A 2-D intercept problem using the neural extended Kalman filter for tracking and linear predictions

The neural extended Kalman filter is a technique that learns unmodeled dynamics while performing state estimation in the feedback loop of a control system. This coupled system performs the standard estimation of the states of the plant while estimating a function to approximate the difference between the given state-coupling function model and the behavior of the true plant dynamics. At each sample step, this new model is added to the existing model to improve the state estimate. The neural extended Kalman filter is applied to a two-dimensional intercept problem and the results are compared to those obtained from a standard tracking system. Since the neural extended Kalman filter better models the dynamic system, the time prediction of the state estimate which is needed for intercept control is superior to that provided by a standard tracking model. In this paper, predictions of one, two, and five time steps are investigated for use as the reference signal.