Having a better motion model in the state estimator is one way to improve target tracking performance. Since the motion model of the target is not known a priori, either robust modeling techniques or adaptive modeling techniques are required. 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 has also been investigated as a target tracking estimation routine. Implementation issues for this adaptive modeling technique, including neural network training parameters, were investigated and an analysis was made of the quality of performance that the technique can have for tracking maneuvering targets.
A --Classification is on imponontpart of the Level I Fusion problem. It provides necessary information to the user to make decisions about the object in question. In the military problem, this decision can be the difference between prosecution of a target and declaration as a noncombatant. A number of approaches use different sources of information oddress classification. Many of these techniques are probabilistic. Some are linguistic interpretations by expert analysis. These classifiers can provide different views of the same object The key is to combine these dispnrate pieces of information, each with different levels of qualify andlor confidence. We propose B technique to combine these various classifiers using the concept of evidence accrual. Informotion can offirm U class' hypothesis, refute it, or have no bearing upon it. Also. each source of evidence has II degree of qunlify or uncertainry about it. Finally, the sources may be represented as numeric or nonnumeric informotion. To oddress these issues, we utilized a set of decoupled fuzqv Kalmanjilters. This bank offilters, similar in nature tofirst-order obseservers, estimates the degree of evidence that each known class achieves. Thepaper outlines the development of our approach, its implementation to emulate (I Bayesian clmxifier, and a sef of examples.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.