Simple Robust Fixed Lag Smoothing
Abstract
This paper introduces a class of robust lag-k smoothers based on simple low order Markov models for the Gaussian trend-like component of signal plus non-Gaussian noise models. The kth order Markov models are of the Kth difference form Delta 'k' x sub t = epsilon sub t where Delta x sub t = x subt - x sub (t-1) and epsilon sub t is a zero-mean white Gaussian noise process with variance sigma-sq. sub epsilon. The nominal additive noise is a zero-mean white Gaussian noise sequence with variance sigma-sq sub 0, while the actual additive noise is non-Gaussian with an outliers generating distribution, e.g., (1-gamma) N(0, sigma-sq sub 0) + gamma H. This setup is particularly appropriate for radar glint noise. Implementation of the smoothers requires estimation of the two parameters sigma-sq sub epsilon and sigma-sub 0. This is accomplished using a robustified maximum likelihood approach. Application to both artificial data sets and to glint noise data reveals that the approach is quite effective. We briefly discuss the choices of lag k for the smoothers and also briefly study the sensitivity of our approach to model mismatch.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 02, 1988
- Accession Number
- ADA207204
Entities
People
- N. D. Le
- R. D. Martin
Organizations
- University of Washington