A HOPFIELD NEURAL TRACKER FOR PHASED-ARRAY ANTENNA

A state estimator based on neural network is applied to phased array t racking. The state estimation is formulated as a dynamic optimization problem and solved using a Hopfield neural network. This neural tracke r has the flexibility for adaptively varying the target-track update r ate as a function of target maneuvering. The value of the update time is dependent on the magnitude of the residual error of the state estim ator. Simulation results show improvement of the new approach over the standard variable update time cu-p filter for phased array tracking.