Autonomous Vehicle Decision-Making Framework for Considering Malicious Behavior at Unsignalized Intersections

TL;DR

This paper presents a Q-learning based decision-making framework for autonomous vehicles to safely and efficiently navigate unsignalized intersections with potentially malicious vehicles, incorporating theory of mind and adaptive safety prioritization.

Contribution

It introduces a novel decision framework that combines first-order theory of mind inferences with adaptive safety weighting for autonomous vehicles at intersections.

Findings

Framework improves safety in simulations

Enhances decision-making under malicious behaviors

Meets predefined performance requirements

Abstract

In this paper, we propose a Q-learning based decision-making framework to improve the safety and efficiency of Autonomous Vehicles when they encounter other maliciously behaving vehicles while passing through unsignalized intersections. In Autonomous Vehicles, conventional reward signals are set as regular rewards regarding feedback factors such as safety and efficiency. In this paper, safety gains are modulated by variable weighting parameters to ensure that safety can be emphasized more in emergency situations. The framework proposed in this paper introduces first-order theory of mind inferences on top of conventional rewards, using first-order beliefs as additional reward signals. The decision framework enables Autonomous Vehicles to make informed decisions when encountering vehicles with potentially malicious behaviors at unsignalized intersections, thereby improving the overall safety and efficiency of Autonomous Vehicle transportation systems. In order to verify the performance of the decision framework, this paper uses Prescan/Simulink co-simulations for simulation, and the results show that the performance of the decision framework can meet the set requirements.