Time- and order-recursive estimation of higher order moments in a linear array

To achieve real-time performance in signal processing applications that require the adaptive estimation of higher order statistics, it is necessary to (a) design new time-recursive algorithms and (b) introduce parallel processing and pipelining. Time-recursive estimation methods of the computation of the higher order moments from the incoming sampled data are presented. The fixed size sliding window estimation algorithm is mapped onto an optimal linear array. The array provides estimates of all the moments up to the fourth order in time- and order-recursive fashion, achieves minimum latency and is suitable for VLSI implementation. A systematic synthesis methodology facilitates the construction of the optimal locally recursive algorithm and the design of the architecture