基于横向和纵向协调控制的车辆队列最优变道研究

OPTIMAL LANE CHANGE ANALYSIS FOR VEHICLE PLATOONING BASED ON LATERAL AND LONGITUDINAL CONTROL

周润发

11849024

硕士

机械工程领域工程

杨再跃

2020-06-04

2020-07-20

哈尔滨工业大学

深圳

在过去的几十年，随着集成和通信技术应用于运输系统，车辆队列逐渐成为一种新的运输方式，人们普遍认为它可以提高运输网络的能源效率和安全性。与人类驾驶相比，队列自动驾驶可以缩短车间距离并允许更多的车辆通过同一路段，因此可以更好地利用道路基础设施。在研究车辆队列系统时，必不可少的任务是通过变道的方法加入新车辆来有效地扩展车辆队列。安全、舒适、节省油耗是车辆队列行驶追求的目标，本文以扩展车辆队列为切入点，在追求上述目标的基础上研究了车辆队列的最优变道问题，全文的主要工作概括如下：本文将所研究问题在纵向和横向上进行了解耦：纵向上基于队列误差模型设计了实现车辆队列链稳定的控制器，以实现车辆队列纵向上的稳定行驶；横向上基于方向盘控制模型采用了模型预测控制（Model Predictive Control, MPC）确保变道的流畅完成。研究了变道因素对队列系统速度波动的影响，设计了一个波动评价指标来评估变道引起的波动。在两种行驶场景（单辆车请求变道加入队列和队列中的车请求加入队列）里，采用控制变量法探究了变道位置和变道速度对队列系统速度波动的影响，并用设计的算法找到了最小波动评价指标对应的最优变道速度和变道距离。接着进一步研究了波动评价指标和车辆油耗、驾驶舒适度、跟踪性能等指标的联系。基于提出的波动评价指标，为了降低变道引起的队列速度波动并提高控制效果，为队列误差模型设计了考虑安全距离、跟车性能、速度恢复能力等因素的MPC控制器。然后以波动评价指标为标准，与通用方法设计的控制器进行了对比，仿真结果表明了新设计方法的有效性。为了协调变道车采用最佳波动评价指标对应的变道速度和距离变道得到的参与队列行驶的纵向轨迹，从横向层面对变道轨迹做出了规划。研究了变道完成时间对变道轨迹的影响，设计了综合考量驾驶舒适度、油耗、变道时间三个因素的目标函数，通过加入动态罚函数的方法减少了目标函数的约束量，并用人工鱼群算法求解出了横向最优变道轨迹。通过结合纵向行驶轨迹和横向最优轨迹得到完整的最优变道轨迹，最后采用模型预测控制对最优变道轨迹进行了跟踪。

In the past few decades, with the development of integrated applications and communication technologies applied to transportation systems, vehicle platoon has gradually become a new mode of transportation. It is generally believed that it can improve the energy efficiency and safety of transportation networks. Compared with human driving, platoon driving can shorten the distance between vehicles and allow more vehicles to pass through the same road segment, so the road infrastructure can be better utilized. When researching the vehicle platoon system, the indispensable goal is to add new vehicles through the lane change method to effectively expand the platoon. Safety, comfort, and fuel saving are the goals pursued by vehicles in platoon. In this paper, the extended vehicle platoon is taken as the introduction point, and the optimal lane change problem is studied based on the pursuit of the above goals. The main work is summarized as follows:This paper decouples the research problem in the longitudinal and lateral directions. Based on the error dynamics in the longitudinal direction, a controller for string stability is designed to achieve stable driving in the longitudinal direction. Based on the steering wheel control model in the lateral direction, model predictive control(MPC) is adopted to ensure the smooth completion of the lane change.The influence of lane changing factors on the speed fluctuation of the platoon system is studied, and a fluctuation evaluation index is designed to evaluate the fluctuation caused by lane changing. In two driving scenarios (single vehicle requests to change lanes to join the platoon and vehicle of the platoon requests to join the platoon), the control variable method is used to explore the influence of lane change position and lane change speed on the speed fluctuation of the platoon system. Using the designed algorithm, the optimal lane change speed and lane change distance corresponding to the minimum fluctuation evaluation index are found. Then this work further investigates the relationship between fluctuation evaluation index and fuel consumption, driving comfort, tracking performance.Based on the proposed fluctuation evaluation index, in order to reduce the platoon speed fluctuation caused by lane change and improve the control effect, an MPC controller considering safety distance, following performance, speed recovery ability is designed for the error dynamics. Compared with the controller designed by the general method using the fluctuation evaluation index as the measurement standard, the simulation results show the effectiveness of the new design method.In order to coordinate the longitudinal driving trajectory obtained by the optimal lane change factors, the lane change trajectory is planned from the lateral level. The effect of lane change completion time on lane change trajectory is studied, and an objective function that comprehensively considered three factors of driving comfort, fuel consumption, and lane change time is designed. The dynamic penalty function was added to reduce the constraint amount of the objective function. Then the artificial fish swarm algorithm is used to find the optimal lane change trajectory. Finally, the model predictive control is used to track the optimal lane change trajectory.

车辆队列 模型预测控制 变道因素 波动评价指标 最优变道轨迹

platoon model predictive control lane change factor fluctuation evaluation index optimal lane change trajectory

中文

联合培养

南方科技大学

周润发. 基于横向和纵向协调控制的车辆队列最优变道研究[D]. 深圳. 哈尔滨工业大学,2020.