Recognize then Resolve: A Hybrid Framework for Understanding Interaction and Cooperative Conflict Resolution in Mixed Traffic

TL;DR

This work tackles unsafe and inefficient CAV-HDV interactions at mixed intersections by introducing the Recognize then Resolve (RtR) framework. RtR combines a data-driven Bilateral Intention Progression Graph (BIPG) to detect interaction breakdown and infer HDV intentions with a constrained Monte Carlo Tree Search (MCTS) to compute an optimal, HDV-aligned passing order, formalized by $O^* = \arg\max_O \sum_{i=1}^N R(S_i,o_i)$ subject to $o_h = I_h \in \{0:\text{rush},1:\text{yield}\}$. Key contributions include the BIPG model, the definition of three interaction breakdown scenarios, and a reward-based constrained MCTS that enforces HDV intentions, together with extensive experiments showing RtR improves safety (via PET metrics) and efficiency across penetration rates while reducing computation by about 50% compared to fully cooperative planning. The results demonstrate RtR's potential to enable scalable, real-time mixed-traffic conflict resolution, with performance approaching consistent cooperation even under substantial HDV heterogeneity. The framework thus offers practical impact for deploying safer, more efficient CAVs in real-world mixed traffic.

Abstract

A lack of understanding of interactions and the inability to effectively resolve conflicts continue to impede the progress of Connected Autonomous Vehicles (CAVs) in their interactions with Human-Driven Vehicles (HDVs). To address this challenge, we propose the Recognize then Resolve (RtR) framework. First, a Bilateral Intention Progression Graph (BIPG) is constructed based on CAV-HDV interaction data to model the evolution of interactions and identify potential HDV intentions. Three typical interaction breakdown scenarios are then categorized, and key moments are defined for triggering cooperative conflict resolution. On this basis, a constrained Monte Carlo Tree Search (MCTS) algorithm is introduced to determine the optimal passage order while accommodating HDV intentions. Experimental results demonstrate that the proposed RtR framework outperforms other cooperative approaches in terms of safety and efficiency across various penetration rates, achieving results close to consistent cooperation while significantly reducing computational resources. Our code and data are available at: https://github.com/FanGShiYuu/RtR-Recognize-then-Resolve/.

Figures (8)

• Figure 1: Main challenges of conflict resolution in mixed traffic.
• Figure 2: Overview of the proposed RtR framework for interaction breakdown recognition and conflict resolution.
• Figure 3: Illustrations of three typical interaction breakdown scenarios.
• Figure 4: A CAV takeover case while interacting with HDV at un-signalized intersection.
• Figure 5: Variations in success rate and computation time under different penetration rates.