Generalised spatial smoothing using partial arrays for multiple target bearing estimation

This paper considers the problem of bearing estimation for a small number of radar targets which cannot be resolved in range or Doppler frequency. Bearing estimation for non-fluctuating targets involves a single "snapshot" resulting from a multi-channel optimum (matched) filtering process. The standard spatial smoothing technique may be applied to this single-snapshot model, but only for uniform linear antenna arrays. Here we introduce a special class of nonuniform geometry with embedded partial arrays and a corresponding generalised spatial smoothing (GSS) algorithm. The partial array characteristics determine the resulting bearing estimation accuracy. A two-stage bearing estimation procedure is proposed. The initialisation stage involves spatial averaging over all suitable partial arrays. The refinement stage uses a local maximum-likelihood search. Typical radar model simulations and Cramer-Rao bound calculations demonstrate the efficiency of this approach compared with standard spatial smoothing using a uniform linear array.