Date of Award

Spring 1982

Project Type


Program or Major

Engineering Signal Processing Area

Degree Name

Doctor of Philosophy


Problems with missing data, sampling irregularities and randomly sampled systems are the topics covered by this dissertation.

The spectral analysis of a series of periodically repeated sampling patterns is developed. Parameter estimation of autoregressive moving average models using partial observations and an algorithm to fill in the missing data are proved and demonstrated by simulation programs. Interpolation of missing data using bandlimiting assumptions and discrete Fourier transform techniques is developed. Representation and analysis of randomly sampled linear systems with independent and identically distributed sampling intervals are studied. The mean, and the mean-square behavior of a multiple-input multiple-output randomly sampled system are found. A definition of and results concerning the power spectral density gain are also given.

A complete FORTRAN simulation package is developed and implemented in a microcomputer environment demonstrating the new algorithms.