异构车联网环境下基于模型预测控制的车辆多性能巡航控制

Multi-performance cruise control with safety, economy and ride comfort has been listed as one of leading directions of automatic driving by the national new generation artificial intelligence development plan.The purpose of this project is to study model predictive control-based multi-performance cruise control methods of vehicles under heterogeneous vehicular network environments (i.e. networked vehicles). It includes: By analyzing transmission characteristics of heterogeneous vehicular networks and characteristics of vehicles cruising, the cruise systems of the heterogeneous networked vehicles are modelled as a class of the discrete-event driven switched models using the switching methods of networked systems; By combining the parameterization idea and receding horizon optimization principle, we study multi-performance cruise MPC schemes of heterogeneous networked vehicles. Then the stability synthesis methods are proposed for the multi-performance cruise MPC by control Lyapunov functions of nonlinear systems and the switching control theory; Using the concept of robust control Lyapunov functions and small gain theorem, we further develop the stability re-design methods of the cruise MPC to raise the robust stability and disturbance rejection level of the multi-performance cruise MPC of heterogeneous networked vehicles under uncertain disturbances. These make a breakthrough in performance bottleneck of cruise systems in heterogeneous vehicular networks and uncertain disturbances. The correctness of the theoretical results obtained in the project is verified by a physical simulation platform. This project will provide theoretical support for research on the automatic driving technologies of networked vehicles. Also, it is of great significance to promote applications of model predictive control in the emerging fields such as automatic driving.

具备安全、经济和舒适的多性能巡航控制被国家新一代人工智能发展规划列为自动驾驶的先导方向。本项目旨在研究异构车联网环境下车辆（即网联车辆）基于模型预测控制的多性能巡航控制方法，包括：分析异构车联网的传输特性和车辆巡航特点，应用网络化系统切换思想，建立异构网联车辆巡航系统的不受控离散事件驱动切换模型；结合参数化思想和滚动时域优化原理，研究异构网联车辆多性能巡航MPC算法，并用非线性系统控制Lyapunov函数和切换控制理论，提出该巡航MPC稳定性综合方法；利用鲁棒控制Lyapunov函数和小增益定理开展巡航MPC稳定性再设计，提高异构网联车辆多性能巡航MPC在不确定性作用下的鲁棒稳定性和扰动抑制水平，突破巡航系统在异构车联网和不确定扰动下的性能瓶颈；利用实物模拟平台，验证项目理论成果的正确性。本项目将为网联车辆自动驾驶技术研究提供理论支撑，对预测控制在自动驾驶等新兴领域的推广应用具有重要意义。

针对智能网联车辆安全、经济和舒适巡航控制问题，本项目研究了异构车联网环境下车辆基于模型预测控制的多性能巡航控制方法。项目应用网络化系统切换思想，建立了异构网联车辆巡航系统的不受控离散事件驱动切换模型，再结合参数化思想和滚动时域优化原理，设计了异构网联车辆多性能巡航MPC算法，并用控制Lyapunov函数和切换控制理论，提出该巡航MPC稳定性综合方法；针对通信过程中可能存在的延时、丢包和有限通信资源问题，设计双向非对称的延时反馈控制器并证明其闭环稳定性；采用伯努利随机过程和具有可变输入延迟的跟踪模型描述具有参数不确定性和丢包、时延的异构通信网联车系统，提出了网联车系统鲁棒协同自适应巡航控制器设计方法，提高了异构网联车辆在不确定性作用下的鲁棒稳定性和扰动抑制水平；针对未知动态特性全自动驾驶车辆队列CACC问题，提出了一种基于动态自适应规划的车队CACC策略，突破巡航系统在异构车联网和不确定扰动下的性能瓶颈；利用实物模拟平台，验证项目理论成果的正确性。项目执行发表文章12篇，其中SCI收录期刊论文8篇，EI收录期刊论文9篇，会议论文2篇；授权国家发明专利4项，申请国家发明专利3项，授权软件著作权1项。获得中国发明家协会发明创业二等奖1项。结合本项目的研究工作，培养硕士研究生5名。