Pride and Prejudice: Toward an Information-Theoretic Framework for Mutually Communicative Driver Behavior Modeling

2026-06-15

AI Summary

This study addresses unsafe and inefficient interactions in mixed traffic arising from misjudged intentions between autonomous and human-driven vehicles. The authors model lane-changing as an implicit bidirectional communication process, wherein agents express their own intentions while inferring others’ preferences to enable cooperative decision-making under cognitive uncertainty. They introduce the novel “Pride-Inquiry” and “Pride-Prejudice” analytical planes to explicitly characterize reciprocal communication and cognitive uncertainty for the first time, alongside an adaptive utility weight for communication efficacy. Their approach integrates level-k Bayesian persuasion games, virtual signaling features, information-theoretic rewards, and communication-aware multi-agent inverse reinforcement learning (C-MIRL). Evaluation on the NGSIM dataset demonstrates a 20% reduction in forced lane-change prediction error, and driver subjective ratings show significant positive correlation with communication variables, validating the model’s effectiveness.

Abstract

Mixed autonomy driving becomes unsafe and inefficient when autonomous vehicles (AVs) and human-driven vehicles (HVs) misread each other's intentions. We study this problem as implicit mutual communication in lane changes. The proposed framework models how the ego vehicle both expresses its intent and probes the other driver's preference under epistemic uncertainty. It combines a level-k Bayesian persuasion game with virtual features for proactive signaling, information-theoretic rewards for mutual communication, and adaptive weights of communication affordances. We further introduce the Pride-Inquiry (P-I) and Pride-Prejudice (P-P) planes to analyze communication intensity and tendency. The model is calibrated with a Communication-Based Multi-Agent Inverse Reinforcement Learning algorithm (C-MIRL) on the naturalistic NGSIM dataset. Compared with the non-communicative baseline, the proposed model reduces the prediction error of mandatory lane changes by up to 20% while maintaining strong generalization. Driver-In-the-Loop questionnaire scores are positively correlated with the calibrated communication variables, supporting the subjective validity of the model. The learned rewards further show that inquiry and listening affordances contribute more than pride and expression alone, and that inquiry preference varies more strongly across drivers. These results support explicit modeling of mutual communication and epistemic uncertainty in interactive driving.

Problem

• mutual communication
• mixed autonomy driving
• epistemic uncertainty
• lane change
• driver intent

Innovation

• mutual communication
• information-theoretic reward
• Bayesian persuasion game
• adaptive communication affordances
• C-MIRL