Interacting Multiple Model algorithm with Quasi-Monte Carlo Kalman Filter

The Interacting Multiple Model (IMM) Algorithm is widely used in multi-model systems over the recent years. It often needs to handle nonlinearity of each mode in the framework of IMM. Compared with particle filter based on sequential Monte Carlo method, the Quasi-Monte Carlo (QMC) method has a superior performance in dealing with nonlinearity. Based on the technique that the QMC method is introduced into the IMM framework to dealing with the nonlinearity in each mode, the IMM algorithm with Quasi-Monte Carlo Kalman Filter (QMC-KF) is proposed in this paper. Meanwhile, the sample number in each mode is decided by the value of the mode probability in order to pay more attention to the dominant mode. Simulation results show that the performance of the proposed IMMQMC-KF is prior to that of the IMMUKF, IMMPF, IMMEPF and IMMUPF. Furthermore, the computing load of the IMMQMC-KF is lower than that of the IMMPF, IMMEPF and IMMUPF.