Research On Robot Localization And Navigation System Based On SLAM

Autonomous navigation in unknown environments has always been a hotspot and difficult issue in the field of robotics research,and it is also an indicator of robot autonomy.Synchronous positioning and map creation(SLAM)technology is an important technique for constructing a high-precision environment map model and achieving high-precision positioning without prior knowledge.This technology is of great significance for the autonomous operation of mobile robots in unmanned situations.This dissertation aims at the autonomous navigation of wheeled robots in indoor environment,and mainly studies the following key technologies.Firstly,the main algorithm framework in synchronous positioning and map creation technology for mobile robots is analyzed based on Kalman filter-based SLAM and particle filter-based SLAM.Through MATLAB simulation of the positioning accuracy of the two algorithms in different noise environments,the particle filter-based SLAM algorithm has better accuracy for the positioning under complex environmental noise,and established the algorithm framework applicable to the platform.Secondly,the traditional dynamic window based DWA(Dynamic Window Approach)path planning algorithm is improved.By analyzing the experiment of DWA algorithm related parameters and the principle,a global path based DWA algorithm is proposed.Compared with the original algorithm,this algorithm has better reliability and solves the problem of the optimal path selection and the navigation failure of traditional DWA algorithm in U-shaped environment.Finally,the wheeled robot is used as an experimental platform to design and implement an autonomous navigation software system based on SLAM in the robot operating system ROS.The SLAM technology is used to realize the positioning and construction of the robot in an unknown environment,and the DWA based on global path optimization is utilized.The algorithm implements path planning and obstacle avoidance in an unfamiliar environment,and tests the system in an actual environment.Experiments show that the system can complete the autonomous navigation function of the wheeled robot in the indoor environment.At the same time,it can construct an accurate indoor two-dimensional grid map with high practicability and robustness.