Search

A simple, necessary and sufficient test is developed for testing whether a single connection in a tree-type NAND network is redundant. A procedure is presented for testing every connection in the network. The computational complexity of the procedure is m(i squared) where m = the number of gates and i = the average number of inputs per gate in the network. The redundancy test is generalized for multi-output tree-type NAND networks and such networks realizing partially specified functions. A dual test is developed for tree-type NOR networks, but NOR networks for partially specified functions are treated by a simpler test than the dual test. The test may be applied to AND-OR networks as well by converting them, at least conceptually, to an equivalent NAND form while preserving redundancy. (Modified author abstract).

The goal of this research project is to develop biocybernetic concepts and techniques useful for the analysis and development of skills needed for the enhancement of concrete images of the 'eidetic' type. The authors treat scanpaths of the eye during observation of visual patterns as indicators of the brain's strategy for the intake of visual information. The authors attempt to determine the features that differentiate visual scanpaths associated with superior imagery from scanpaths associated with inferior imagery, and simultaneously, to differentiate the electroencephalographic features correlated with superior imagery from those correlated with inferior imagery. Incorporating this information, a closely-coupled man-machine system to generate image enhancement and to train the individual to exert greater voluntary control over his own imagery will be designed and implemented.

Given any set of signals for use with a one-way communication channel, a feedback channel can be used to decrease the average power used by the transmitter while keeping the error probability the same, or equivalently, decreasing the error probability for the same average transmitter power. The case of orthogonal transmitter signals is analyzed in detail. The feedback channel is used only to tell the transmitter at certain specified times (prior to a final decision) which signal the receiver believes is being sent. The feedback information is used by the transmitter to modify its transmission according to a rule known also at the receiver. The feedback communication scheme is analyzed for both a noiseless and a noisy feedback link. In the case of noiseless feedback if only an average transmitter power constraint is imposed and orthogonal signals are used, it is possible to have the error probability go down in an N-fold exponential manner with increasing signal duration, where N-1 corresponds to the number of feedback iterations per message. This is in contrast to only a double exponential form for the error probability with some previously proposed feedback schemes. If a peak transmitter power constraint is also imposed, neither the N-fold exponential nor the double exponential behavior can be obtained. The peak power constraint limits the error probability expression to only a single-exponential form. (Author).