INSTINCT: Instance-Level Interaction Architecture for Query-Based Collaborative Perception

2025-09-28

AI Summary

Existing LiDAR-based multi-vehicle collaborative perception methods suffer from limited accuracy in long-range detection and occluded scenarios, while being severely constrained by stringent communication bandwidth requirements. To address these challenges, this paper proposes a query-driven, instance-level collaborative perception framework. Its core contributions are: (1) a quality-aware dynamic query selection mechanism that suppresses transmission of low signal-to-noise-ratio instances; (2) a dual-branch detection routing strategy that decouples local inference from collaborative optimization; and (3) a cross-agent local instance fusion module enabling lightweight and efficient feature alignment. Evaluated on DAIR-V2X and V2V4Real benchmarks, the method achieves absolute mAP improvements of 13.23% and 33.08%, respectively, while reducing communication bandwidth to merely 1/281 and 1/264 of the current state-of-the-art—demonstrating unprecedented balance between accuracy and efficiency.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Multi-modal Vision
• Multiagent Systems: Coordination and Collaboration

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs

Abstract

Collaborative perception systems overcome single-vehicle limitations in long-range detection and occlusion scenarios by integrating multi-agent sensory data, improving accuracy and safety. However, frequent cooperative interactions and real-time requirements impose stringent bandwidth constraints. Previous works proves that query-based instance-level interaction reduces bandwidth demands and manual priors, however, LiDAR-focused implementations in collaborative perception remain underdeveloped, with performance still trailing state-of-the-art approaches. To bridge this gap, we propose INSTINCT (INSTance-level INteraCtion ArchiTecture), a novel collaborative perception framework featuring three core components: 1) a quality-aware filtering mechanism for high-quality instance feature selection; 2) a dual-branch detection routing scheme to decouple collaboration-irrelevant and collaboration-relevant instances; and 3) a Cross Agent Local Instance Fusion module to aggregate local hybrid instance features. Additionally, we enhance the ground truth (GT) sampling technique to facilitate training with diverse hybrid instance features. Extensive experiments across multiple datasets demonstrate that INSTINCT achieves superior performance. Specifically, our method achieves an improvement in accuracy 13.23%/33.08% in DAIR-V2X and V2V4Real while reducing the communication bandwidth to 1/281 and 1/264 compared to state-of-the-art methods. The code is available at https://github.com/CrazyShout/INSTINCT.

Problem

• Reducing bandwidth demands in collaborative perception systems
• Improving LiDAR-based query interaction for occlusion scenarios
• Enhancing detection accuracy while minimizing communication overhead

Innovation

• Quality-aware filtering mechanism selects high-quality instance features
• Dual-branch detection routing decouples collaboration-relevant instances
• Cross Agent Local Instance Fusion aggregates hybrid instance features