Research On Indoor Environment Modeling And Autonomous Navigation Of Mobile Robot Based On 2D Laser Sensor

Mobile robot is a comprehensive system that integrates environmental perception,behavioral decision-making and motion control.With the development of technology,robots have entered many aspects of human life,such as sweeping robots,service robots,unmanned aerial vehicle,driverless cars.The working environment of the indoor mobile robot determines that the robot must safely move from one place to another in an unknown and changing environment without human intervention,and complete the task assigned by the owner.Therefore,autonomous navigation is a must-have function for indoor mobile robots.This paper takes ActivMedia's robot simulation platform MobileSim and Robot Operating System(ROS)as the research platform.Firstly,it studies the Simultaneous Localization and Mapping(SLAM)problem of mobile robot in indoor environment,and then realizes the path planning of mobile robot on the constructed grid map.And a mobile robot feedback controller designed with the Composite Nonlinear Feedback(CNF)method is proposed for the robot to track the optimal trajectory.The main work of this paper can be divided into the following three parts:The environment modeling module: In this module,the indoor environment is modeled by EKF-SLAM.Firstly,the mathematical principle of the extended Kalman filter and the mathematical derivation of EKF-SLAM are introduced.Aiming at the point cloud information of the laser sensor and the various line features(such as walls and tables)existing in the actual indoor environment,We consider to use the Random Sample Consensus(RANSAC)algorithm to extract the straight line from the point cloud information as the landmark in the EKF-SLAM process to obtain the relatively accurate location of the robot,and then combined with sensor information we can build a grid map.The path planning module: Firstly,based on the grid map obtained in the environment modeling module,the global optimal path of the static environment is obtained by using the A* algorithm.Then we construct the cost function related to the global optimal path for the dynamic window approach(DWA),so the robot can obtain a smooth path and dynamically avoid obstacle in the real time while ensuring the global optimality of the planned path.Moreover,the output of DWA algorithm is linear velocity and angular velocity,which isbeneficial to the motion control of the robot.The motion control module: At the bottom,we need to design a controller to ensure that the robot can track the trajectory planned by the path planning module.In this paper,the trajectory tracking problem is firstly transformed into the leader-follower problem,which can be transformed into the output regulation problem.We use the composite nonlinear feedback controller to solve the output regulation problem and achieve the trajectory tracking.