Multi-Sensors Based Robotics Follow System On Unconstructed Environment

With the development of robotics,the idea of "Human-Machine Integration" has increasingly become one of the important directions pursued by science and technology personnel.In some occasions where goods need to be moved frequently such as factories and warehouses,the demand of the handling robot that can autonomously follow users is also increasing.However,the existing follower robots on the market are only suitable for the structured environments,and there are very few related researches and products in complex terrain environments.Considering the technical landing and practicality,this thesis proposes an improved autonomous tracking method for trajectory following.This method meets the requirements of robots operating in unknown environments.The main contents of this article are as follows:(1)Determine the autonomous following robot system based on trajectory following.The robot advances strictly in accordance with the target motion trajectory and completes the following task.(2)Solve the problem of positioning the target trajectory in the trajectory following solution.First,the robot is equipped with an inertial navigation system to complete its own track calculation,and then the "absolute positioning" of the robot in the world coordinate system can be obtained.Then combine the description of the target's pose in the robot coordinate system with the robot's pose in the world coordinate system,and use coordinate transformation to get the target's pose in the world coordinate system.(3)The method of multi-sensor data fusion is adopted to improve the stability of the system and reduce the amount of calculation of the system.First,the convex combination algorithm is used to fuse the track estimation algorithm scheme and the UWB scheme to obtain the robot trajectory;then,for the detection of the target,the lidar data and camera data are fused to reduce the amount of system calculations.(4)When a dynamic obstacle appears on the robot's forward path,the robot completes the local path planning.This thesis proposes a local path planning algorithm based on an improved artificial potential field method.In terms of principle experiment verification,the task of this subject requires software simulation,without providing actual sensors and experimental platforms.Therefore,this thesis designs a set of simulation environment.The environment has information such as maps of complex terrain,various sensor simulation data and so on.Among them,through the classification and modeling of complex terrain features,a method for building random maps under complex terrain based on raster maps is proposed.Finally,based on the established simulation environment of the following robot,the simulation experiments are carried out for the autonomous follow-up robot algorithm proposed in this thesis.The experimental results show that the proposed autonomous following algorithm can follow the target well in the complex terrain environment.