CAREER: Socially-Aware Control of Autonomy: Reshaping Urban Mobility in Traffic Networks with Mixed Vehicle Autonomy

随着城市在各地发展，城市道路变得不堪重负，需要有效的战略来改善城市流动性。随着自动驾驶汽车的出现，如果采取积极的控制和规划方法，就有机会回收城市流动性。在这个项目中，我们的目标是开发基础理论和领域驱动技术，以利用自动驾驶汽车提供的机会来实现移动高效的智慧城市。我们考虑混合自动驾驶交通网络，这是由人类驾驶和自动汽车共享的道路网络。我们将开发一套算法，用于这些系统中自动汽车的社会感知控制。更确切地说，我们开发的控制算法考虑到了人类驾驶和自动汽车共存的社会影响，并保证了整体社会利益。我们的项目实现了从当前现实到更光明未来的转变，其中自动驾驶汽车与道路和网络交通控制系统和谐互动，以提高机动性。为了释放自动汽车的移动潜力，一个关键的控制挑战是考虑人类如何适应和响应自动汽车的行动，如路线决策。我们专注于旅行者的路由决策，并开发设计和控制算法，诱导有效的路由决策。我们使用路由游戏建模旅客的交通网络的路线选择，并解决与控制混合自治网络的核心组件相关的研究挑战。(i)我们将开发自动汽车路线决策的利他控制算法，也就是说，自动汽车的路线对社会有利。(ii)我们将开发定价算法，影响人类的路由决策，使网络转向理想的状态。(iii)最后，我们将开发规划策略，网络拓扑结构设计和容量调节，诱导有效的路由决策的旅行者。我们提出的控制策略也将通过广泛的真实世界的实验和交通模拟验证。这个奖项反映了NSF的法定使命，并已被认为是值得通过使用基金会的智力价值和更广泛的影响审查标准进行评估的支持。

As cities grow everywhere, and urban roadways become overburdened, efficient strategies are required for improving urban mobility. With the emergence of autonomous cars, there is an opportunity to reclaim urban mobility provided that a proactive control and planning approach is taken. In this project, our goal is to develop fundamental theory and domain-driven techniques for leveraging the opportunities that autonomous cars provide to achieve mobility-efficient smart cities. We consider mixed-autonomy traffic networks which are road networks that are shared by human-driven and autonomous cars. We will develop a set of algorithms for socially-aware control of autonomous cars in these systems. More precisely, we develop control algorithms that take into account the social implications of the co-existence of human-driven and autonomous cars and guarantee overall societal good. Our project enables the transformation from the current reality to the brighter future, in which autonomous cars harmoniously interact with the road and network traffic control systems to increase mobility. To unlock the mobility potentials of autonomous cars, a key control challenge is to account for how humans adapt and respond to autonomous cars’ actions such as routing decisions. We focus on travelers’ routing decisions and develop design and control algorithms that induce efficient routing decisions. We use routing games for modeling travelers’ route choices over traffic networks and tackle research challenges associated with controlling the core components of a mixed-autonomy network. (i) We will develop algorithms for altruistic control of autonomous cars’ routing decisions, that is, autonomous cars that are routed in the favor of society. (ii) We will develop pricing algorithms that affect humans' routing decisions such that the network is steered towards desirable states. (iii) Finally, we will develop planning strategies for network topology design and capacity regulation that induce efficient routing decisions by travelers. Our proposed control strategies will also be validated through extensive real-world experiments and traffic simulations.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.