Research on Fusion Filtering Algorithm for Lost Observation System

Aiming at the difficulty of information fusion estimation for multi-sensor linear discrete stochastic systems with unknown loss observation rate. In this paper, a novel information fusion estimation method based on matrix weighted optimal fusion filtering algorithm with observation loss rate is proposed. Firstly, using a group of random variables which satisfies the Bernoulli distribution to describe the phenomenon of sensor observation loss. Secondly, the correlation function is applied to identify the unknown loss rate online under the condition of the sensor observation loss rate is unknown. Finally, the identified observation loss rate is introduced into the proposed matrix weighted optimal fusion filtering algorithm. Compared with traditional multi-sensor matrix weighted Kalman filtering algorithm, the matrix weighted optimal fusion algorithm with missing observers has taken into account the missing observation data, this method uses the statistical principles to compensate for these missing observations. Therefore, it has obvious advantages, the simulation examples verify the effectiveness of the proposed algorithm.