Bruce R. Musicus

Fast MLM power spectrum estimation from uniformly spaced correlations

We derive a fast algorithm for calculating the Capon maximum likelihood method (MLM) power spectrum estimate when given uniformly spaced samples of the correlation function. This algorithm computes a weighted correlation of the predictor coefficients found by running Levinson recursion. The Fourier transform of the result gives the MLM spectrum. This approach also suggests an interesting new comparison between MLM and MEM.

A local relaxation method for solving PDEs on mesh-connected arrays

A local relaxation method for solving linear elliptic PDEs with

Fast fault-tolerant digital convolution using a polynomial residue number system

O(N)

Maximum likelihood parameter estimation of noisy data

processors and

Iterative algorithms for optimal signal reconstruction and parameter identification given noisy and incomplete data

O(\sqrt N )

Estimate-maximize algorithms for multichannel time delay and signal estimation

computation time is proposed. We first examine the implementation of traditional relaxation algorithms for solving elliptic PDEs on mesh-connected processor arrays, which require

Generalised multiprocessor scheduling using optimal control

O(N)

The optimal control approach to generalized multiprocessor scheduling

processors and

Hierarchical compilation of macro dataflow graphs for multiprocessors with local memory

O(N)

Maximum entropy pole-zero estimation

computation time. The disadvantage of these implementations is that the determination of the acceleration factors requires some global communication at each iteration. The high communication cost increases the computation time per iteration significantly. Therefore, a local relaxation scheme is proposed to achieve the acceleration effect with very little global communication in the loading stage. We use a Fourier analysis approach to analyze the local relaxation method and also show its convergence. The convergence rate of the local relaxation method is studied by computer simulation.