The Motion Planning And Experimental Research For Autonomous Sweeping Vehicle

With the rapid development of China’s economy and the release and implementation of "made in China 2025",driverless vehicles have gradually stepped into people’s vision.Intelligent Internet connected vehicle is the general trend,and driverless is accumulating energy.In recent years,scholars at home and abroad have invested in the research and experiment in the related direction of automatic driving,of which motion planning is the difficulty.The motion planning technology of driverless vehicle involves many fields such as automobile engineering,computer field,robot field,information technology and so on.Therefore,the research on driverless motion planning is challenging and very important.Based on the real vehicle platform provided by an enterprise,this paper studies and designs the automatic driving scheme and motion planning of articulated driverless sweeper,taking the articulated driverless sweeper as the research object,considering its kinematic constraints,operation indicators and other requirements.The work of this paper is as follows:According to the characteristics of low cost and driving by the side of the road of articulated driverless sweeper,the automatic driving scheme is designed,the software and hardware platform is constructed,and the research contents and design requirements of motion planning are clarified.Based on the characteristics of articulated vehicles,a motion planning framework is designed.This motion planning method mainly consists of two parts: based on trajectory generation of model prediction and trapezoidal velicity planning.The trajectory planning is decomposed into space trajectory generation and velocity curve,and then coupled to output the optimal trajectory.Then according to the characteristics of the sweeper running close to the edge and articulated vehicles,the constraints of vehicles are analyzed,design terminal sampling based on reference path,a large number of trajectory clusters are generated by using the method of model prediction and trajectory generation of vehicle model,The pre stored tracks are looked up in the off-line table and the initial guess is obtained from it.When planning the tracks,the tracks cluster is generated according to the initial guess,and then combined with the actual situation of sweeper operation,the trajectory evaluation function is constructed,the optimal trajectory is selected,and the result of trajectory planning is improved to ensure the safety of the selected trajectory,feasibility and real-time.At the same time,based on the upper trajectory planning,a trajectory pure tracking control system based on adaptive preview distance is designed to make the sweeper track the desired trajectory stably and quickly.Then the speed is decomposed from the track planning,and the sweeper is planned separately to analyse the constraints in the speed planning.Based on the advantages of trapezoidal velocity planning,the acceleration deceleration section is improved by using the parameterized third-order Bessel curve,and the dynamic obstacles are analyzed and treated.Then,the anti integral saturation PID longitudinal speed controller is designed to Analyze the switch strategy of accelerating braking to make the vehicle track the expected speed smoothly and in real time.Finally,based on the real vehicle platform of the enterprise,the test scenario and experimental scheme are designed to verify and analyze the motion planning method of articulated driverless sweeper under multiple working conditions.The experimental results show that the trajectory generated by the designed motion planning method is safe,real-time and adaptive,and can meet the needs of low-speed and stable edge trimming operation of sweeper ask.