Abstract :
Conventional tracking with a towed array is carried out by employing a Kalman filter using a straight-line constant-speed kinematic model for the target. The measurement is made up of at least two bearing estimations, usually obtained using a conventional beamformer. This paper describes a new approach to the problem. Instead of using a plane-wave model for the target signal, a near-field signal model is employed which implicitly includes the array motion and the curvature of the wavefront. The importance of including the array motion is that it contains information that is ignored by a conventional beamformer. The problem is treated as a joint estimation of the range, bearing and source frequency for a moving source. The near field geometry, coupled with the synthetic aperture effect, renders these three quantities observable without necessitating a maneuver. A set of examples is given, using a simulated narrowband signal, in which it is shown that the method provides excellent performance for range-to-aperture ratios much larger than required by the conventional wavefront-curvature method.
Keywords :
Kalman filters; array signal processing; direction-of-arrival estimation; sonar tracking; synthetic aperture sonar; target tracking; Kalman filter; array motion; bearing estimation; near field geometry; near field signal model; range estimation; source frequency estimation; straight line constant speed kinematic model; synthetic aperture effect; towed array tracking; wavefront curvature; Delay effects; Direction of arrival estimation; Equations; Frequency estimation; Geometry; Kinematics; Narrowband; Target tracking; Tin; USA Councils; Model-Based Processing; Passive Synthetic Aperture; Towed Array; Tracking;