Rethinking the Role of Infrastructure in Collaborative Perception

2024-10-15

arXiv.org

AI Summary

This work quantitatively evaluates the role of roadside infrastructure data in collaborative perception (CP), clarifying its performance gains under both vehicle-centric and infrastructure-centric paradigms. We propose a novel infrastructure-centric CP paradigm, incorporating multi-source sensor fusion, cross-view feature alignment, and heterogeneous vehicle–infrastructure co-modeling, alongside a robustness evaluation framework. Through systematic experiments—first of their kind—we quantify the value of infrastructure data: its standalone incorporation improves 3D object detection AP by up to 10.30%; compared to conventional vehicle-centric CP, the infrastructure-centric paradigm achieves up to a 46.47% AP gain and demonstrates markedly superior robustness under sensor noise. These findings provide theoretical foundations and technical support for the deployment of intelligent connected infrastructure.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Multi-modal Vision
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Collaborative Perception (CP) is a process in which an ego agent receives and fuses sensor information from surrounding vehicles and infrastructure to enhance its perception capability. To evaluate the need for infrastructure equipped with sensors, extensive and quantitative analysis of the role of infrastructure data in CP is crucial, yet remains underexplored. To address this gap, we first quantitatively assess the importance of infrastructure data in existing vehicle-centric CP, where the ego agent is a vehicle. Furthermore, we compare vehicle-centric CP with infra-centric CP, where the ego agent is now the infrastructure, to evaluate the effectiveness of each approach. Our results demonstrate that incorporating infrastructure data improves 3D detection accuracy by up to 10.30%, and infra-centric CP shows enhanced noise robustness and increases accuracy by up to 46.47% compared with vehicle-centric CP.

Problem

• Quantify infrastructure data importance in vehicle-centric CP
• Compare vehicle-centric and infra-centric CP effectiveness
• Assess infrastructure's impact on 3D detection accuracy

Innovation

• Quantitatively assesses infrastructure data importance
• Compares vehicle-centric and infra-centric CP approaches
• Infra-centric CP enhances noise robustness significantly