Title :
Simulation, MMSE estimation, and interpolation of sampled continuous-time signals using factor graphs
Author :
Bolliger, Lukas ; Loeliger, Hans-Andrea ; Vogel, Christian
Author_Institution :
Dept. of Inf. Technol. & Electr. Eng., ETH Zurich, Zurich, Switzerland
fDate :
Jan. 31 2010-Feb. 5 2010
Abstract :
The paper proposes a factor graph representation of finite-dimensional continuous-time linear systems /filters driven by white Gaussian noise. Applications of such factor graphs include (i) simulation of the output signal at arbitrary discrete instants and (ii) MAP /MMSE /LMMSE estimation of the input signal, or of the state, or of the output signal at arbitrary discrete instants from noisy observations of the output signal at arbitrary discrete instants. Moreover, it is suggested that the MMSE estimate of the input signal from the output signal may be viewed as a generalization of a bandlimited estimate that is suited for real filters.
Keywords :
Gaussian noise; continuous time systems; interpolation; least mean squares methods; linear systems; signal processing; white noise; MMSE estimation; arbitrary discrete instants; factor graph representation; finite dimensional continuous time linear systems; input signal; interpolation; output signal; sampled continuous time signals; white Gaussian noise; Autocorrelation; Gaussian noise; Information technology; Interpolation; Linear systems; Nonlinear filters; Sampling methods; Speech; State estimation; Vectors;
Conference_Titel :
Information Theory and Applications Workshop (ITA), 2010
Conference_Location :
San Diego, CA
Print_ISBN :
978-1-4244-7012-9
Electronic_ISBN :
978-1-4244-7014-3
DOI :
10.1109/ITA.2010.5454111