Modeling Interactions of Autonomous Vehicles and Pedestrians with Deep Multi-Agent Reinforcement Learning for Collision Avoidance

TL;DR

This paper develops a deep multi-agent reinforcement learning framework to model and improve collision avoidance systems for autonomous vehicles interacting with pedestrians, accounting for human-like behavior and uncertainty.

Contribution

It introduces a novel DMARL approach for AV-pedestrian interaction modeling, including a realistic pedestrian DRL agent and benchmarking collision mitigation performance.

Findings

AVs can effectively mitigate collisions in most scenarios.

The pedestrian DRL model learns realistic crossing behaviors.

Observation uncertainty impacts decision-making efficiency.

Abstract

Reliable pedestrian crash avoidance mitigation (PCAM) systems are crucial components of safe autonomous vehicles (AVs). The nature of the vehicle-pedestrian interaction where decisions of one agent directly affect the other agent's optimal behavior, and vice versa, is a challenging yet often neglected aspect of such systems. We address this issue by modeling a Markov decision process (MDP) for a simulated AV-pedestrian interaction at an unmarked crosswalk. The AV's PCAM decision policy is learned through deep reinforcement learning (DRL). Since modeling pedestrians realistically is challenging, we compare two levels of intelligent pedestrian behavior. While the baseline model follows a predefined strategy, our advanced pedestrian model is defined as a second DRL agent. This model captures continuous learning and the uncertainty inherent in human behavior, making the AV-pedestrian interaction a deep multi-agent reinforcement learning (DMARL) problem. We benchmark the developed PCAM systems according to the collision rate and the resulting traffic flow efficiency with a focus on the influence of observation uncertainty on the decision-making of the agents. The results show that the AV is able to completely mitigate collisions under the majority of the investigated conditions and that the DRL pedestrian model learns an intelligent crossing behavior.