Multiple Model Adaptive Tracking Algorithm for a High Energy Laser Weapon System

摘要

This thesis considers replacing a standard correlation tracker with a hybrid Kalman filter/enhanced correlation tracker in a high energy laser weapon system. Dynamic airborne targets are tracked by a Bayesian multiple model adaptive filtering (MMAF) algorithm, which processes the outputs of a matrix-type array of infrared sensing detectors. Emphasis is placed on extending the adaptive potential of the tracking algorithm. This is accomplished by processing measurements from various field of view (FOV) sizes and shapes, and by incorporating direction-dependent target dynamics in some of the elemental Kalman filters within the multiple model structure. A sensor to target range tuning algorithm is derived which can be used for on line adaptive filter tuning should the tracker be provided range information, (even at low sample rates and/or precision), possibly via laser ranging. Also, the problem of initial target acquisition is explored through an algorithm which acquires the target in the center of the FOV despite initial sensor pointing errors. Two different target dynamics models are considered for the elemental Kalman filters: a linear, Gauss-Markov acceleration model, and a nonlinear, constant turn-rate model.