Robustifying 3D Perception through Least-Squares Multi-Agent Graphs Object Tracking

2025-07-07

AI Summary

To address the insufficient robustness of 3D perception in edge-intelligent systems (e.g., autonomous driving) under adversarial attacks, this paper proposes the first multi-agent collaborative 3D object tracking framework based on least-squares graph optimization. Methodologically, it innovatively integrates differential coordinates and anchor mechanisms into LiDAR-based tracking—enabling intrinsic suppression of adversarial noise without auxiliary defense modules. Inter-vehicle collaboration is modeled via a fully connected graph, and multi-view bounding boxes, overlap-region analysis, and two-stage data association are jointly leveraged to achieve accurate, persistent multi-object tracking in dynamic scenes. Evaluated on the V2V4Real dataset, our method achieves a 23.3% improvement in detection and tracking accuracy over state-of-the-art methods under stringent adversarial conditions, significantly enhancing the security and robustness of edge-side 3D perception.

Technology Category

• Computer Vision: Motion & Tracking
• Multiagent Systems: Adversarial Agents
• Machine Learning: Adversarial Learning & Robustness

Application Category

• Security and Privacy: Tracking, profiling, and countermeasures against them
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

The critical perception capabilities of EdgeAI systems, such as autonomous vehicles, are required to be resilient against adversarial threats, by enabling accurate identification and localization of multiple objects in the scene over time, mitigating their impact. Single-agent tracking offers resilience to adversarial attacks but lacks situational awareness, underscoring the need for multi-agent cooperation to enhance context understanding and robustness. This paper proposes a novel mitigation framework on 3D LiDAR scene against adversarial noise by tracking objects based on least-squares graph on multi-agent adversarial bounding boxes. Specifically, we employ the least-squares graph tool to reduce the induced positional error of each detection's centroid utilizing overlapped bounding boxes on a fully connected graph via differential coordinates and anchor points. Hence, the multi-vehicle detections are fused and refined mitigating the adversarial impact, and associated with existing tracks in two stages performing tracking to further suppress the adversarial threat. An extensive evaluation study on the real-world V2V4Real dataset demonstrates that the proposed method significantly outperforms both state-of-the-art single and multi-agent tracking frameworks by up to 23.3% under challenging adversarial conditions, operating as a resilient approach without relying on additional defense mechanisms.

Problem

• Enhancing 3D object tracking robustness against adversarial noise
• Improving multi-agent cooperation for resilient perception in EdgeAI
• Reducing adversarial impact via least-squares graph fusion and refinement

Innovation

• Multi-agent least-squares graph for 3D tracking
• Differential coordinates refine adversarial bounding boxes
• Two-stage fusion suppresses adversarial noise impact