Skip to Main Content
Radio Frequency (RF) source location methods for wide bandwidth sources are well known. Such methods typically assume that a wide bandwidth array is available in order to collect the signal data. A number of techniques have been derived from the coherent signal subspace method, but often suffer from limitations such as the requirement for preliminary source location estimates, iteration of the technique, computational expense or others. The common theme among many of these is to divide the collected data into a number of sub-bands, then process the data in each subband in such a way that they can be coherently combined and processed using signal subspace techniques. This paper introduces a method to process data from a bank of true time delay beamformers that are inherently frequency invariant. The spatial diversity of the beamformer bank alleviates the need for a preliminary estimate while simultaneously reducing the dimensionality of subsequent signal subspace processing resulting in computational efficiency since the data can be processed in beam-space rather than element-space. Other methods to use frequency invariance have been described but often remain computationally expensive. We briefly review previous methods, pointing out their strengths and limitations, introduce the present method and demonstrate results compared to several previous methods. We also discuss an analysis of performance in relation to the Cramer-Rao Lower Bound (CRLB).