Bearing estimation with a single moving Pressure-Vector sensor

Abstract

By combining a pressure sensor measurement with a three-dimensional vector sensor measurement on a moving platform, it is shown that by including the motion in the signal model, a significant performance improvement in the estimation of the polar angles of the incoming signal is achieved by the exploitation of the bearing information in the Doppler. The method considers the four outputs from a Pressure-Vector (PV) sensor as the measurement system in an Unscented Kalman Filter (UKF), and the source frequency and the two polar angles as elements of a three-dimensional Kalman state vector in a random-walk state model. It is shown, using simulated data, that an improvement in the signal-to-noise performance of the bearing estimate on the order of 10 dB is possible.