Low power residue number system using lookup table decomposition and finite state machine based post computation

In this paper, memory optimization and architectural level modifications are introduced for realizing the low power <span lang="EN-US">residue number system (RNS) with improved flexibility electroencephalograph (EEG) signal classification. The proposed RNS framework is intended to maximize reconfigurability of high-performance finite impulse response (FIR) filter design. By replacing existing power-hungry RAM-based reverse conversion model a highly decomposed lookup table (LUT) which can produce results without using any post accumulation process. block modified an appropriate functional unit accommodate FIR convolution results. approach established develop execute pre-calculated inverters various module sets. Therefore, LUT-decomposition multiplication-based post-accumulation technology provides architecture that allows different frequency configuration elements. Experimental shows superior performance decomposing LUT-based direct over other techniques adopted energy-efficient implementations. When compared conventional design FSM based FIR, logic elements (LEs) were reduced by 4.57%, component increased 31.79%, LUTs 42.85%, dissipation was 13.83%.</span>