Research map / Multi-agent systems and games / Human-AI and goal-conditioned
CoLMDriver: LLM-based Negotiation Benefits Cooperative Autonomous Driving
Changxing Liu, Genjia Liu, Zijun Wang, Jinchang Yang, Siheng Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add LLM negotiation without grouping (ablation) | CoLMDriver without negotiation · InterDrive-total | Driving Score (DS): 47.64 -> 9.33 | Hurt |
| add dynamic graph grouping to negotiation (ablation) | CoLMDriver negotiation · InterDrive-total | Driving Score (DS): 9.33 -> 77.29 | Improved |
| add dynamic graph grouping to negotiation (ablation) | CoLMDriver negotiation · InterDrive-total | Success Rate (SR): 0.000 -> 0.652 | Improved |
| add negotiation with grouping over no negotiation (ablation) | CoLMDriver · InterDrive-total | Success Rate (SR): 0.130 -> 0.652 | Improved |
| add critic feedback (consensus score) to negotiation (ablation) | CoLMDriver negotiation + grouping · InterDrive-total | Driving Score (DS): 77.29 -> 83.46 | Improved |
| add safety/efficiency score to critic (ablation) | CoLMDriver negotiation + grouping + consensus critic · InterDrive-total | Driving Score (DS): 83.46 -> 88.53 | Improved |
| add safety/efficiency score to critic (ablation) | CoLMDriver negotiation + grouping + consensus critic · InterDrive-total | Success Rate (SR): 0.739 -> 0.804 | Improved |
| add safety/efficiency score to critic (ablation) | CoLMDriver negotiation + grouping + consensus critic · InterDrive-total | Route Completion (RC): 91.93 -> 94.05 | Hurt |
| use CoLMDriver instead of rule-based negotiation substitute | Rule-based · InterDrive-total | Driving Score (DS): 78.38 -> 88.53 | Improved |
| use CoLMDriver instead of CoDriving | CoDriving · InterDrive-total | Driving Score (DS): 74.13 -> 88.53 | Improved |
| use CoLMDriver instead of TCP in lane changing scenarios | TCP · InterDrive-LC | Success Rate (SR): 0.00 -> 0.50 | Improved |
| use CoLMDriver instead of Rule-based in intersection crossing | Rule-based · InterDrive-IC | Success Rate (SR): 0.72 -> 0.72 | No clear effect |
Closest papers by meaning
- End-to-End Autonomous Driving Through V2X Cooperation (AAAI 2025 · similarity 0.86)
- AlignDrive: Aligned Lateral-Longitudinal Planning for End-to-End Autonomous Driving (NeurIPS 2026 · similarity 0.85)
- CoLLMLight: Cooperative Large Language Model Agents for Network-Wide Traffic Signal Control (ICLR 2026 · similarity 0.85)
- UniMotion: A Unified Motion Framework for Simulation, Prediction and Planning (NeurIPS 2025 · similarity 0.85)
- HCRMP: An LLM-Hinted Contextual Reinforcement Learning Framework for Autonomous Driving (NeurIPS 2025 · similarity 0.85)
- IntersectionZoo: Eco-driving for Benchmarking Multi-Agent Contextual Reinforcement Learning (ICLR 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
