Multi-Agent Reinforcement Learning for Safe Autonomous Driving Under Pedestrian Behavioral Uncertainty

TL;DR

This paper uses multi-agent reinforcement learning to create more realistic pedestrian behaviors in autonomous driving simulations, improving safety assessment accuracy.

Contribution

It introduces a MARL environment with co-trained pedestrians and an SDC, capturing pedestrian heterogeneity and behavioral uncertainty, especially in jaywalking scenarios.

Findings

Co-trained SDC achieved 78% goal success with 14% collision rate.

Jaywalking caused 62% of collisions despite being 13% of crossings.

MARL pedestrian training reduced collisions by 30% compared to single-agent RL.

Abstract

Simulation-based testing of self-driving cars (SDCs) typically relies on scripted or simplified pedestrian models that do not capture the heterogeneity and uncertainty of real human crossing behavior. This limits the realism of safety assessments, especially in scenarios involving jaywalking, which is governed by latent personality traits that the vehicle cannot observe. We hypothesize that jointly training pedestrians and the SDC with multi-agent reinforcement learning (MARL) produces more realistic interaction scenarios than training the SDC against fixed pedestrian policies, and that the resulting behavior gap between predictable and unpredictable crossings can be measured directly from trajectories. This paper describes a MARL environment in which an SDC and 12 pedestrians are co-trained using Multi-Agent Proximal Policy Optimization (MAPPO). Pedestrian locomotion follows scripted Dijkstra pathfinding, while an RL policy controls high-level go/wait decisions. Jaywalking probability depends on a per-pedestrian personality trait sampled at episode start and hidden from the SDC. In 500-episode evaluations, the co-trained SDC reached 78% of goals with a 14% collision rate, compared to 35% goals and 33% collisions for the best rule-based baseline. A speed differential metric shows that the SDC traveled 2.65 m/s faster near jaywalkers than near crosswalk users at close range (0-3 m), indicating that jaywalking encounters were not anticipated. Jaywalking accounted for 13% of crossing events but was associated with 62% of collisions. Co-training with MARL pedestrians reduced collisions by 30% relative to single-agent RL, as pedestrians learned to wait when the SDC approached at speed.