【作者】 李勇；

【导师】 熊光明；

【作者基本信息】 北京理工大学 ， 机械工程， 2015， 硕士

【摘要】 无人驾驶车辆技术已经发展到一定水平，但完全应用于现实生活还有很长一段距离。尤其在混合交通场景中，多个无人驾驶车辆之间、无人驾驶车辆与有人驾驶车辆之间如何交互还是一个亟待解决的问题。本文研究无信号灯十字交叉路口环境下的多车协作控制，在考虑车辆动力学特性的基础上，针对无人驾驶车辆与有人驾驶车辆及多无人驾驶车辆之间的协作与控制问题展开研究。无人驾驶车辆与有人驾驶车辆之间进行交互时，首先通过无线通信获取有人驾驶车辆的状态信息，并建立交叉口车辆行为预测的有限状态机模型，对交叉口有人驾驶车辆的行为进行预测。构建应用于无人驾驶车辆的混合状态系统，结合无人驾驶车辆交叉口安全条件运行规则，实现对无人驾驶车辆离散状态的分析及车辆控制。最后运用PreScan和Simulink/Stateflow实现交叉口联合仿真。仿真结果表明，该方法可以使无人驾驶车辆与有人驾驶车辆进行交互协作，并安全地通过路口。多无人驾驶车辆通过路口时，基于分层控制理念，搭建了路口控制系统。该系统分为宏观路口控制器、车辆顶层控制器以及车辆底层控制器。其中宏观路口控制器通过无线通信的方式获取进入路口的各个无人驾驶车辆的状态信息，并根据各个车辆的速度和位置等信息预测车辆在路口潜在的冲突，计算出各个车辆避免冲突的目标速度，而车辆顶层和底层控制器则控制车辆跟踪目标速度。通过PreScan和Simulink搭建仿真实验，实验结果表明，基于冲突预测的路口协作算法在保证安全的前提下提高了路口通行效率。为了采用真实的无人驾驶车辆验证多车路口协作算法，基于东风猛士平台设计无人驾驶车辆的纵向控制系统，其中包括电控液压制动系统、电控油门系统、车速的采集以及综合速度的控制，实现无人驾驶车辆的纵向控制。通过该无人驾驶车辆和一辆有人驾驶车辆搭建无信号灯多车路口协作研究实车实验平台，该平台中的车辆采用GPS和惯导进行组合定位，采用ZigBee模块进行无人驾驶车辆和有人驾驶车辆之间的通信。实车实验验证了路口协作算法在实际无信号灯路口环境中的可行性，提高了无人驾驶车辆的行驶安全性。更多还原

【Abstract】 The technology of unmanned vehicle has been to a certain level, but it still has a longdistance to be completely applied to our real life. Especially in the mixed traffic scenario, itreserves to be a serious problem that how to cooperate between the multi-unmanned vehiclesor between the unmanned vehicle and the manned vehicle. In this thesis, the multi-vehiclecollaborative control was carried out under an intersection environment without traffic lights.The proposed framework forces on cooperation and control among the multi-unmannedvehicles, as well as the unmanned vehicle and the manned vehicle, during which vehicledynamics are considered.The unmanned vehicle interacting with the manned vehicle through cooperation in anintersection was firstly studied. In this situation, the unmanned vehicle obtained the statusinformation of the manned vehicle via wireless communication and then the finite statemachine (FSM) model was established to predict the behaviors of manned vehicles in theintersection. Then the ego-unmanned vehicle states were taken into considerationcomprehensively, the hybrid state system (HSS) for an unmanned vehicle was built and theFSM model was used to analyze the district states combined with the safety rules ofunmanned vehicle and to control the vehicle under the safety condition. Then the co-simulation with Simulink/Stateflow and PreScan were built to verify the algorithm. The resultof the simulation showed that it was able to help the unmanned vehicle to interact with themanned vehicle safely in the intersection without traffic light.In the scenario of multiple unmanned vehicles driving through the intersection, theintersection control system was set up based on the idea of hierarchical control. The proposedsystem can be divided into macroscopic intersection controller (MIC), top level controller(TLC) and the low level controller (LLC). Among them, MIC obtained the states informationof the unmanned vehicles approaching intersection through wireless, calculated theunmanned vehicles’ safety speed via the collaboration algorithm, and transmitted the safetyspeed to the unmanned vehicle respectively in the real-time. The TLC and LLC helpedvehicle to track the target velocity. The simulation experiments were built with PreScan and Matlab/Simulink. The experiments result showed that the collaboration algorithm based oncollision prediction and avoidance improves the efficiency of the road traffic safety whileguarantees the vehicles’ safety.In order to verify the multi-vehicle collaboration algorithm under the real condition (thereal unmanned vehicle in the real circumstance), the longitudinal control system of unmannedvehicle was designed based on the platform of Dong Feng Motors (DFM). The proposedframework to realize the longitudinal control of the unmanned vehicle includes the electroniccontrol brake system, electronic control throttle system, the acquisition of velocity and aimvelocity control. The experiment platform for research combines two vehicles--the unmannedvehicle and a manned vehicle. GPS and Inertial Navigation System (INS) were used to getthe vehicles’ position and heading information in the platform, and ZigBee modules wereused to communicate between the vehicles. The experiments results showed that thecollaborative algorithm was able to be used in the real intersection without traffic lights, andimproved the driving safety of the unmanned vehicle.更多还原