Research On Path Planning Of Autonomous Driving Sweeper

Environmental issues have always been a hot issue that people are concerned about.The environmental protection department will arrange sweepers to clean up the road network.In the near future,self-driving sweepers have begun to be tested in many cities.The self-driving sweeper can disregard labor costs,and its service time is not restricted by driver factors,there is no fatigue driving,and carbon emissions are reduced.In addition,the self-driving sweeper can also provide services at any time during peak periods,solving the problem of congestion caused by slow-moving vehicles that affect normal-speed driving on the main road.This paper takes the path planning of autonomous sweeper as the research object.The problem is divided into global path planning and local path planning of autonomous sweeper.Global path planning is divided into closed multi depot automatic sweeper global path planning and open multi depot automatic sweeper global path planning.Local path planning is divided into local detour path planning of autonomous sweeper and local detour path simulation of autonomous sweeper.In this paper,a mathematical model is established based on the characteristics of the path planning of the automatic driving sweeper in a closed multi-yard.According to the characteristics of the closed multi-car yard that requires the automatic driving sweeper to return to the original departure yard after completing the cleaning task,a strategy for returning to the yard is formulated,and the service lane and vehicle driving direction are restricted.The model is solved by designing a two-stage method.First,the path is divided into various parking lots by clustering analysis of the global road network.Then use real number coding to encode chromosomes,design competition selection operator,multi-segment crossover operator and gene swap mutation operator,and formulate service strategy while walking,return no service strategy and path segmentation strategy for decoding.Through the analysis of calculation examples,it can be found that the effective working driving time of the first autonomous driving sweeper in the two areas reached about 90% during the driving process.The distribution of the remaining demand arc for the second car’s service is relatively fragmented,resulting in a decrease in the proportion of the second car’s working time,but both are higher than 60%.This paper establishes a mathematical model based on the characteristics of the path planning of the open multi-yard automatic driving sweeper.The charging selection strategy is formulated according to the characteristics of open multi-parking yards that there is no need to return to the departure parking yard after the self-driving sweeper completes the service.There are parking lots and charging parking spaces in the road network.When the self-driving sweeper is faced with insufficient power during the service process,it can choose to go to the nearest parking space with charging piles or return to the nearest parking lot.Therefore,in the decoding part of the algorithm,in addition to the selection,crossover,and mutation operators of closed multi-parking,the randomness of the departure parking lot and the returning parking lot or charging pile is increased,and the diversity of chromosomes is maintained.Finally,the one-stage method is used to directly solve the path that each autonomous driving sweeper needs to serve.Through the analysis of calculation examples,it can be found that in the comparison of multiple results,the average optimal total driving time when the number of iterations is 6000 is 56% less than the optimal average time of the initial chromosome.In this paper,the detour track of the self-driving sweeper to avoid static obstacles is studied.According to the characteristics of autonomous sweeper,the multi-objective models of vehicle safety risk minimization and area minimization are established to study the local detour path of autonomous sweeper.The genetic algorithm is used to generate the random nodes of the detour trajectory.The improved distance metric function is used to screen the initial nodes and improve the smoothness of the trajectory.The non-dominated sorting method and congestion calculation are constructed to solve the problem.Based on the advantages and disadvantages of the trajectory planning method,the quintic polynomial trajectory planning algorithm and the improved NSGAⅡ algorithm are used for trajectory simulation.According to the characteristics of self-driving sweeper,the two methods are improved respectively.The randomness of lane changing end point is increased in quintic polynomial,and the possibility of lane changing end point is increased.The trimming function is added to NSGAⅡ algorithm to optimize the smoothness of the trajectory and ues the B-spline function to fit the trajectory.Finally,the appropriate path is selected by solving the Pareto solution set.The two simulation methods are compared and analyzed.