THE GENERALIZED KALMAN FILTER APPROACH TO ADAPTIVE UNDERWATER TARGET TRACKING

A novel approach for solving the underwater target tracking problem is proposed. We review a model relating the target states to time-delay measurements as the basis for a linearized measurement model. Conventional Kalman filters (CKF's) have been used in adaptive target tracking. CKF's have difficulties in maintaining lateral continuity (stability), as a result of each record being analyzed separately (single channel operation). The proposed generalized Kalman filter (GKF) is based on the formulation developed by Lagarias and Aminzadeh establishing a trade-off between the cost associated with estimation error and the cost related to the lateral discontinuity of the estimates. By assigning proper weights for accuracy and stability in the objective function the desired balance between accuracy of estimates and lateral continuity is achieved. We offer computational results to illustrate the performance of the technique. Conclusions as to the effects of the accuracy and stability weights are drawn.