Modified Cerebellar Model Articulation Controller (MCMAC) As An Amplitude Spectral Estimator for Speech Enhancement - DSP for In-Vehicle and Mobile Systems - page 123

Chapter 8

Modified Cerebellar Model Articulation Controller
(MCMAC) As An Amplitude Spectral Estimator for
Speech Enhancement

Abdul Wahab 1 , Tan Eng Chong 1 , and Hüseyin Abut 2

1 School of Computer Engineering, Nanyang Technological University Nanyang Avenue,
Singapore; 2 Electrical and Computer Engineering Department, San Diego State University,
San Diego, CA 92182, USA. Email: [e-mail]

Abstract:

In this chapter, we present a modified cerebellar model articulation controller
(MCMAC) to be used together with the amplitude spectral estimator (ASE) for
enhancing noisy speech. The MCMAC training overcomes the limitations of
the CMAC technique we have employed noise/echo cancellation in a vehicular
environment. While the CMAC in the training mode has trained only the
trajectory it has visited by controlling the reference input, the modified
MCMAC-ASE system architecture proposed in this work includes multiple
MCMAC memory trainable for different noise sources.

Keywords:

Cerebellar model articulation controller (CMAC), speech enhancement, echo
cancellation, in-car noise, amplitude spectral estimation, Wiener filtering,
Kohonen's self-organizing neural network (SFON), Grossberg learning rule,
neighborhood function, and MOS.

INTRODUCTION AND CEREBELLAR MODEL
ARITICULATION CONTROLLER

1.

In this chapter, we present first a cerebellar model articulation controller
(CMAC) block diagram as shown in Figure 8-1, which can be described as
an associative memory that can be trained to implement non-linear functional
mappings.