Mobile Robot Path Planning Based On Multi-sensor Information Fusion

Since modern times,robot technology has developed rapidly and its working environment has become more and more complex.Path planning has become one of the key research areas of mobile robots.By analyzing related literature,aiming at the problems of weak path planning,poor real-time performance,low accuracy,and low intelligence of mobile robots in unknown environments,a path planning method for mobile robots based on multi-sensor information fusion is proposed.Based on the unstructured unknown environment,a mobile robot experimental platform is built.The robot finds the target position by itself in the unknown environment using vision sensors,detects the position of obstacles using ultrasonic sensors,and plans a collision-free path from the starting point to the target position according to the fuzzy control algorithm and T-S type fuzzy neural network algorithm.The specific research contents of this article include:(1)In the traditional method,all the sensor data analysis is completed in the main controller,which causes the main controller to calculate a lot and the mobile robot’s real-time performance is poor.This article starts with the analysis of the kinematics model of the mobile robot,adopts a modular and distributed design for the mobile robot control system,separates the vision processing module from the main controller,and reduces the amount of calculation of the main controller.Each module works in parallel,which improves the processing speed and computing power of the entire system.(2)On the basis of maintaining the inherent advantages of the fuzzy algorithm,it is proposed to integrate fuzzy control into behavior-based control technology,and decompose the robot path planning into a series of relatively independent behaviors such as finding the target,autonomous obstacle avoidance,and approaching the target,so as to improve the robot Path planning ability in unknown environment.(3)This article deeply researches the robot target recognition method.This article adds visual sensors to the robot to find the target location and return the information.There is no need to specify the specific coordinate position of the target in advance before path planning.Even if the target position is dynamically changing,the robot can perform path planning,so that the robot can adapt to the unknown environment.(4)Aiming at the problem of the lack of self-learning ability of fuzzy control,a fuzzy neural network system based on the T-S model is designed in combination with the robot model.The neural network learning algorithm mechanism is used to train the initial membership function offline,and the trained system integrates fuzzy control reasoning ability and adaptive and self-learning capabilities of neural network,the perception and decision-making of obstacle distance are closer to the driver’s experience.(5)Use Matlab toolbox to simulate and verify the path planning algorithm based on fuzzy control and T-S type fuzzy neural network,and finally perform physical verification.The verification shows that in the unstructured unknown environment,the mobile robot can independently plan a better path to avoid obstacles and reach the target by using two control algorithms.Compared with the fuzzy control algorithm,the robot running path based on the T-S fuzzy neural network algorithm is smoother and keeps a safer distance from obstacles.It can not only meet the path planning requirements of mobile robots,but also has higher control accuracy and response speed than fuzzy control.The validity and feasibility of the mobile robot path planning algorithm based on multi-sensor information fusion in this paper is verified theoretically and practically.The advantages of the method in this paper are: for the unstructured unknown environment,target recognition makes the robot dynamic path planning possible,the system has strong real-time performance,high control precision,and the robot trajectory is smoother.