A PARALLEL BINARY STRUCTURED LMS ALGORITHM FOR TRANSVERSAL ADAPTIVE FILTERS

The adaptation process in digital filters requires extensive calculati on. This computation makes adaptation a slow and time consuming proces s. Simple, but versatile, parallel algorithms for adaptive filters, su itable for VLSI implementation, are in demand. In this paper a regular and modular parallel algorithm for an adaptive filter is presented. T his parallel structure is based on the Gradient Vector Estimation Algo rithm, which minimizes the Mean Square Error. In the parallel method, the tap weights of the adaptive filter are updated every s steps, wher eas in the recursive algorithms, the tap weights are updated at each s tep. For s step update, bit strings of length s are used to derive the terms with which the tap weights of the adaptive filter are calculate d. The algorithm presented computes the tap weights at time n + s as a function of the tap weights at time n, the inputs from time n + 1 to n + s - 1, and the desired output from time n + 1 to n + s - 1. The al gorithm also can be mapped to a VLSI architecture that is both regular and modular and allows either expansion of the order of the filter or the degree of parallelism obtainable. A comparison between the perfor mance of the sequential LMS algorithm, Fast Exact LMS algorithm, and t he parallel binary structured LMS algorithm is presented.