Target Acceleration Estimation from Radar Position Data using Neural Network

Abstract

This work is a preliminary investigation on target manoeuvre estimation in real-time from the available measurements of noisy position data from tracking radar using an artificial neural network (ANN). Recently, simulation study of target manoeuvre estimation in real-time from the same position alone measurement using extended Kalman filter has been carried out in a simulated environment using measurements at 100 ms interval. The results reveal that the estimated acceleration consists of substantial error and lag, which is a stumbling block for guidance accuracy in real-time. So, the target acceleration has been estimated using the ANN with less error and lag than the same using Kalman estimator.