High speed and Power efficient Digital VLSI Architecture of Artificial
Neural Network for reliable in-situ Water Quality Application
Abstract
This paper presents a low-power portable Digital VLSI architecture using
an Artificial Neural Network (ANN) for Water Quality Monitoring. The
study uses Posit number representation to implement ANN on both FPGA and
ASIC platform. The proposed ANN Posit architecture has 50% improvement
over IEEE 754 in terms of Power and Silicon Area, and 13% improvement
in speed it achieves the comparable accuracy. The same design using FPGA
consumes 6 orders higher power than VLSI architecture.