Research On Water Obstacle Recognition And Local Obstacle Avoidance Planning For Unmanned Vehicles Based On Vehicle Dynamics

In the wild,swamps,wetlands,and large-scale water barriers caused by severe weather conditions such as heavy rains pose a significant threat to UGVs.The scope of these areas is generally large,and the pavement adhesion conditions are not so good.Currently,optical devices such as cameras and lidar cannot accurately identify such obstacles: marshes are so hidden that they are hard to see even with the naked eye;The impassable area of the road after a rainstorm contains a lot of sediment,resulting in uneven water surface,which can also make detection difficult.Moreover,these devices cannot meet the demand of all-weather detection,such as limited camera recognition in rain and snow weather,limited Li DAR recognition in haze weather,etc.Such water obstacles are very common in unstructured scenes,so the detection of such obstacles by unmanned ground vehicles is one of the key points for the detection of obstacles in the field of unmanned vehicles.To make up for the limitations of such obstacle detection and meet the wild unmanned ground vehicles to the requirement of all-weather obstacle detection,considering these areas of tire-road friction coefficient is lower than the surrounding roads,thus can be used as a feature different from other traffic area,and then use the vehicle dynamics is used to identify the tireroad friction coefficient,according to the result of recognition implement to recognize and avoid obstacles.Vehicles at the edge of the barrier region recognition to the tire-road friction coefficient is low and gradually reduce the tire-road friction coefficient,the vehicles will be taken in operation,such as,reversing around the direction of the adhesion coefficient reduces,combined with the local obstacle avoidance algorithm,can cause the vehicle to avoid into the swamp,wetland,and zone depth after the rain the road impassable.In view of the above scenarios,this paper mainly carried out the following research work through the technology of road adhesion coefficient to identify water obstacles and local obstacle avoidance technology:Firstly,mathematical derivation is carried out on the vehicle dynamics model used.Based on Matlab/Simulink as the simulation platform,the vehicle dynamics model is established,and the theoretical verification of the model is carried out by designing the scene.Second,research to identify water obstacle edge of tire-road friction coefficient of road surface condition recognition algorithm,based on the vehicle for detection of all-weather and adapt to a variety of conditions,study drive,steering and braking condition recognition algorithm and design the tire-road friction coefficient,among them: for driving condition design based on state observer and recognition algorithm of recursive least squares estimate,design for steering conditions recognition algorithm based on fuzzy rules,aiming at braking condition design based on the Angle of vehicle deceleration tire-road friction coefficient of recognition algorithm and simulation.Then,a three-dimensional information grid obstacle map containing the pavement adhesion coefficient information was constructed.Obstacle avoidance algorithm based on safety threshold is designed for convex and concave obstacles.Through the three-dimensional information grid obstacle map with the road adhesion coefficient information,the simulation was carried out based on the Matlab/Simulink platform,and the effectiveness of the algorithm was analyzed according to the simulation results.Finally,based on the Car Sim platform,the vehicle,transmission system,tire and other models are built.The interface with Matlab/Simulink co-simulation is set and the co-simulation experiment is carried out to analyze the simulation results.