Recursive robust filtering for uncertain systems with delayed measurements

A recursive robust filtering method based on the orthogonal projection and an innovation analysis approach is proposed for stochastic uncertain systems with delayed measurements. First, the original system is transformed into a stochastic uncertain system with correlated measurement noises, and the innovations are recalculated. Then, by applying the mathematical induction, the newly obtained innovations are proofed to be uncorrelated with each other. Finally, the desired recursive robust filter is designed via an innovation analysis approach and the orthogonal projection theorem. Without resorting to state augmentation, the proposed filter treats the noise correlated across time directly. Therefore, the desired filter has less computational burden, but a high computational accuracy. Simulation results demonstrate the effectiveness of the proposed approach. © 2017, Editorial Board of Jilin University. All right reserved.