Learned Multitarget Tracking by Feature Recognition

Ninth Midwest Artificial intelligence and Cognitive Science Conference

Volume

Ninth Midwest Artificial intelligence and Cognitive Science Conference

Contents

A tactical pilot typically experiences difficulty in maintaining an accurate lock on multiple-interacting targets in the presence of clutter. If the pilot is trained to focus on object features and predict the relative motions of targets, the anticipated target tracking methodology allows the pilot to identify multiple targets simultaneously. By assessing features, the paper seeks to demonstrate how feature recognition enhances multisensor integration of multitarget scenarios.