---
title: "G-MARK: Grounded Multi-Agent Reasoning for Cooperative Driving via Knowledge Graphs"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.19964v1"
permalink: "https://feed.astralor.com/items/itm_e7998e6f596bee133396fd4c"
published: "2026-08-20T12:35:12.000Z"
score: 77
channel: "ai"
tags: ["G-MARK","Vehicle-to-Vehicle Cooperation","Multi-Agent Reasoning","Knowledge Graph","Autonomous Driving"]
---

# G-MARK: Grounded Multi-Agent Reasoning for Cooperative Driving via Knowledge Graphs

## AI commentary

> What stands out here is the shift from compressed latent features to explicit provenance-aware knowledge, which lets the system answer who saw what—critical for safety-critical autonomous driving. It is especially useful for researchers working on V2X cooperative perception, occlusion reasoning, or interpretable AI.

## At a glance

Autonomous driving systems operate under partial observability, where safety-critical objects may be occluded or visible only to neighboring connected vehicles. Existing cooperative driving methods often compress multi-agent evidence into latent features, obscuring source attribution and conflicting evidence. G-MARK converts cooperative object-centric observations into provenance-aware knowledge graphs that preserve object hypotheses with source attribution, ego-versus-partner visibility, uncertainty, conflicts, spatial relations, and planning context. Lightweight task heads built on the shared KG representation support object reasoning, motion prediction, control selection, and trajectory forecasting. Compared with the state-of-the-art baseline, G-MARK improves occlusion reasoning accuracy by 42.2%, reduces control-selection error by 13.1%, and achieves comparable trajectory-planning accuracy with a 25.6x smaller structured communication payload.

## Key points

- Proposes an explicit, interpretable cooperative driving reasoning framework that converts multi-vehicle observations into provenance-aware knowledge graphs.
- Preserves occlusion, conflict, and uncertainty information to support object reasoning, motion prediction, control selection, and trajectory forecasting.
- Improves occlusion reasoning accuracy by 42.2% and reduces control-selection error by 13.1% over the SOTA baseline.
- Reduces structured communication payload by 25.6x while maintaining comparable trajectory-planning accuracy.

## Full text

Autonomous driving systems must operate under partial observability, where safety-critical objects may be occluded or visible only to neighboring connected vehicles. Vehicle-to-vehicle cooperation can reduce this uncertainty, but existing cooperative driving methods often compress multi-agent evidence into latent features or hidden multimodal states. As a result, they obscure which agent observed each object, whether the object is visible to the ego vehicle, and how conflicting evidence affects downstream decisions. We propose G-MARK, a grounded multi-agent reasoning framework that converts cooperative object-centric observations into explicit provenance-aware knowledge graphs (KGs). The resulting KGs preserve object hypotheses together with their source attribution, ego-versus-partner visibility, uncertainty, conflicts, spatial relations, and planning-relevant context. G-MARK then derives a shared feature representation from these KGs, enabling lightweight task heads to support object reasoning, motion prediction, control selection, and trajectory forecasting. Compared with the state-of-the-art baseline, GMARK improves occlusion reasoning accuracy by 42.2%, reduces control-selection error by 13.1%, and achieves comparable trajectory-planning accuracy with a 25.6x smaller structured communication payload. Our code is available at https://github.com/bhavyagupta98/g-mark.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_e7998e6f596bee133396fd4c