智能网联重型载货卡车队列分布式协同控制

The intelligent and connected heavy-duty trucks platooning has great potential to improve fuel economy, driving safety and reduce traffic congestion. Based on the transformation relation between time, velocity and displacement, a discrete-displacement based nonlinear heavy-duty truck longitudinal dynamic model and a time delay spacing policy are established. With the preview information of road slope within limited distance, a long-short distance scale and hierarchy receding fashion optimization method of economic speed and gear-shift schedule are proposed by the decoupling of the speed ratio between engine and transmission. Considering the effect of gear-shift delay on speed, dynamic energy compensation strategy is presented. Optimal speed trajectory tracking, gap policy control and gear-shift trigger are fulfilled by distributed model predictive control (DMPC) with vehicle-to-vehicle information interaction and synchronous updated strategy. With consideration of model uncertainty, control input of adjacent car is set as bounded disturbance. Time variant terminal constraint set and robust feasible constraint set are constructed based on invariant set theory. A feasible robust DMPC is designed based on nominal model under the condition of robust strongly feasible. Sufficient condition for the robust stability of the closed-loop system is given. Parameter design of control algorithm is conducted through co-simulation of heavy-duty truck platooning in a hilly terrain highway as a typical complex traffic condition. An optimal control method featured by multi-scale global optimization coupled with distributed real-time control is finally proposed.

智能网联重型载货卡车队列具有改善燃油经济性、提高驾驶安全性和减缓交通拥堵的巨大潜力。基于时间/位移/车速转换关系，建立离散位移下非线性重卡纵向动力学模型和等时差车间距离模型。将发动机、变速器速比解耦，结合有限距离道路坡度信息，提出采用长-短距离尺度、分层构架的经济性车速及换挡规律滚动优化方法。考虑换挡延迟对车速的影响，提出换挡点能量动态补偿策略。设计车-车信息交互机制和同步更新策略，通过分布式模型预测控制器实现车速轨迹跟踪、车间距离控制，并触发换挡操作。为考虑模型不确定性，将邻车控制输入作为有界扰动，利用不变集理论构造时变终端约束集和鲁棒可行约束集以满足鲁棒强可行条件，提出采用标称模型的鲁棒可行预测控制方法，并给出系统闭环稳定的充分条件。以丘陵地形高速公路作为典型复杂交通场景，对重卡队列进行联合仿真和控制算法参数设计，提出基于“多尺度全局优化+分布式实时控制”的重卡队列最优控制方法。

丘陵地形高速公路是重卡队列所面对的典型复杂交通场景，本项目从重卡队列模型改进出发，针对车速及换挡规律优化、车速轨迹跟踪及车间距离控制、队列鲁棒稳定性分析三个方面对重卡队列的建模和最优控制问题进行深入研究，为节能型智能网联重卡队列的优化控制提供了一种新颖的思路。本项目的创新之处体现在以下三方面：（1）在车辆队列动力学方面，基于离散位移下的非线性重卡纵向动力学模型与等时差车间距离模型，将车速与换挡规律优化问题解耦，采用滚动优化思想，提出一种考虑时变坡度高速公路下的经济性车速优化方法。为降低换挡延迟引起的车速波动，提出换挡点动态能量补偿策略以优化车速轨迹跟踪性能。（2）在预测控制综合理论方面，以两辆重卡作为队列子系统，将前车控制输入视为有界扰动。运用鲁棒不变集理论构建满足鲁棒强可行条件的终端约束集和鲁棒可行约束集，增大状态可行域，提出一种采用标称模型并附加时变输入和鲁棒性约束的鲁棒可行预测控制方法并给出闭环系统稳定的充分条件。（3）在分布式控制构架设计方面，针对多参数、多约束、弱耦合、受扰动的重卡队列系统，提出一种面向工程应用的基于“多尺度滚动优化+分布式实时控制”分层构架的重卡队列最优控制方法。上层基于长-短距离尺度对车速和换挡规律进行滚动优化以确保全局最优，下层采用分布式预测控制对最优车速轨迹进行实时跟踪，并确保跟车安全和队列稳定。.本项目以第一作者发表SCI论文4篇（含JCR Ⅰ区论文3篇），EI论文1篇，参加国内学术会议3次，申请发明专利26件，包含授权发明专利10件，培养研究生10名。