A Multi-Mode Surface Sensing Simulation Framework for UAS Integrating Radio-Frequency, Infrared and Optical Domains

Abstract

This paper presents the Multi-Mode Surface Sensing Simulation Framework (MM3S) for Unmanned Aerial Systems a physics-based simulation framework that generates perfectly co-registered synthetic data across radio-frequency (RF), thermal infrared (IR), and visible spectral domains from a maneuvering Unmanned Aerial Vehicle (UAV). The pipeline supports arbitrary terrain maps, material-specific electromagnetic properties, full-3D antenna pattern synthesis for complex arrays, and accelerated ray-tracing with novel early-ray-culling and precomputed normals. Doppler velocity signatures are computed from platform motion and superimposed on RF returns. The system outputs quad-view video sequences (RF power, IR radiance, visible RGB, Doppler) suitable for training cross-modal perception algorithms. Validation on complex scenarios demonstrates a 2x speedup over baseline Möller-Trumbore implementations while preserving physical accuracy. MM3S addresses the scarcity of labeled multi-modal UAV datasets and enables reproducible evaluation of sensor fusion techniques in challenged environments.