Study On Dynamic Obstacle Avoidance Method For Mobile Robot In Partly Unknown Environment

This paper focuses on dynamic obstacle avoidance of mobile robot and relativeproblems in part unknown environment. To ensure that the robot can be able to copewith the complex dynamic environment in the walking process and reach thedesired destination without collision with obstacles. Dynamic obstacle avoidance isone of the important hotpots in the research of robotics, and also the key of mobilerobot towarding to intelligent. The navigation technique also provides anexperimental platform for signal detection and processing, multi-sensor informationfusion, and motion control technology and so on.Firstly, a brief analysis and induction about the research backgrounds andcurrent situation of the obstacle avoidance is proposed, which makes goodpreparation for the in-depth study of the subject. Secondly, according to the partunknown environment, the grid method is applied to create a global map withknown environmental information. Then the local environment information usinglaser scanner and infrared sensors is detected. The complete map is composed bythe global map and local map.This paper focuses on using the artificial potential field method to avoid thecollisions with dynamic obstacles. There are three defects inherent in the potentialfield method, which are Goals Nonreachable, Local Minimum, and DynamicReal-time. Profound causes of these three defects are analyzed. The conventionalpotential field method is improved several times to solve the above problems. Inorder to improve the comprehensive performance of the dynamic obstacleavoidance system, a closed-loop control system of fuzzy logic controller basing onartificial potential function is proposed. The closed-loop control system not onlyachieves the dynamic obstacle avoidance, but also reduces the shock in front of theobatacle some extent, which improved the stability of the system.In the end of this paper, some simulation experiments of the improvedalgorithm under complex dynamic environment are carried out in MATLABplatform. In these simulation experiments, the mobile robot overcomes the abovedefects and reaches the destination without acollision with obstacles. Theeffectiveness of the algorithm is verified by the simulation experiment.