Mixture weighted gating and its application to target tracking

This paper presents a new approach to gating in target tracking. Here, the gate is defined as the region in which the tracked target is expected to exist. The observation vector in the gate is used for tracking. In tracking in a two-dimensional plane, various shapes can be conceived as a gate, including a rectangle, circle, and ellipse. Elliptical gate is optimal for linear Gaussian modeled systems and suboptimal for non-Gaussian systems. This paper introduces mixture weighted gating by generalizing Kalman filter to non-Gaussian linear systems. This non-elliptical gate is applied in target tracking and its advantage respect to elliptical gate is shown