Research On Location And Obstacle Avoidance Of AGV Based On Information Fusion

With the rapid development of today's society,information fusion technology has been widely used in many fields and which has been applied to mobile robots which provides theoretical support for the further improvement of the level of robot intelligence.Location and obstacle avoidance is an important guarantee for automatic guided vehicle(AGV)to accomplish various tasks.Multi-sensor information fusion technology is applied to AGV positioning and obstacle avoidance system.It can provide an effective technical solution for AGV to work in a variety of complex,dynamic,uncertain or unknown environments.Therefore,this paper focuses on information fusion technology to study AGV location and obstacle avoidance.In the aspect of AGV location,the Kalman filter can only solve the linear system model;it is not suitable for the nonlinear system model.But in reality,AGV is a nonlinear system,so there exists some weakness when using Kalman filter to solve the localization problem of AGV,therefore the extended Kalman localization algorithm is used in this paper.First,the motion model of AGV and the observation model of sensor are designed,and then the information collected by the odometer and ultrasonic sensor is fused.By using the control model to estimate the initial position and pose of the AGV,and to detect the surrounding environment by ultrasonic sensor,the estimation error can be corrected in time,and the positioning accuracy can be improved.Finally,the simulation results show that the combination of the mileage meter and ultrasonic sensor information fusion can effectively eliminate the accumulation error of the mileage meter during the AGV movement,and improve the positioning accuracy of the AGV.In the aspect of obstacle avoidance of AGV,because the information provided by single sensor is not enough to make AGV avoid obstacle quickly and accurately,in this paper,many ultrasonic sensors and infrared sensors are used to obtain the information of obstacles in the operation of AGV.The position and target azimuth information of the AGV is obtained by using electronic compass in real time.Using information fusion technology can make the information obtained by the system more accurate and comprehensive.In this paper,the fuzzy control algorithm and the fuzzy neural network algorithm are verified.It is found that the fuzzy neural network algorithm has more advantages and can better achieve the AGV in the unknown environment of autonomous obstacle avoidance.Therefore,the input and output variables,membership function and fuzzy control rule base are designed for the environment of AGV,and Matlab simulation is carried out.The simulation results show that the fuzzy neural network obstacle avoidance algorithm can approach the nonlinear system with high precision,high obstacle avoidance accuracy,fast response speed,and the fuzzy neural network algorithm is superior to the fuzzy control algorithm.In this paper,the information fusion method is used to realize the location and obstacle avoidance of AGV,which can ensure the safe and accurate operation of AGV in complex and unknown environment.