In urban driving scenarios, autonomous vehicles are expected to conform to traffic rules covering traffic lights, traversable and non-traversable traffic lines, etc. In this article, we propose an optimization-based integrated decision-making and control scheme for urban autonomous driving. Inherently, to ensure the compliance with traffic rules, an innovative design of potential functions (PFs) is presented to characterize various traffic rules that are commonly encountered in urban driving scenarios, and these PFs are further incorporated as part of the model predictive control (MPC) formulation. In this sense, it circumvents the necessity of typical hand-crafted rule design, and high-level decision-making is attained implicitly along with control as an integrated architecture, facilitating flexible maneuvers with safety guarantees. As demonstrated from a series of simulations in CARLA, it is noteworthy that the proposed framework admits real-time performance and high generalizability.

翻译：在城市驾驶场景中，自动驾驶车辆需遵循涵盖交通信号灯、可通行及不可通行交通标线等在内的交通规则。本文提出一种基于优化的城市自动驾驶集成决策与控制方案。本质上，为确保交通规则合规性，本文创新性地设计了势函数用于刻画城市驾驶场景中常见的各类交通规则，并将这些势函数进一步纳入模型预测控制框架中。由此，该方法规避了典型的手工规则设计需求，并通过集成架构隐式实现了高层决策与控制的一体化，从而在保障安全性的前提下实现灵活机动。经CARLA仿真平台系列实验验证，所提框架兼具实时性与高泛化能力。