Detection and estimation in non-stationary environments
T.M. Toolan, D.W. Tufts · 2004
In this paper we describe a matrix based method for estimating both a signal subspace dimension as well as signal parameters when one has an array of sensors, multiple exponential signals, and significant signal changes after a small number of snapshots. This method combines the technique of creating Hankel or Toeplitz matrices from single-channel data with methods for sensor-array processing using multiple array snapshots.