DeepSync: A Learning Framework for Pervasive Localization using Code Synchronization on Compressed Cellular Spectrum

TL;DR

DeepSync tackles pervasive, low-power localization by leveraging compressed cellular spectrum and precise sub-sample timing estimation. It frames synchronization as a template-matching problem and solves it with a cross-attention-enhanced, temporally aware CNN architecture that aggregates information across multiple frames. A physics-informed data-generation pipeline combines RF ray-tracing with real cellular data, enabling robust training for severely degraded SNR conditions and intermodulation effects. The approach delivers real-world median localization around 2.13 m with substantial power savings and outperforms prior compressed-spectrum methods, offering GPS-like localization for asset tracking and IoT under strict energy budgets.

Abstract

Pervasive localization is essential for continuous tracking applications, yet existing solutions face challenges in balancing power consumption and accuracy. GPS, while precise, is impractical for continuous tracking of micro-assets due to high power requirements. Recent advances in non-linear compressed spectrum sensing offer low-power alternatives, but existing implementations achieve only coarse positioning through Received Signal Strength Indicator (RSSI) measurements. We present DeepSync, a deep learning framework that enables precise localization using compressed cellular spectrum. Our key technical insight lies in formulating sub-sample timing estimation as a template matching problem, solved through a novel architecture combining temporal CNN encoders for multi-frame processing with cross-attention mechanisms. The system processes non-linear inter-modulated spectrum through hierarchical feature extraction, achieving robust performance at SNR levels below -10dB -- a regime where conventional timing estimation fails. By integrating real cellular infrastructure data with physics-based ray-tracing simulations, DeepSync achieves 2.128-meter median accuracy while consuming significantly less power than conventional systems. Real-world evaluations demonstrate 10x improvement over existing compressed spectrum approaches, establishing a new paradigm for ultra-low-power localization.

Figures (7)

• Figure 1: DeepSync's architecture
• Figure 2: Localization performance of DeepSync vs garg2024litefoot in a simulated urban setting.
• Figure 3: CDF of localization
• Figure 4: Offset estimation
• Figure 5: CDF error of sample offset: (a) With different offset size, (b) With different input sizes.