Suicidal Pedestrian: Generation of Safety-Critical Scenarios for Autonomous Vehicles

TL;DR

This paper introduces a reinforcement learning-based suicidal pedestrian agent in the CARLA simulator to generate diverse safety-critical scenarios for testing autonomous vehicles, revealing decision errors in state-of-the-art algorithms.

Contribution

It presents a novel RL pedestrian model for automatic scenario generation, enabling free roaming and high-velocity collisions to test AV safety in diverse, realistic situations.

Findings

Effective in identifying decision errors in autonomous driving algorithms.

Uses collision-oriented metrics for scenario assessment.

Demonstrates robustness across different AV models.

Abstract

Developing reliable autonomous driving algorithms poses challenges in testing, particularly when it comes to safety-critical traffic scenarios involving pedestrians. An open question is how to simulate rare events, not necessarily found in autonomous driving datasets or scripted simulations, but which can occur in testing, and, in the end may lead to severe pedestrian related accidents. This paper presents a method for designing a suicidal pedestrian agent within the CARLA simulator, enabling the automatic generation of traffic scenarios for testing safety of autonomous vehicles (AVs) in dangerous situations with pedestrians. The pedestrian is modeled as a reinforcement learning (RL) agent with two custom reward functions that allow the agent to either arbitrarily or with high velocity to collide with the AV. Instead of significantly constraining the initial locations and the pedestrian behavior, we allow the pedestrian and autonomous car to be placed anywhere in the environment and the pedestrian to roam freely to generate diverse scenarios. To assess the performance of the suicidal pedestrian and the target vehicle during testing, we propose three collision-oriented evaluation metrics. Experimental results involving two state-of-the-art autonomous driving algorithms trained end-to-end with imitation learning from sensor data demonstrate the effectiveness of the suicidal pedestrian in identifying decision errors made by autonomous vehicles controlled by the algorithms.