Research On The Path Planning And Obstacle Avoidance Strategy Of Intelligent Vehicles For Park

Driverless technology combines sensor,computer,artificial intelligence,communication,navigation and positioning,pattern recognition,machine vision,intelligent control and other technologies,which has played a decisive role in the development of intelligent technology for industries such as chemical inspection,urban road traffic,unmanned mining operations and urban sanitation.The core technology of driverless can be divided into three levels: perception,decision making and control,while path planning is one of its core technologies.At this stage,compared to the urban road scenario with complex vehicle conditions and high vehicle speeds,the complex and variable road types and weather types of noninstitutionalized roads such as closed chemical park inspection and intelligent distribution in fixed areas can lead to perceptual limitations and have a greater impact on the path planning of intelligent vehicles.Therefore,this paper takes the inspection intelligent vehicles in the park as the main body and conducts research on the path planning of unstructured roads in the closed park,and the main research contents are as follows:First,the kinematic analysis is carried out for the four-wheeled vehicle of Ackermann structure,and the kinematic model is established.According to the steering characteristics of the Ackerman structure vehicle,the kinematic model is simplified to a "bicycle" model including the front wheel turning angle,and the maximum turning angle and curvature constraints of the vehicle are obtained.On this basis,the park operation environment is analyzed and the mainstream environment modeling methods are compared,and the occupancy grid map is selected as the environment model.Secondly,the path planning based on the steering model of the inspection vehicle is studied,and the path planning principle of the traditional RRT algorithm is elaborated,and based on this,the improved bidirectional RRT algorithm is proposed.The kinematic constraints of the vehicle are introduced in the bi-directional RRT node extension,and the smooth guide path is obtained by using the cubic B-sample curve based on the removal of redundant nodes on the path,which shortens the path length and optimizes the path quality.Again,for the problems of obstacle avoidance in the tracing process of inspection vehicles,on the basic of global path planning,study the local path planning and obstacle avoidance strategies of inspection vehicles.Starting from the analysis of the kinematic model of the inspection vehicle,the improved VFH+algorithm is proposed.Propose to build a vehicle corner constraint model to obtain the front wheel corner range of the vehicle,and generate a global path connecting the current position of the vehicle with the target position as the direction-guided path of the VFH+ algorithm by introducing the angle constraints of the extended nodes in the bidirectional RRT algorithm.Finally,a hardware and software platform was constructed for real-world testing and validation,and a bidirectional RRT path planning algorithm and an improved VFH+ obstacle avoidance strategy were engineered and implemented on the software platform,and real-world road tests were conducted in a closed campus environment.The results show that the intelligent vehicle path planning in the lowspeed scenario can effectively avoid falling into the environmental dead zone and ensure that the planned path can reach the target location and satisfy the vehicle’s cornering constraints.