Abstract

The rate equation for the antenna weights and the covariance matrix of adaptive phased array radar are compared to the neural slab rate equation and the synaptic connection strength matrix of an artificial neural network model. It is shown that the mathematics are nearly identical except for a few key differences. These differences are traced to the basic goals of each system: The purpose of the radar systems to null out a given input distribution while that of the neural model is to identify and discriminate various input distributions. The second part of the paper considers some of the optical techniques appropriate for each system and discusses how they might be applied to the other. Other optical approaches are developed and shown to be compatible with adaptive phased array and/or artificial neural networks.

© 1985 Optical Society of America