无人驾驶车辆换道控制方法研究

发布时间：2018-11-25 11:52

【摘要】：随着人们对汽车智能化要求的不断提高,无人驾驶汽车作为智能驾驶的核心成为了最为关注的技术。而车道变换是无人驾驶中最常见的行为,很大程度上影响道路的流通性,快速而安全的换道操作对减少交通事故具有极强的现实应用意义。本文以无人驾驶汽车为研究对象,针对换道工况下的路径规划和轨迹跟踪进行了研究,设计出一种简单有效的路径规划方法和轨迹跟踪控制器,以优化跟踪效果和降低控制器在线计算量。同时探讨了考虑侧倾运动时的轨迹跟踪效果。主要的研究内容包括以下四个方面：(1)针对行驶时需要换道的不同工况,设计了综合的换道策略。给出前后车换道时不发生碰撞的换道安全车距,再利用多项式设计出汽车在没有障碍物和有障碍物两种工况下的换道路径。(2)建立无人驾驶汽车的运动学模型、三自由度整车动力学模型和魔术轮胎模型,搭建Simulink模型并与CarSim模型进行联合仿真,仿真结果表明Simulink数学模型和CarSim模型具有相近的动力学特性。(3)将非线性汽车动力学系统进行线性化处理,设定约束条件和优化目标函数,在此基础上设计基于Laguerre函数的模型预测器和基于Legendre函数的模型预测器以跟踪换道轨迹,并运用Simulink和CarSim进行联合仿真,仿真结果表明基于Legendre函数的模型预测控制器对路面附着系数具有鲁棒性,同时还具有降低汽车动力学参数波动和在线计算量的优点。(4)建立了包含侧倾的四自由度整车动力学模型,设计了基于Legendre函数的模型预测控制器,并运用Simulink和CarSim对模型预测控制系统进行联合仿真,仿真结果表明考虑侧倾的无人驾驶系统跟踪效果更好,同时系统响应速度也更快。
[Abstract]:With the increasing demand for intelligent automobile, driverless vehicle as the core of intelligent driving has become the most concerned technology. Lane change is the most common behavior in driverless traffic, which largely affects the flow of roads. Rapid and safe operation of changing lanes has a strong practical significance in reducing traffic accidents. In this paper, a simple and effective path planning method and a trajectory tracking controller are designed to study the path planning and trajectory tracking of driverless vehicles. In order to optimize the tracking effect and reduce the on-line calculation of the controller. At the same time, the track tracking effect considering the roll motion is discussed. The main research contents include the following four aspects: (1) according to the different driving conditions, a comprehensive changing strategy is designed. In this paper, the safe distance between the front and rear vehicles without collision is given, and then the path of the vehicle is designed by polynomial. (2) the kinematics model of the driverless vehicle is established, and the path of the vehicle is designed under the condition of no obstacle and no obstacle. (2) the kinematics model of the driverless vehicle is established. The three-degree-of-freedom vehicle dynamics model and the magic tire model are built, the Simulink model is built and the CarSim model is simulated jointly. The simulation results show that the Simulink mathematical model and the CarSim model have similar dynamic characteristics. (3) the nonlinear vehicle dynamics system is linearized, the constraint conditions are set and the objective function is optimized. On this basis, a model predictor based on Laguerre function and a model predictor based on Legendre function are designed to track the track change, and Simulink and CarSim are used for joint simulation. The simulation results show that the model predictive controller based on Legendre function is robust to the road adhesion coefficient. At the same time, it has the advantages of reducing the fluctuation of vehicle dynamics parameters and the amount of calculation on line. (4) A four-degree-of-freedom vehicle dynamics model including roll is established, and a model predictive controller based on Legendre function is designed. The model predictive control system is simulated by Simulink and CarSim. The simulation results show that the driverless system with roll tilting has better tracking effect and faster response speed.
【学位授予单位】：东南大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：U463.6

【相似文献】