Sciweavers

TBCAS
2016

Ultra-Low Power Dynamic Knob in Adaptive Compressed Sensing Towards Biosignal Dynamics

7 years 11 months ago
Ultra-Low Power Dynamic Knob in Adaptive Compressed Sensing Towards Biosignal Dynamics
—Compressed sensing (CS) is an emerging sampling paradigm in data acquisition. Its integrated analog-to-information structure can perform simultaneous data sensing and compression with low-complexity hardware. To date, most of the existing CS implementations have a fixed architectural setup, which lacks flexibility and adaptivity for efficient dynamic data sensing. In this paper, we propose a dynamic knob (DK) design to effectively reconfigure the CS architecture by recognizing the biosignals. Specifically, the dynamic knob design is a template-based structure that comprises a supervised learning module and a look-up table module. We model the DK performance in a closed analytic form and optimize the design via a dynamic programming formulation. We present the design on a 130 nm process, with a 0.058 mm fingerprint and a 187.88 nJ/event energy-consumption. Furthermore, we benchmark the design performance using a publicly available dataset. Given the energy constraint in wireles...
Aosen Wang, Feng Lin, Zhanpeng Jin, Wenyao Xu
Added 10 Apr 2016
Updated 10 Apr 2016
Type Journal
Year 2016
Where TBCAS
Authors Aosen Wang, Feng Lin, Zhanpeng Jin, Wenyao Xu
Comments (0)