Multistatic moving target detection in unknown coloured Gaussian interference

Abstract

One of the main interferers for a Doppler radar has always been the radar׳s own signal being reflected off the surroundings. This creates the problem of searching for a target in a coloured noise and interference environment. Traditional space–time adaptive processing (STAP) deals with the problem by using target-free training data to study this environment and build its characteristic covariance matrix. The maximum likelihood estimation detector (MLED) and its generalised counterpart (GMLED) are two reduced-rank STAP algorithms that eliminate the need for training data when mapping the statistics of the background interference. In this work the MLED and GMLED solutions to a multistatic scenario are derived. A hybrid multiple-input multiple-output (MIMO) system where each receiver is a coherent STAP radar has been employed. The focus of the work is the spatial diversity provided by the wide separation of the individual transmitter and receiver platforms. It is proven that this configuration does not affect the constant false alarm rate (CFAR) property of the bistatic radar case. A Gaussian approximation to the statistics of the multistatic algorithms is derived in order to provide a more in-depth analysis. The viability of the theoretical models and their approximations are tested against a numerical simulation.