Study On Autonomous Mobile Robot Obstacle Avoidance In Unknown Environment

Study on autonomous mobile robot obstacle avoidance in unknown environment is oneof the core contents in mobile robot field. The main task is that robot should automaticallycomplete a safety and optimal path selection from the starting to target in unknownenvironment, which requires robot to complete the avoidance of static and dynamic obstacles.In this research, experiment platform of4WD mobile robot is set, on which an in-depth studyon algorithm of autonomous mobile robot obstacle avoidance in unknown environment islaunched.First of all, on the in-depth research and improvement of rolling window obstacleavoidance, algorithm of the G-rolling window obstacle avoidance, adopting the "polarcoordinate vector method" for obstacle detection and modeling of external environment, isproposed. Vector density modal method, which greatly reduces the processing amount ofrobot, is put forward to optimize the rolling window. So is the method of distance comparisonand vector coincident, which contributes to solve the question of boundary followingtermination point and optimize the escape route with the local minimum. Compared with thetraditional rolling window obstacle avoidance algorithm, the improved one, which makesenvironment modeling more accurate, choice of path more optimized, is more practical.Then, experiment platform of4WD mobile robot that consists of energy layer, drivelayer, sensor layer and main control layer is built in this paper. The platform has an integrateduse of ultrasonic distance sensors, infrared sensors, information fusion technology of multisensors to build the environment model, electronic compass sensor for the control of real-timenavigation, the encoder control for robot’s pose adjustment and global localization combiningwith electronic compass.Finally, the mobile robot obstacle avoidance simulation tests are performed in MATLAB.The algorithm of G-rolling window obstacle avoidance is validated by simulation of selectingstatic, dynamic obstacles and some other representative environment. The ability of robotobstacle avoidance is tested on the experiment platform of4WD mobile robot in real andunknown environment. All these results prove the efficiency and optimization of the improvealgorithm. In a word, autonomous robots achieve the autonomous obstacle avoidance andnavigation in unknown environment and have safety paths with the algorithm of obstacleavoidance under G-rolling window.