Scalable RF Simulation in Generative 4D Worlds

Radio Frequency (RF) sensing has emerged as a powerful, privacy-preserving alternative to vision-based methods for various perception tasks. However, building high-quality RF datasets in dynamic and diverse environments remains a major challenge. To address this, we introduce WaveVerse, a prompt-based, scalable framework that simulates realistic RF signals from generated indoor scenes with human motions guided by spatial paths, enabling diverse and feasible behaviors without manual trajectory design. WaveVerse features a language-guided 4D world generator and a physics-based signal simulator that enables realistic simulation of RF signals in diverse environments. It employs a phase-coherent ray tracer that preserves both spatial and temporal phase consistency. The simulated signals show high fidelity on phase-sensitive benchmarks, and closely align with both real-world collected measurements and simulations from a proprietary electromagnetic solver. When used for data augmentation, WaveVerse consistently improves performance in downstream tasks like RF imaging and human activity recognition, with gains that grow with the amount of simulated data and surpass existing methods. Code and additional materials are available on the webpage.

Wireless signals can help computers understand what is happening in an indoor space without using cameras, making them useful for privacy-preserving applications such as activity recognition and indoor perception. However, collecting large and diverse wireless-sensing datasets is difficult because real-world environments, human movements, and signal reflections are hard to control and reproduce. We introduce WaveVerse, a framework that creates realistic indoor worlds and human motions from text prompts, then simulates the wireless signals that would be received in those scenes. Instead of manually designing every room and movement path, WaveVerse uses language guidance to generate diverse environments and feasible human behaviors. It also uses a physics-based simulator to preserve important signal details, including how signal phase changes over space and time. Our simulated signals closely match both real measurements and high-quality electromagnetic simulations. When added to training data, WaveVerse improves performance on tasks such as RF imaging and human activity recognition. This makes it easier to build stronger RF-sensing systems without requiring large amounts of costly real-world data collection.