HybridSim simulator generates realistic mmWave radar signals for human sensing

Researchers have developed HybridSim, a novel physics-learning hybrid simulator designed to generate realistic mmWave radar signals for dynamic human motion. This tool synthesizes signals by decoupling propagation into direct and indirect paths, using a tri-plane representation and Graph Convolutional Network for human features and a microfacet BRDF for surface reflections. HybridSim also approximates indirect paths with 3D Gaussian Splatting and a virtual-receiver geometry, significantly reducing computational costs compared to full ray tracing. The simulator demonstrates improved agreement with physical models and enhances downstream radar-based human sensing tasks through site-specific data augmentation. AI

IMPACT Enables more efficient development and augmentation of radar-based human sensing models.

RANK_REASON The cluster contains a research paper detailing a new simulation method. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• 3D Gaussian splatting
• arXiv
• bidirectional reflectance distribution function
• graph convolutional network
• mmWave sensing

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →