Optimal Filtering of a Gaussian Signal in the Presence of Lévy Noise
Hyungsok Ahn, Raisa Epstein Feldman · SIAM Journal on Applied Mathematics · 1999
Many engineering applications require extracting a signal from observations corrupted by additive noise, possibly heavy-tailed. We assume that the observation noise is a Lévy process, while the signal is Gaussian, and derives a nonlinear recursive filter that minimizes the L 2 error. A suboptimal filter is proposed for numerical purposes, and simulations show that it outperforms the existing linear filter.