2024 NSDI NSDI 2024

Spectrumize: Spectrum-efficient Satellite Networks for the Internet of Things

Abstract

Low Earth Orbit satellite constellations are gaining traction for providing connectivity to low-power outdoor Internet of Things (IoT) devices. This is made possible by the development of low-cost, low-complexity pico-satellites that can be easily launched, offering global connectivity without the need for Earth-based gateways. In this paper, we report the space-to-Earth communication bottlenecks derived from our experience of deploying an IoT satellite. Specifically, we characterize the challenges posed by the low link budgets, satellite motion, and packet collisions. To address these challenges, we design a new class of technique that use the Doppler shift caused by the satellite's motion as a unique signature for packet detection and decoding, even at low signal-to-noise ratios and in the presence of collisions. We integrate these techniques into our system, called Spectrumize, and evaluate its performance through both simulations and real-world deployments. Our evaluation shows that Spectrumize performs 3x better compared to classic approach in detecting packet with over 80% average accuracy in decoding.

🧭 Keyword Pioneer — doppler shift detection
🐝 Cross-Pollinator — Artificial Intelligence, Computer Science, Computer Vision, Data Science & Analytics, Deep Learning, Healthcare & Medicine, Interdisciplinary, Knowledge & Reasoning, Machine Learning, Mathematics & Optimization, Natural Language Processing, Reinforcement Learning, Robotics, Security & Privacy, Speech & Audio