Hardware Neural Network Implementation of Tracking System
Neural Networks for Signal Processing - Proceedings of the IEEE Workshop
A neural network (NN) filter/target-tracking system that accepts and inputs signal data to a noise/target classifier is reported. In its initial design, spectral estimation techniques were used to distinguish noise from real targets. NN was used to calculate the coefficients of an auto regressive linear predictive filter. The current evolution of the design uses Lagrange Multiplier methods to incorporate known character of the noise vs. signal. A (linear) Hopfield NN is used to perform the constrained optimization to solve for the filter coefficients. This algorithm has been demonstrated on real stochastic data.
Locate the Document
Lendaris, G. G., Pap, R. M., Saeks, R. E., Thomas, C. R., & Akita, R. M. (1994, September). Hardware neural network implementation of tracking system. In Proceedings of IEEE Workshop on Neural Networks for Signal Processing (pp. 451-460). IEEE.