NRI: Addressing Safe Interaction Between Autonomous and Human-Driven Vehicles

联网和自动化车辆是机器人系统，表现出显著的计算能力和物理复杂性。当它们在复杂的环境中相互作用时，它们有能力在没有人为干预的情况下独立做出情境决定。车辆与万物通信的数据和共享信息带来了重大的技术挑战，并将建模和控制的复杂性提升到了一个新的水平。预计联网和自动驾驶汽车将逐渐渗透市场，与人类驾驶的汽车互动，并应对车辆到一切的通信限制，例如带宽，辍学，错误和延迟。然而，联网和自动驾驶汽车的不同普及率会显著改变运输效率和安全性。这笔拨款将协同整合人类驾驶行为与控制理论和学习，以开发数据驱动的方法，使联网和自动驾驶车辆具有变革性的新功能，从而安全有效地与人类驾驶的车辆进行交互。在教育和推广方面，这项研究是激发人们对科学、技术、工程和数学学科兴趣的极好催化剂。这项研究的结果将提供新的方法来解决网联和自动驾驶汽车的最优轨迹规划和安全关键控制之间的根本差距。研究框架是在三个相互依存的维度，即人类驾驶行为、控制理论和学习的交叉点组织起来的。人类驾驶维度将增强我们对人类驾驶员如何应对不同驾驶场景的理解。控制理论维度将创建有关适当处方函数的知识，这些函数将根据人类驾驶行为为联网和自动驾驶车辆提供最佳决策和规划。学习维度将创造有关联网和自动驾驶汽车如何在遇到与人类驾驶不同的行为时学习调整决策和计划的知识。因此，这个维度不仅可以提高联网和自动驾驶车辆的鲁棒性，还可以提高它们在遇到任何不同驾驶行为时的操作范围。此次研究的预期成果将是制造与人类驾驶的车辆相协调的联网和自动化车辆，以提高安全性，减少污染、能源消耗和旅行延误。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。

Connected and automated vehicles are robotic systems which exhibit significant levels of computational capability and physical complexity. They have the capacity to make contextually decisions independently, without human intervention, while they interact in a complex environment. The data and shared information through vehicle-to-everything communication are associated with significant technical challenges and gives rise to a new level of complexity in modeling and control. It is expected that connected and automated vehicles will gradually penetrate the market, interact with human-driven vehicles, and contend with vehicle-to-everything communication limitations, e.g., bandwidth, dropouts, errors, and delays. However, different penetration rates of connected and automated vehicles can significantly alter transportation efficiency and safety. This grant will synergistically integrate human-driving behavior with control theory and learning in developing data-driven approaches that will enable a transformative new functionality of connected and automated vehicles to interact with human-driven vehicles safely and efficiently. On the education front and outreach, the research is an excellent catalyst for motivating interest in science, technology, engineering, and mathematics disciplines. The outcome of this research will deliver new methods to address a fundamental gap between optimal trajectory planning of and safe-critical control in connected and automated vehicles. The researched framework is organized at the intersection of three interdependent dimensions, namely, human-driving behavior, control theory, and learning. The human-driving dimension will enhance our understanding on how human drivers will respond to different driving scenarios. The control theory dimension will create knowledge on the appropriate prescription functions that will yield the optimal decisions and planning of connected and automated vehicles with respect to human driving behavior. The learning dimension will create knowledge of how connected and automated vehicles can learn to adapt their decisions and planning in situations where they encounter different behavior from what they already know about human driving. Thus, this dimension will not only improve the robustness of connected and automated vehicles but also their operation range with respect to any different driving behavior that they might encounter. The expected outcome of this research will aim at making connected and automated vehicles to coordinate with human-driven vehicles to improve safety and reduce pollution, energy consumption, and travel delays.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.