Towards ultra-low power bio-inspired processing
File(s)
Accepted version
Author(s)
Shepherd, LM
Constandinou, TG
Toumazou, C
Type
Chapter
Abstract
The natural world is analogue and yet the modern microelectronic world with which we interact represents real world data using discrete quantities manipulated by logic. In the human space, we are entering a new wave of body-worn biosensor technology for medical diagnostics and therapy. This new trend is beginning to see the processing interface move back to using continuous quantities, which are more or less in line with the biological processes. We label this computational paradigm “bio-inspired” because of the ability of silicon chip technology which enables the use of inherent device physics, allowing us to approach the computational efficiencies of biology. From a conceptual viewpoint, this has led to a number of more specific morphologies including neuromorphic and retinomorphic processing. These have led scientists to model biological systems such as the cochlea and retina and gain not only superior computational resource efficiency (to conventional hearing aid or camera technology), but also an increased understanding of biological and neurological processes.
Date Issued
2014
Citation
Body Sensor Networks, 2014, pp.273-299
ISBN
9781447163732
Publisher
Springer London
Start Page
273
End Page
299
Journal / Book Title
Body Sensor Networks
Copyright Statement
© Springer-Verlag London 2014. The final publication is available at Springer via https://doi.org/10.1007/978-1-4471-6374-9_7
Identifier
http://link.springer.com/chapter/10.1007/978-1-4471-6374-9_7
Publication Status
Published
Article Number
7