Research On Stable Positioning And Trajectory Control Of Tracked Robot In Outdoor Partial Occlusion Environment

With the development of technology,intelligent mobile robot has become an important production tool,which affects people's way of life and prod uction.It has a wide range of application prospects and has been a hot research content in robot technology.It is very important to realize the stable and accurate positioning and smooth movement of mobile robot,which is an important basis for robot navigation,monitoring and other functions.But in the harsh environment of petroleum,smelting,military and so on,all kinds of production equipment or obstacles are closely connected,which block the positioning satellite signal and affect the satellite positioning of robot.The poor road condition and many obstacles restrict the robot's movement.It brings new challenges to the application of mobile robot.Therefore,this paper studies the stable positioning and trajectory control algorithm of tracked robot which is more suitable for complex environment.(1)The overall structure of the tracked robot is analyzed and designed.According to the characteristics of the working environment,the appropriate sensors are selected to build the tracked robot platform.The power supply circuit and communication interface are designed,and the motion control of the robot is realized in ROS.Then,the wheel odometer model with slip is established,and the parameters are optimized according to other sensor information to obtain a better wheel odometer,which is verified on the self-developed tracked robot.(2)Aiming at the problems of low positioning accuracy of GNSS in occluded environment,limited relative positioning stability of laser and point cloud map,and difficult relocation after position loss,a multi-sensor fusion positioning algorithm based on neural network is proposed to obtain better positioning results.This algorithm combines laser IMU odometer positioning results,wheel odometer positioning results and GNSS positioning results to obtain more stable and accurate robot pose,which lays a good foundation for other functions of mobile robot,such as navigation,trajectory control and so on.The algorithm is verified on KITTI data set,and the results show that the algorithm can realize relocation within 2 seconds after the robot's posture is lost,and the positioning accuracy is0.25 m,which is 63% higher than the laser positioning algorithm.(3)In order to realize the trajectory tracking control of tracked r obot,the trajectory tracking error model is established.To solve the slip problem of tracked robot in the turning process,motion compensation is added to the adaptive sliding film control algorithm with high robustness and parameter insensitivity.The adaptive sliding film tracking controller based on motion compensation is obtained,and the stability of the algorithm is proved;Finally,the experimental verification is carried out on the simulation platform.The results show that the position tracking error of the robot is about 0.1m and the heading angle tracking error is about 0.02 rad after adding the motion compensation.The control accuracy and rapidity are improved to a certain extent.(4)Finally,the experiment is carried out on the self-developed tracked robot.The results of fusion positioning on the 4.5km long path of the Science Park of Harbin Institute of technology show that the robot can realize repositioning within2 seconds after losing its position posture,and the positioning accuracy i s 0.28 m,which is 77.4% higher than the satellite positioning accuracy and 57% higher than the laser positioning algorithm.In the open environment,GNSS positioning data is used as the robot position feedback to make the robot track two moving tracks.The experimental results show that the position tracking error of the robot is about0.15 m,and the heading angle tracking error is about 0.05 rad.In the tracking experiments of the two kinds of tracks,the position error is reduced by 60% and25%,and the heading angle tracking error is reduced by 50% and 20% respectively.Finally,the fusion positioning results are used as the robot's pose feedback to make the robot track a given trajectory,and the GNSS is used to record the actual running pose of the robot.The experimental results show that in this case,the integrated positioning accuracy of the fusion positioning algorithm is 0.19 m,and the tracking accuracy is about 0.24 m.