Statistical Multistatic Radar with Imperfect Time Synchronization

Abstract

A radar system observes radio-frequency (RF) reflections in order to detect targets and estimate where they are, what they are, and how they are moving. Distributed multistatic radar refers to a type of radar system in which signals transmitted and received at multiple spatially separated sites are processed jointly. A practical issue in multistatic radar is establishing a common reference for time and frequency between distributed sites. Without ideal timing coordination, signals at distributed sites will be misaligned in time, complicating processing and causing additional error in range measurement. Without ideal frequency synchronization, target Doppler frequency measurements will contain additional error. In this thesis, the fundamental topics of detection and estimation using a multistatic system are examined when synchronization is imperfect. A flexible system model is developed to describe multistatic configurations in a standardized way. Based on this system model, a signal model is derived to describe the signals received across the system, including the impact of inter-site synchronization errors. Frameworks for analyzing the core radar capabilities of detection and estimation are developed using the aforementioned signal model, with simulation analysis focused on time synchronization. In the detection portion, a single transmitter detection framework utilizing waveform diversity to accommodate coarse synchronization is proposed and studied. In the estimation portion, the Cramer-Rao Lower Bound (CRLB) is derived for a single-shot multistatic estimation scenario with Gaussian timing errors modeled in a novel way. An iterative Gauss-Newton position estimator is developed and compared to this bound, assuming statistics of the synchronization error are known.