Towards Low-Power On-chip Auditory Processing
- PDF / 2,007,317 Bytes
- 11 Pages / 600 x 792 pts Page_size
- 104 Downloads / 163 Views
Towards Low-Power on-Chip Auditory Processing Sourabh Ravindran Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0250, USA Email: [email protected]
Paul Smith Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0250, USA Email: [email protected]
David Graham Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0250, USA Email: [email protected]
Varinthira Duangudom Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0250, USA Email: [email protected]
David V. Anderson Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0250, USA Email: [email protected]
Paul Hasler Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0250, USA Email: [email protected] Received 18 September 2003; Revised 16 August 2004 Machine perception is a difficult problem both from a practical or implementation point of view as well as from a theoretical or algorithmic point of view. Machine perception systems based on biological perception systems show great promise in many areas but they often have processing requirements and/or data flow requirements that are difficult to implement, especially in small or low-power systems. We propose a system design approach that makes it possible to implement complex functionality using cooperative analog-digital signal processing to lower power requirements dramatically over digital-only systems, as well as provide an architecture facilitating the development of biologically motivated perception systems. We show the architecture and application development approach. We also present several reference systems for speech recognition, noise suppression, and audio classification. Keywords and phrases: low power, noise suppression, classification, speech recognition, cooperative analog digital, feature extraction.
1.
INTRODUCTION
This paper describes our efforts toward making on-chip auditory perception systems. With these systems we hope to achieve human-like performance in various auditory tasks. Part of the motivation in this arena comes from the fact that humans outperform machines in most tasks of audio perception; therefore, one way to improve the current machine perception implementations is to mirror biological systems as best as we can in hopes of obtaining comparable results.
In keeping with the current mainstream ideas in audio signal processing, the characteristics of biology could be programmed into a digital processing system, and the results would likely be good. However, this is at the cost of high power consumption and longer time requirements due to the computational complexity. It is time we looked at different perspectives of performing these tasks both in terms of hardware innovation and algorithm development [1]. A major consideration while designing on-chip machine perception i
Data Loading...
