GLR Algorithms for Detecting and Estimating Abrupt Maneuvers in ASMD Scenarios Using a Decomposition of the Maneuver Signature Matrix

摘要

We have presented two recursive GLR algorithms for detecting and estimating maneuver states and parameters in the engagement between an anti-ship cruise missile and a ship defense interceptor. A decomposition of the maneuver signature matrix is used to derive the algorithms. The computational and storage requirements are substantially less than those of other GLR algorithms. The decomposition divides the maneuver signature matrix into the product of two matrices. One matrix depends only on the current observation time while the other depends only on the jump time. The product of the latter matrix and the jump magnitude vector provides a jump error state vector which is constant. This constancy facilitates using the GLR approach. The other matrix of the decomposition represents the new maneuver signature matrix for the new constant jump error state vector. The nondependency of the maneuver signature matrix on the jump time avoids storing large matrices and computing large matrix products for each past time observation time.