Performance Analysis of Shift-and-Add Architecture using RBA Adder for Low Latency Biomedical Signal Processing Applications

In biomedical signal processing, the most prevalent operations are FIR/IIR, Shift-and-Add, DWT, DCT, differentiation and moving average. Human bio signals such as muscle or brain activity are used to support fast, and natural human-robot interaction via EMG/EEG, respectively, with low latency. This paper proposes a multiplexer-based shift-and-add architecture for biomedical signal processing applications with low latency. This architecture replaces the conventional adders and subtractors with redundant arithmetic concepts namely Redundant Binary Adder (RBA), consists of Redundant Arithmetic Plus-Plus-Minus (RA-PPM) adder. Minus-Minus-Plus (MMP) Subtractor is also designed. Shift-and-Add structure based on the principle of redundant arithmetic attains a low latency, with minimized hardware. The simulation is performed on the proposed architecture to compare conventional structure with power, area, and latency as the performance metrics. Compared to conventional structures, RA-PPM adder-based shift-and-add structure attains low latency with a minimized area on an IC and these kinds of architectures are more suitable for easy hardware realizations. Hence it can be used in diverse areas such as DCT, signal and image processing, and biomedical applications. The numbers of cells are reduced up to 36.6% for the proposed RA-PPM adder compared to CSA and the total power is also reduced up to 50%.