自动代客泊车（Automated Valet Parking, AVP）能够将驾驶员从泊车场景 中解放出来，该系统能够实现车辆的自动驾驶、泊车和充电，因而自动代客泊车 也增大了人们对电动车辆的吸引力。通过解决无人化泊车的驾驶任务，能够带来 城市区域自动驾驶的一次伟大飞跃。 本论文的研究工作主要集中在运动规划模块的第一步：全局路径规划器。因此依据实际的研究方法，开发并评估全局路径规划器。基于占据栅格地图（Occupancy grid map, OGM），利用 A*算法求解最优路径。然后在每个转折点处单独应用极多项式函数进行平滑处理，获得可驾驶路径。最后将全局路径规划的结果用于路径跟踪的输入。 为了获得合理的仿真结果，设计了基于预瞄距离与节点检测的车辆控制器， 使车辆跟踪期望路径。因此，需要利用车辆当前位置、航向角与道路信息，求解 预瞄点及相关偏差量。控制器利用输入的偏差量求解转向参数用于导航车辆回到 期望路径。 通过 Matlab/Simulink 与 CarSim 的联合仿真，详细分析路径跟踪与车辆行为，仿真的主要目的是在合理的准确度下跟踪理想的路径。因此，本文对关键参数与其影响进行了分析推导。

Automated Valet Parking (AVP) can relieve drivers from having to park. It enables vehicles to drive, park and charge fully automatically. Due to this fact Automated Valet Parking also increases the allure of electric vehicles. By solving the driving tasks of driverless parking, a great leap can be done to automated driving in urban areas. This research work focuses on the first working step of the motion planning module, the global path planner. Therefore, a global path planner will be developed and evaluated based on actual research methods. A route will be computed by using an A* algorithm based on the occupancy grid map. Subsequently, a smoothing operation will follow to obtain a drivable path. By implementing a single polar polynomial function each vertex is smoothened. The resulting global path serves as an input for path tracking. For a reasonable simulation, a vehicle controller with a lock-ahead distance and a node-check will be designed in order to keep the vehicle on the desired path. Therefore, the position and the heading angle of the vehicle as well as the path will be needed to calculate the errors and the preview point. The errors will then serve as an input for controllers, which will adjust defined steering parameters to navigate the vehicle back. A combined simulation of the designed components in Matlab/Simulink and CarSim will be carried out to analyze the path tracking and vehicle behavior in detail. The main target for the simulation is to track the desired path with a reasonable accuracy. Hence, important parameters and their influence will be derived.