Space-time adaptive processing for airborne surveillance radar systems

Abstract

In this thesis, we investigate the computational complexity, realization requirements and detection performance of Space-Time Adaptive Processing (STAP) for radar surveillance radar system. We identify the signal components comprising the transmitted and received waveforms and organize it in a space-time data vector. SMI, RLS, EVD and QRD are presented as different implementation of the optimum, in the SINR sense, STAP. QRD avoids the computationally intensive inversion of covariance matrix performed by SMI, RLS and EVD while offering numerical stability to the calculation. Computationally reduced, suboptimum STAP approaches including factored and beamspace techniques are investigated. Drastic detection performance is lost in factored space-time configuration when the interference arrives at the same angular direction as target object, while, in factored time-space configuration when the interference arrives at the same Doppler frequency as the target object. Joint domain localization responses are symmetric to interference with the same angular direction and Doppler frequency as the target. In addition to alleviating the sample support requirement and non-homogeneity in the secondary range bins, deterministic and recursive approaches to STAP deals with present of discrete interference source in the target range bins. Target detection schemes including Bayes detectors, AMF and GLRT are utilized to evaluate the detection performance of each STAP algorithms.