Improved Hilbert-Huang Transform and its application in acoustic vector signal processing

Hilbert-Huang Transform (HHT) is a kind of signal analysis method suitable for nonstationary signal, whose core algorithm is empirical mode decomposition (EMD). To improve the results and decomposition speed of EMD, this paper proposes a new method of obtaining mean curve for empirical mode decomposition(EMD), in which mean curve is obtained based on piecewise high order spline fitting. The method obtains more accurate mean points by using piecewise high order spline curve, but overcomes the disadvantage of too much computation in high order spline fitting. The mean curve is achieved by directly fitting mean points sequence, and the effect of overshoot and undershoot is reduced on the premise of guaranteeing the smoothness of the mean curve. The improved algorithm is used in processing acoustic vector signal, and data obtained from a certain sea experiment in Dalian is analyzed by improved HHT. The waterfall plot accurately shows the azimuth of target ship, which validates of the algorithm.