Implementation of multiple-model multi-Bernoulli filter for nonlinear models

Two implementations of the multiple-model multi-Bernoulli (MM-MB) filter were proposed for nonlinear system models in the presence of clutter and missed detection. Based on the cubature Kalman filtering algorithm, the cubature Kalman Gaussian mixture (CK-GM) implementation of the MM-MB filter was proposed. Moreover, to improve the numerical stability of the CK-GM-MM-MB filtering algorithm, the square-root CK-GM implementation of the MM-MB filter was proposed. Finally, a numerical example was presented to verify the effectiveness of the proposed approaches.