基于滚动优化的车辆自适应巡航控制

发布时间：2018-02-27 18:33

本文关键词： 自适应巡航控制(ACC) 多目标优化模型预测控制(MPC) 变权重系数方法多模式的自适应巡航系统　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：自适应巡航(ACC)系统能够代替驾驶员对车辆的纵向进行控制,有效地减少了人为因素引发的交通事故,近年来得到了学者和汽车厂商的广泛关注。针对目前ACC系统在设计过程中对舒适性和燃油经济性考虑不充分,且只能在特定简单工况下实现跟车功能的特点,本文提出在模型预测控制(MPC)框架下进行系统设计,同时兼顾车辆行驶过程中的安全性、跟车性、舒适性以及燃油经济性。为了适应城市工况复杂多变的场景,提出变权重系数的方法以适应不同工况的驾驶需求,同时针对多模式的ACC系统也展开研究。由于直接式控制结构会影响系统的鲁棒性,因此本文采用分层的控制方案来完成跟车系统的设计。对于上位控制器需要考虑车辆行驶过程中的多个目标,包括安全性、跟踪性、舒适性以及燃油经济性,同时还要考虑实际车辆存在的各种约束,因此,选用能有效处理多目标和系统约束的模型预测控制(MPC)方法;对于下位控制器要求实际加速度快速跟踪上期望值,因此,选用前馈加反馈的控制方法,前馈控制有助于消除系统中非线性项的影响,反馈控制采用PID控制算法,能够满足系统的快速性要求。为了提高ACC系统对环境的适应能力,本文提出一种变权重系数的方法,针对不同的行驶工况,采用模糊理论在线辨识目标函数中跟车性以及舒适性的权重系数,达到不同工况下对二者侧重不同的目的。为了验证分层控制策略的有效性,针对五种典型的交通场景分别进行多组离线仿真验证。结果表明,在不同的场景下本文设计的ACC系统均具有较好的跟踪效果,同时车辆在行驶过程中能够兼顾安全性与舒适性等多个行驶目标,车辆的状态也具有平滑的瞬态响应。由于真实的道路环境信息是复杂多变的,并且各种因素都可能对驾驶员的决策产生影响,而单一控制器参数相对固定,从而限制了ACC系统的适用范围,因此,本文也对多模式的ACC系统展开研究。首先根据车辆行驶的紧急程度不同,将行驶工况具体划分为巡航、加速接近前车、稳态跟随、减速接近前车以及紧急跟随这五种不同的模式。针对每种模式重新调节目标函数中的权重系数以及相应的约束范围,完成每种模式下控制器的设计。为了验证多模式ACC系统的有效性,首先在五种典型交通场景下进行验证,其次,设置两组极限行驶工况,仿真结果表明,在典型场景下多模式ACC系统具有很好的跟踪效果,极限工况下,多模式ACC系统能够及早的采取相应的制动措施,有效避免碰撞事故的发生,同时能够兼顾车辆行驶的多个目标,其控制具有一定的平顺性。本文以汽车ACC系统为研究对象,根据车间运动学的高阶模型,在MPC框架下设计考虑多目标上层控制器,面向车辆逆纵向动力学模型设计前馈加反馈的下层控制器,同时为了增加对道路交通的适应能力提出变权重系数的方法。为了扩大ACC系统的适用范围,提出多模式控制的思想。多组离线仿真结果表明,典型场景下两种控制方案都具有良好的跟踪效果和动态响应,极限工况下,多模式ACC系统能够通过紧急制动避免碰撞事故的发生,从而扩大了ACC系统的适用范围。
[Abstract]:Adaptive cruise control (ACC) system can replace the driver for longitudinal control of the vehicle, and effectively reduce the traffic accidents caused by human factors, in recent years has been widespread concern of scholars and automobile manufacturers. In view of the current ACC system in the design process of comfort and fuel economy is not adequately considered, and can only be achieved with the car features the function in certain simple conditions, this paper proposed in model predictive control (MPC) system design framework, taking into account the traffic safety in the process, with the car, comfort and fuel economy. In order to adapt to the complex conditions of the city scene, put forward method of variable weight coefficient to adapt to different conditions at the same time for driving demand, multi mode ACC system is also studied. Because of the direct type control structure will affect the robustness of the system, this paper adopts the hierarchical control scheme to complete The car with the design of the system. For the host controller requires multiple objectives into consideration in the vehicle running process, including safety, tracking, comfort and fuel economy, but also to consider a variety of constraints, the actual vehicle in use can effectively handle multiple objectives and system constraints, model predictive control (MPC) method; for the requirements of the actual controller acceleration fast track the expectations, therefore, the control method of feedforward and feedback, help to eliminate the effect of nonlinearity in the system of feedforward control, feedback control using the PID control algorithm can meet the fast system requirements. In order to improve the ability to adapt to the ACC system on the environment, this paper proposes a method of variable weight coefficient, according to different driving conditions, using the fuzzy theory for online identification of objective function with the car and the weight coefficient of comfort, reached under different conditions The purpose of the two is different. In order to verify the effectiveness of the hierarchical control strategy, aiming at five typical traffic scenes were multi group offline simulation. Results show that in different situations this paper designs the ACC system has a better tracking effect, while the vehicle can satisfy the safety and comfort of a number moving target in the running process, the transient response of vehicle state also has a smooth. Because the information real road environment is complex and changeable, and various factors may affect the driver's decision, and a single controller parameter is relatively fixed, thus limiting the scope of the ACC system, so this paper also begin to study more the model of ACC system. According to the level of emergency vehicles in different driving conditions, will be divided into specific cruise speed, close to the car in front, with steady deceleration, close to the car in front and tight Anxious to follow these five different modes. For each mode re adjust the weight coefficients in the objective function and constraint range, complete the design of controller for each mode. In order to verify the effectiveness of multi mode ACC system, first verify, in five typical traffic scenarios, secondly, set up two sets of limit driving conditions, the simulation results show that in typical scenarios of multi mode ACC system has a good tracking effect, under extreme conditions, the multi mode ACC system can take corresponding measures to brake, effectively avoiding collision, and can take into account the multiple target vehicle, the control has a certain comfort. Based on vehicle ACC system as the research object, according to the high order model workshop kinematics, multi targets upper controller design under the framework of MPC, the vehicle longitudinal dynamics model for inverse design of feedforward inverse The lower controller is fed at the same time, in order to increase the ability to adapt to road traffic is proposed variable weight coefficient. In order to expand the scope of the ACC system, the multi mode control idea. Many groups of offline simulation results show that the two control schemes of typical scenes has good tracking effect and dynamic response, under extreme conditions, multi mode ACC system can use the emergency brake to avoid collision accidents, so as to expand the scope of the ACC system.

【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U463.6;TP273

【相似文献】