Automatic Intersection Management in Mixed Traffic Using Reinforcement Learning and Graph Neural Networks

Abstract

Connected automated driving has the potential to significantly improve urban traffic efficiency, e.g., by alleviating issues due to occlusion. Cooperative behavior planning can be employed to jointly optimize the motion of multiple vehicles. Most existing approaches to automatic intersection management, however, only consider fully automated traffic. In practice, mixed traffic, i.e., the simultaneous road usage by automated and human-driven vehicles, will be prevalent. The present work proposes to leverage reinforcement learning and a graph-based scene representation for cooperative multi-agent planning. We build upon our previous works that showed the applicability of such machine learning methods to fully automated traffic. The scene representation is extended for mixed traffic and considers uncertainty in the human drivers' intentions. In the simulationbased evaluation, we model measurement uncertainties through noise processes that are tuned using real-world data. The paper evaluates the proposed method against an enhanced first infirst out scheme, our baseline for mixed traffic management. With increasing share of automated vehicles, the learned planner significantly increases the vehicle throughput and reduces the delay due to interaction. Non-automated vehicles benefit virtually alike.

Cooperative Maneuver Planning for Mixed Traffic at Unsignalized Intersections Using Probabilistic Predictions

“…They assign an artificial priority between cooperative vehicles, temporarily deviating from usual right-of-way regulations when it can be coordinated safely. Cooperative maneuvers can significantly increase the traffic throughput and lower waiting times, depending on the number of cooperative vehicles [2]- [4]. However, traffic control systems depend on a realistic longterm (i.e., ≥ 10 s) prediction of the road users to correctly estimate the efficiency and safety impact of a maneuver.…”

“…In early individual traffic control approaches, all vehicles are assumed to be cooperative and automated, so a simple prediction model is sufficient without compromising on safety [2]. However, in the medium term, we will still face mixed traffic with human drivers, whose behavior cannot be controlled and has to be predicted [3], [4]. This is challenging, since not only the continuous velocity of a vehicle has to be estimated, but also the discrete decision of whether it will cross or merge onto another lane, commonly referred to as gap acceptance, has to be predicted.…”