Obstacle Perception And Obstacle-aviding Strategy Research Of Omni-directional Mobile Robot

Environmental awareness and self-obstacle avoidance is one of the core contents of intelligent mobile robots in the research field,and it is the basis for robots to achieve advanced tasks.In the condition of lacking prior map,the robot only uses its own sensors to perceive the surrounding environment.According to the obstacle information,the robot plans out the next moment of movement and trajectory,and ultimately reaches the mission starting point as safely and efficiently as possible.This topic is combined with the National Natural Science Foundation of China(31470714).On the RM robot platform,taking the RPLIDAR A2 laser radar produced by Shanghai SLAMTEC Co.,Ltd.as the main distance measuring sensor,it focuses on obstacle perception and autonomous obstacle avoidance,conducts in-depth research and discussion in theory and practice.It will focus on solving the traditional threshold sensitivity problem of the VFH algorithm and the path optimization problem of the local obstacle avoidance algorithm.First,on the basis of the RM robot system platform,the coordinate system model of the robot and laser radar sensors is established.The coordinate transformation relationship between the robot coordinate system,the world coordinate system and the sensor coordinate system is deduced and analyzed,which realize the robot's conversion of sensor data in each coordinate system during obstacle avoidance.In order to improve the efficiency of obstacle recognition and the real-time performance of the obstacle avoidance system,a new obstacle model is proposed to describe the information of the surrounding environment.The selection of the model is based on the polar coordinate vector method instead of the traditional grid method,which make the data storage greatly reduced.We determined whether the distance between the obstacles meets the requirements for the robot to pass to classify the obstacles into a group of obstacles.Secondly,for the obstacle avoidance effect of the traditional VFH algorithm and the obstacle avoidance effect of the improved adaptive threshold algorithm,careful analysis and discussion are made.We found two defects.First,the VFH algorithm has a threshold-sensitive problem.Secondly,the VFH algorithm or the algorithm improved by the VFH algorithm uses a feasible sector angle division to obtain feasible directions,which causes the robot's path to be less than optimal.Aiming at the above problems,this paper proposes a local obstacle avoidance algorithm for the neighboring edge vertical outer expansion method.The selection of the feasible direction refers to the obstacle and the adjacent side of the robot,which to a certain extent,solves the threshold-sensitive problem,and ensures that the path of robot motion is better.Finally,the obstacle perception model proposed in this paper is verified by simulation and experiment.Experimental results show that using this model can reduce data storage space and improve system response.In the aspect of verification obstacle avoidance algorithm,based on the Matlab and Visual Studio experimental platforms,the feasibility of the neighboring vertical outreach method is verified from different perspectives.The experimental content includes the analysis of obstacle avoidance effects of single and multiple obstacles,the effect of threshold size on obstacle avoidance performance,and the comparison and analysis of obstacle avoidance results of different obstacle avoidance algorithms.The experimental results show that the adjacent outer vertical extension method has good obstacle avoidance performance under large and small thresholds.As the threshold increases,the smoother and more continuous trajectories planned by the obstacle avoidance algorithm are obtained,and the motion path obtained by the obstacle avoidance algorithm presented in this paper is better than the traditional obstacle avoidance algorithm.