On Multi-sensor Information Fusion

With the wide application of mobile robots in industrial production and daily life,people have higher and higher requirements for their safety.Mobile robots will encounter a variety of obstacles when moving in unknown environments.How to identify and avoid these obstacles quickly and effectively to complete the established tasks has always been a research hotspot in the field of autonomous mobile robots.Single sensor often has limitations in practical application.The rapid development of multi-sensor information fusion technology provides a solution.Nowadays,information fusion technology has attracted extensive attention of researchers all over the world.Applying multi-sensor information fusion technology to obstacle avoidance task of mobile robot,the robot can acquire redundant and complementary information of environment,and the reliability and real-time degree of information have been greatly improved.Therefore,this paper takes the self-developed mobile robot as the platform to study the application of multi-sensor information fusion technology in obstacle avoidance of mobile robots.The specific contents of this paper are as follows:1.Based on the analysis of obstacle avoidance requirement of mobile robot,the overall scheme of obstacle avoidance system of mobile robot is designed,and the multi-sensor system and information fusion framework are designed.2.Based on the analysis of crosstalk in ultrasonic ranging,a method of ultrasonic ranging based on 2PSK modulation is proposed,which eliminates crosstalk and improves ranging accuracy.3.The Extended Kalman Filter(EKF)algorithm is studied.A data fusion algorithm based on EKF is designed,which combines the data of inertial sensors,encoders and optical mouse sensors for positioning.The accuracy and validity of the algorithm are verified by experimental tests.4.The algorithm of fuzzy neural network is studied.The EKF-based algorithm of fuzzy neural network is applied in obstacle avoidance system.The sparse auto-encoder is introduced to extract the feature of dimensionality reduction from the high-dimensional data obtained by various sensors,which solves the problems of long training time,complex structure and easy over-fitting of the fuzzy neural network in the training process.The feasibility of the algorithm is verified by experimental analysis.This paper studies the obstacle avoidance system of mobile robot,designs a method of eliminating ultrasonic crosstalk.EKF is used to fuse three sensors including photoelectric mouse sensor to locate the obstacle,and combined with sparse automatic encoder and EKF-based fuzzy neural network algorithm to avoid obstacle.The mobile robot platform is built,and the ultrasonic ranging method based on 2PSK modulation is validated by the obstacle avoidance platform,which enhances the anti-interference ability.The validity of the location fusion algorithm is verified,and the positioning accuracy is improved.The feasibility of the obstacle avoidance algorithm is also verified.