Magnetic dipole signal detection and localization using subspace method

The likelihood ratio method has been reported for signal processing involving the detection problem (magnetic anomaly detection) of magnetic dipole target (such as a sunken ship) by a magnetic sensor on board an aircraft. However, the observed waveform of the magnetic dipole target is not uniquely determined and varies with the magnetic dipole moment direction of the target. Therefore, many reference functions are needed for detection. It is a challenge to carry out stable detection with a small computational overhead. In this paper, based on the fact that the mathematical model of the magnetic dipole target is a linear combination of functions of three types electromagnetically, a method of detection based on the subspace method is proposed. Use of the wavelet transform as a preliminary process is discussed for improvement of the detection capability in the background noise. Further, in order to estimate the signal source of the magnetic dipole target, the subspace method is extended and a signal source estimation method is proposed that does not require complex nonlinear search. Finally, by means of a computer experiment, it is shown that the detection based on the subspace method is effective for detection of a magnetic dipole target with a low S/N ratio and that robust signal detection of a magnetic dipole target is possible by the proposed signal detection and localization method. © 2002 Scripta Technica, Electron Comm Jpn Pt 3, 85(5): 23–34, 2002