Research On Obstacle Avoidance Method Of Unmanned Surface Vessel Based On Lidar

Marine engineering equipment has been developed to be unmanned and intelligent with the increasing investment of our country to marine engineering in the past few years.Unmanned surface vessel(hereinafter referred to as "USV")has rapidly become a research hotspot.Environmental perception is the basis for USV to realize autonomous driving and obstacle avoidance,and plays an important role in improving the intelligent level of the USV.As a kind of sensor that acquires environmental information by emitting light beam,Li DAR has the characteristics of fast speed,high accuracy,and all-weather,which is very suitable for the needs of the USV.Therefore,this thesis uses Li DAR as a detection method to build an USV obstacle avoidance system,so as to improve the level of intelligence of the USV.In terms of obstacle perception,this thesis uses Li DAR to obtain point cloud data and filter the data based on the actual conditions of the water surface,through the clustering and segmentation of the point cloud,effective obstacle information is obtained.The motion state of the obstacle is estimated through the combination of point cloud geometric feature matching and nearest neighbor method;In terms of static obstacle avoidance,based on the basic idea of the VFH algorithm and considering the delay in the actual operation of the USV,this thesis proposes a "dynamic sector method" that sets the sector radius and sector center angle according to the current state of the USV.,by setting dynamic sectors,reduces the risk of collision of the USV due to delays and other factors,and overcomes the shortcomings of the traditional VFH algorithm;In terms of dynamic obstacle avoidance,this thesis improves the DWA algorithm,from the perspective of safety avoidance,quantifies the safety of different obstacle avoidance trajectories when the USV avoids obstacles.Combining the kinematics model of the USV,the sampling space is designed,and an obstacle avoidance method considering the safety of trajectory is proposed,which reduces the risk of collision when the USV encounters a dynamic obstacle.On the basis of the above work,this thesis built an USV platform,carried out static obstacle avoidance experiments and dynamic obstacle avoidance experiments on the water to verify the effect of the algorithm.The research in this thesis realizes the sensing and decision making of the USV to the surrounding environment,and builds the USV obstacle avoidance system from the "sensingdecision making" level,so that the USV can make decisions autonomously based on the sensing results of the surrounding environment by Li DAR,which improves the level of intelligence of the USV.