Collaborative Research: Smart Vehicle Platooning Built upon Real-Time Learning and Distributed Optimization

新兴的联网和自动驾驶汽车（CAV）技术为减少交通拥堵和提高交通效率提供了巨大的潜力。然而，许多与自动驾驶汽车相关的工作都集中在单个车辆的安全性上，当混合交通（自动驾驶汽车和人工驾驶汽车）在道路上相互作用时，这会影响交通效率。本项目旨在研究一组自动驾驶汽车如何通过协同速度或加速度控制来应对人为驾驶车辆、变道请求以及异常交通和网络条件引起的外部干扰。该研究将提高未来包括自动驾驶汽车在内的交通系统的道路安全和交通效率。该项目将向更广泛的受众传播研究和教育成果，包括代表性不足的大学和K-12学生，特别关注少数民族学生。本项目的具体研究目标是在各种正常或异常交通和/或网络条件下，开发以车辆队列为中心的最优、自适应和弹性车辆队列控制。该项目将开发(a)集成分布式优化的先进模型预测控制，在正常交通/网络条件下实现最佳车辆队列控制；(b)基于混合整数规划的自适应变道请求的最优车辆队列控制模型预测控制；(c)在异常交通和/或网络条件下集成实时学习和分布式优化的弹性车辆队列控制。该项目将整合多个领域的最新技术，包括交通流、控制、优化、学习和分布式计算，并将为复杂现实交通条件下的协调和自动化车辆排中心交通控制奠定跨学科基础。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Emerging connected and autonomous vehicle (CAV) technologies offer great potentials to reduce traffic congestion and improve traffic efficiency. However, much of the CAV related work focuses on individual vehicles's safety, which compromises traffic efficiency when mixed traffic (CAVs and human-driven vehicles) are on the road interacting with each other. This project aims to study how a group of CAVs can respond to exogenous disturbances resulting from human-driven vehicles, lane change requests and abnormal traffic and cyber conditions through cooperative speed or acceleration control. The research will improve road safety and traffic efficiency of future transportation systems involving CAVs. This project will disseminate research and education outcomes to broader audiences, including under-represented college and K-12 students with a particular focus on minority students. The specific research objectives of this project are to develop vehicle platoon centered optimal, adaptive, and resilient vehicle platooning control under various normal or abnormal traffic and/or cyber conditions. The project will develop (a) advanced model predictive control integrating distributed optimization for optimal vehicle platooning control under normal traffic/cyber conditions; (b) mixed integer programming based model predictive control for optimal vehicle platooning control adaptive to lane change requests; (c) resilient vehicle platooning control integrating real-time learning and distributed optimization under abnormal traffic and/or cyber conditions. The project will integrate the state of the art from multiple fields including traffic flows, control, optimization, learning, and distributed computation and will establish an interdisciplinary foundation for coordinated and automated vehicle platoon centered traffic control under complex real-world traffic conditions.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.