Abstract
This paper presents an innovative pipelined architecture for the implementation of adaptive filter based on distributed arithmetic (DA) with low-area, low-power. The high throughput rate of the proposed design is achieved by updating the lookup table simultaneously and parallel implementation of filtering and weight-update operations. In order to reduce the sampling period and area complexity, the proposed method uses conditional signed carry-save accumulation for the purpose of DA-based inner-product computation in the place of conventional adder-based shift accumulation. In order to reduce the power consumption, the proposed design uses a faster bit clock for carry-save accumulation but it uses a much slower clock for the remaining operations. The proposed design involves the same number of multiplexers but a smaller LUT and the number of adders used reduces to half when compared to the existing DA-based design.