Distributed UAV-SAR Synchronization in GPS-Denied Environments: Oscillator Phase Noise Modeling and Performance Analysis

Abstract

This work quantifies the impact of stochastic oscillator phase noise on distributed radar synchronization in GPS-denied networks. A unified oscillator model captures both short-term jitter and long-term wander phase noise components by jointly fitting manufacturer-provided phase noise power spectral density (PSD) and Allan deviation parameters from commercial off-the-shelf (COTS) oscillators through a weighted-least-squares (WLS) approach. Time-domain realizations are generated via frequency-domain synthesis and extended to a fast- and slow-time oscillator simulation framework that allows for efficient simulation across timescales for distributed UAV-SAR. The oscillator simulator drives a two-stage frequency syntonization and fine time/phase alignment system, allowing evaluation of how oscillator class and network topology impact system-level coherence. Two inter-event tracking approaches--a linear window estimator and Kalman filter--are introduced to further reduce time wander between synchronization events. Together, these contributions extend prior work by incorporating a more realistic oscillator model inclusive of frequency, time, and phase offsets as temporal stochastic processes, rather than as constants over a single model run. Monte-Carlo simulation results across representative oscillator classes, signal-to-noise ratios (SNRs), and synchronization repetition frequencies (SRFs) distinguish noise-limited and oscillator-limited regimes and provide concrete update-rate guidelines for maintaining coherent multi-static operation.