A set based approach to target tracking in clutter

In this paper a set based estimation approach to target tracking in clutter is considered. To ensure a guaranteed bound on target location is maintained at all times, a track split approach using all feasible measurements is adopted. This approach suffers from an exponential growth in the number of feasible target location bounds over time, even with modest clutter densities. An overbounding procedure is introduced to group sets of bounds in a similar locale into a single bound. Above a clutter density threshold, this provides a more computationally efficient option while ensuring a guaranteed bound is maintained.