Optimal filtering and smoothing for speech recognition using a stochastic target model

Author

Ramsay, G. ; Deng, L.

Author_Institution

Dept. of Electr. & Comput. Eng., Waterloo Univ., Ont., Canada

Volume

2

fYear

1996

fDate

3-6 Oct 1996

Firstpage

1113

Abstract

Presents a stochastic target model of speech production, where articulator motion in the vocal tract is represented by the state of a Markov-modulated linear dynamical system, driven by a piecewise-deterministic control trajectory and observed through a non-linear function representing the articulatory-acoustic mapping. Optimal filtering and smoothing algorithms for estimating the hidden states of the model from acoustic measurements are derived using a measure-change technique and require the solution of recursive integral equations. A sub-optimal approximation is developed and illustrated using examples taken from real speech

Keywords

acoustic variables measurement; hidden Markov models; integral equations; optimal control; optimisation; smoothing methods; speech recognition; state estimation; stochastic systems; suboptimal control; Markov-modulated linear dynamical system; acoustic measurements; articulator motion; articulatory-acoustic mapping; hidden state estimation; measure-change technique; nonlinear function; optimal filtering; piecewise-deterministic control trajectory; recursive integral equations; smoothing algorithms; speech production; speech recognition; stochastic target model; sub-optimal approximation; vocal tract; Acoustic measurements; Filtering algorithms; Motion control; Nonlinear control systems; Production systems; Recursive estimation; Smoothing methods; Speech recognition; Stochastic systems; Trajectory;

fLanguage

English

Publisher

ieee

Conference_Titel

Spoken Language, 1996. ICSLP 96. Proceedings., Fourth International Conference on

Conference_Location

Philadelphia, PA

Print_ISBN

0-7803-3555-4

Type

conf

DOI

10.1109/ICSLP.1996.607801

Filename

607801