Mumbai University > Electronics and Telecommunication Engineering > Sem 6 > Data Communication

Marks: 5 Marks

Year: Dec 2016

Optimal filtering in Signal Processing

Wiener Filter: Nobert Wiener (MIT) 1940s:

Model Y = S + W, S is signal W is noise.

Widely used in LMMSE detection.

Kalman Filter: (1960s) Model S and N in time domain (state space models). The Kalman filter is probably the single most used algorithm in signal processing.

Hidden Markov Filter: Developed by statisticians (L. Baum, T. Petrie) in 1960s Significant application in Electrical Engg in 1990s in speech recognition, channel equalization, tracking, etc

Sequential Markov Chain Monte Carlo Methods: Particle filters – randomized (simulation based) algorithms – applications in target tracking – late 1990s

Stochastic Filtering theory studies optimal filtering. Also called recursive Bayesian estimation.