Towards Efficient 3D Object Detection for Vehicle-Infrastructure Collaboration via Risk-Intent Selection

2026-01-06

arXiv.org

AI Summary

This work addresses the tension between limited communication bandwidth and redundant feature transmission in vehicle-infrastructure cooperative perception by proposing a Risk-aware Intention-guided Selection mechanism (RiSe). RiSe uniquely integrates motion risk awareness and driving intention into feature selection. It employs a Potential Field–Trajectory Correlation Model (PTCM), grounded in potential field theory, to assess interaction risk, and couples it with an Intention-Driven BEV Area Prediction Module (IDAPM) to selectively transmit only high-fidelity features from high-risk, highly interactive regions. This paradigm shifts feature transmission from “visible areas” to “risk-critical areas.” Evaluated on the DeepAccident dataset, RiSe reduces communication volume to 0.71% of full-feature sharing while maintaining state-of-the-art detection accuracy, achieving a Pareto-optimal trade-off between bandwidth efficiency and perception performance.

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: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Security and Privacy: Large-scale security measurements

Abstract

Vehicle-Infrastructure Collaborative Perception (VICP) is pivotal for resolving occlusion in autonomous driving, yet the trade-off between communication bandwidth and feature redundancy remains a critical bottleneck. While intermediate fusion mitigates data volume compared to raw sharing, existing frameworks typically rely on spatial compression or static confidence maps, which inefficiently transmit spatially redundant features from non-critical background regions. To address this, we propose Risk-intent Selective detection (RiSe), an interaction-aware framework that shifts the paradigm from identifying visible regions to prioritizing risk-critical ones. Specifically, we introduce a Potential Field-Trajectory Correlation Model (PTCM) grounded in potential field theory to quantitatively assess kinematic risks. Complementing this, an Intention-Driven Area Prediction Module (IDAPM) leverages ego-motion priors to proactively predict and filter key Bird's-Eye-View (BEV) areas essential for decision-making. By integrating these components, RiSe implements a semantic-selective fusion scheme that transmits high-fidelity features only from high-interaction regions, effectively acting as a feature denoiser. Extensive experiments on the DeepAccident dataset demonstrate that our method reduces communication volume to 0.71\% of full feature sharing while maintaining state-of-the-art detection accuracy, establishing a competitive Pareto frontier between bandwidth efficiency and perception performance.

Problem

• Vehicle-Infrastructure Collaboration
• 3D Object Detection
• Communication Efficiency
• Feature Redundancy
• Occlusion

Innovation

• Risk-intent Selection
• Vehicle-Infrastructure Collaboration
• Semantic-selective Fusion
• Potential Field-Trajectory Correlation
• BEV Area Prediction