Development of a ground truth simulator and application of a generalized multiple-model adaptive estimation approach to tune a state estimation filter

In this thesis, a maritime scenario simulator is developed and a data processing/filtering algorithm is applied to estimate the ground truth of the simulated scenario from noisy measurements and system model for the Hierarchical High Level Information Fusion Technologies (H2LIFT) project. H2LIFT is an adaptable information fusion frame- work which takes as input Levels 0/1 (local) data and performs fusion at Levels two and three (distributed, and network centric) hierarchically, in different stages, to provide real- time situational/impact assessment efficiently while avoiding the overload of information to the human decision maker. First, a simulator is developed that imitates a naval threat from an incoming vessel (such as a cargo ship containing a weapon of mass destruction), included in a group of non-threatening vessels. The developed simulations are used as evaluation metrics and performance platforms providing an operational utility assessment tool for the H2LIFT algortithm. Next, a Generalized Multiple-Model Adaptive Estima- tion (GMMAE) technique is used to estimate the unknown parameters involved with a Probability Data Association Filter (PDAF) which includes a Kalman Filter (KF). The properly tuned state estimator is used to provide estimates of the ground truth data from the noisy sensor measurements and incomplete system model. These estimates are used as inputs to the H2LIFT algorithm and can be tested against the known ground truth to gauge filter performance. A demonstration of the process is provided in the simulation section