Neural network for direction or frequency estimating and tracking

Summary form only given, as follows. The authors discuss a neural estimator and its implementations of the parameters of superimposed sinusoidal signals in noise. The formation of the proposed estimator is based on an annealed linear programming neural network. When compared with traditional methods, the neural network has the following advantages: (1) it can obtain a real-time solution by setting up an associated analog circuit: (2) it is more suitable for adaptive processing needed to track nonstationary signal parameters; (3) neither data sampling (snapshot) nor an analog-to-digital converter is required: and (4) the neural network can yield better convergence to the global minima of the parameter estimation problems. Simulation results were obtained to illustrate the excellent performance of the neural networks for estimating and tracking the directions of arrival of plane waves