Sampling strategies of deterministic sampling nonlinear filters based on invariant cubature formula

In order to improve the filter precision, this paper studies on sampling strategies of the deterministic sampling nonlinear filters. Firstly, we propose a general sampling points selection method based on invariant cubature formula. Secondly, several sampling strategies with 3 order and 5 order accuracy by the proposed method are obtained and analyzed. Finally, simulations based on the bearings-only tracking model and target tracking model show that the proposed sampling strategies are valid. Simulation results show that when using different sampling strategies, the estimation precision of the filters are quite different, and select the appropriate sampling strategies can improve the filter precision effectively.