Implementation of a sequential feature selection algorithm for visual evoked potential data
Kane, Cathryn M. Hartmann
Master of Science
There exists a widely held belief that the full potential of LSI technology will be realized by arrays of logic circuits. A novel machine architecture is proposed that consists of a uniform array of identical cells with the property that the array executes a high-level programming language directly. The central idea of this thesis is that it is possible to exploit the geometry of an iterative machine architecture to achieve some degree of efficiency when dealing with languages that are structure-oriented. To this end, synthesis techniques for such languages (machines) are developed. Some of the more important features of such machines are their inherent abilities to sustain concurrent processes, to obtain a high degree of reliability through dynamic fault analysis, and to maintain efficient information storage. A detailed development and simulation of a language-machine in this architectural style is presented.