S.C. Stubberud and K.A. Kramer (USA)
Neural Networks, Kalman Filter, Target Tracking, Target Intercept, Control
One approach to reduce the amount of energy to control an intercept vehicle to its target is to predict the location of the target at some point in the flight and to send the intercept vehicle to that point. This long term prediction or prognostication is predicated on the knowledge of the flight path of the target. If the target flight path is unknown a priori, a target tracking system must provide the trajectory of the information. One approach to target tracking is to use the neural extended Kalman filter. 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 improved model of the neural extended Kalman filter can be used to provide both a predictive and prognosticative estimate of the target location. This estimate can be used to put the interceptor in the general location of the target which can then be homed in on based on updated measurements from that point. In this paper, the control of an interceptor using a prognostic and predictive estimate of target location using the neural extended Kalman filter for control of interceptor to engage a maneuvering ballistic target is demonstrated.
Important Links:
Go Back
