Considering uncertain system parameters in multitarget space surveillance tracking

Consider analysis is an estimation technique that emerged in the 1960s to account for errors in system parameters while simultaneously reducing system dimensionality, and accordingly real-time computational cost, and/or guarding against issues of observability surrounding the parameters. The multitarget joint estimation problem is one whose dynamical and observational systems contain such parameter errors, and these errors can drastically impact the performance of a suboptimal recursion, such as the probability hypothesis density (PHD) filter. A consider formulation of the Gaussian mixture PHD filter is proposed to treat such problems while accounting for errors in system parameters without neglecting or directly estimating them. The proposed algorithm is applied to an example that illustrates its value in space object tracking and orbit determination.