Research On Path Planning Of Automated Vehicles In An Urban Road Network Environment

In recent years,China's per capita car ownership has been rising,and there have been many social problems such as traffic congestion,traffic accidents and air pollution,especially in urban areas.To alleviate the pressure of these problems,vehicles need to travel safely and efficiently in urban areas.With the development of autonomous driving technology,autonomous vehicle will occupy an important position in the car market for the foreseeable future.To this end,this paper studies the path planning problem of autonomous vehicle in urban road network environment,which mainly includes two parts: global path planning and local path planning.The former establishes the framework of the planned path,and the latter makes up for the details,and the two parts complement each other.The second chapter mainly researches a global optimal path search algorithm in an urban road network environment.In order to search the optimal path in an urban road network,a time-varying weighting directed graph model with a limited searching area is established.In a limited searching area,the model firstly introduces the searching direction factor,and comprehensively considers three factors: the time-varying vehicle flow density,the constant space distance and the searching direction,which is more suitable for actual traffic conditions of an urban road network.Meanwhile,the corresponding optimal path searching algorithm is also given.Compared with the traditional optimal path search algorithms in an urban road network,our proposed algorithm has two merits:(i)it lessens the searching scope,and then reduces the amount of computation;(ii)it adaptively selects the optimal path according to the varying traffic flow,and thus ensures that the final selected path is currently optimal in the present environment.Simulation results show that our algorithm is effective,adaptive and real-time.The third chapter mainly researches a local path planning model algorithm for autonomous vehicles in an urban road network environment.The incentives for vehicles to perform local path planning during a driving process include obstacle avoidance,overtaking,and turning(i.e.need to make a left or right turn at the next intersection,and move to left-turn or right-turn lane ahead).Performing a lane change maneuver is required to accomplish these goals.Therefore,the core of local path planning is to design a lane change strategy.The main steps of the proposed lane change strategy include: a)acquire the surrounding driving environment data in real time by using the onboard data acquisition system of the auto-driving vehicle;b)calculate the shortest safe distance between ego and the adjacent vehicles by using the acquired data,and adjusting really its vehicle speed to ensure driving safety;c)judge whether it is necessary to perform lane change maneuver according to the given global route planning or cost saving requirements;d)use the turning signal and lateral movement as the signal to communicate with the vehicle behind the target lane in order to get the chance of lane change;e)calculate the real-time parameters of the lane change process and generates the local path of the lane change process.Finally,the feasibility,security,real-time and reliability of the proposed algorithm are verified by simulation.The fourth chapter gives the summary of the thesis and the prediction about future works.