Research On Localization And Navigation System Of Outdoor Mobile Robot Based On LiDAR

In recent years,the development of robots has become more and more intelligent.Compared with traditional robots,intelligent mobile robots with environment perception and autonomous decision-making are more popular among users.The most basic question that an intelligent robot needs to solve is: where is oneself? Where are you going? How to reach the destination safely? These issues involve key technologies such as autonomous localization of robots,construction of environmental maps,and navigation.Aiming at the design and implementation of patrol robots in outdoor scenes,this article mainly completes the following tasks:(1)The patrol robot system hardware platform was designed and built.This article started with the purchase of the robot chassis,and successively purchased hardware equipment such as lithium batteries,power adapters,industrial computers,threedimensional lidar sensors,and monocular cameras.The hardware platform of the patrol robot was built in one step,and the hardware equipment was debugged in time to make it work normally.A crawler mobile robot was successfully built.The robot is large in size,1020*750*405mm,and has strong endurance,the cruising range is up to 100 km.(2)The odometer-based 3D laser point cloud data correction is realized.Aiming at the point cloud data distortion caused by the motion of the 3D lidar,this article uses the wheel odometer that comes with the motor driver of the patrol robot to remove the motion distortion of the 3D point cloud data.The wheel odometer can record the wheel speed information in real time,and calculate the moving distance of the robot according to the size of the wheel.It has high accuracy in a short time and short distance,and the update frequency of the odometer is high.This advantage of the design is to estimate the position and attitude changes of the lidar,remove the movement distortion of the threedimensional point cloud data,to obtain more accurate external environment information,and provide high-quality point cloud data for subsequent autonomous localization and navigation of the robot.(3)Aiming at the simultaneous localization and mapping method of robots based on LOAM(Lidar Odometry and Mapping),this paper proposes an improved method.Since the LOAM method only uses the laser point cloud data of two adjacent frames to estimate the pose of the robot,the localization accuracy of the robot is low.In this paper,3D lidar is used as a sensor for patrol robots to obtain information about the external environment,combined with wheeled odometer,to calculate the relative motion of the robot,assist in point cloud matching to realize the robot's pose estimation,improve the accuracy of pose estimation,and use improved The SLAM(Simultaneous Localization and Mapping)method realizes the simultaneous localization and mapping of the patrol robot in the outdoor environment.(4)In response to the needs of patrol robots for autonomous path planning and obstacle avoidance in outdoor environments,the current autonomous path planning algorithms in the field of robotics are studied.This paper mainly studies the static global path planning based on the A* shortest path algorithm.And based on DWA(Dynamic Window Approach)algorithm of dynamic local path planning and automatic obstacle avoidance,and use MATLAB to carry out simulation experiments.