Combining H∞ filter and cost-reference particle filter for conditionally linear dynamic systems in unknown non-Gaussian noises

This paper develops a state estimation method for conditionally linear dynamic systems in unknown non-Gaussian noises, which is a combination of the H∞H∞ filter and the cost-reference particle filter (PF). The proposed method has similar algorithmic structure as the mixture Kalman filter (MKF), which is a combination of the Kalman filter and the standard PF for conditionally linear dynamic Gaussian systems. The MKF requires the knowledge of the noise distributions and the noises are Gaussian or conditional Gaussian with known parameters in the model, which might not hold in many practical applications, while the proposed method does not require the knowledge of the noise distributions and the noises can be non-Gaussian, so it is more flexible and has less limitation in applications. Two applications of the proposed method in telecommunications, as well as the computer simulation results, are provided to illustrate the performance of the proposed method.

► A state estimation method is developed for conditionally linear dynamic systems.
► The method does not require the knowledge of the noise distributions.
► Two applications of the method in telecommunications are provided.