A fast nonlinear filtering algorithm for tracking a target in clutter using the wavelet transform

Abstract

We present a fast nonlinear filtering algorithm that can track a single target in multiple clutter points. The proposed algorithm propagates the entire conditional probability density functions recursively, but in a computationally efficient manner using either the fast Fourier transform or the fast discrete wavelet-based convolution. Our algorithm does not need the explicit data association step which is in most multiple target tracking filters, and therefore appears to be more natural and robust.