High throughput adaptive block FIR filter using distributed arithmetic

High throughput is the crucial parameter of the adaptive filters. Here, we present an adaptive block finite impulse response (FIR) filter with high throughput using the concept of distributed arithmetic (DA). In the proposed work, block processing is applied to the adaptive filters that has helped in improving the throughput. In this paper, computation of partial products of filter output takes place in 1 clock cycle unlike clock cycles equivalent to coefficient size(B) required by previous structures. Here B parallel DA based look-up tables (LUT) are used in order to reduce the clock cycles. This parallel structure helps in increasing the throughput by 9.11 times from the existing structure. Input matrix required for computing the output of filter and weight increment vector have minimum registers. Memory reuse approach is exploited while designing registers for generation of required input samples.