Multi-sensor Data Fusion And Simulation For Time-delay Systems

Multi-sensor information fusion technology can get more information to form the more reliable and complete the integration information than a single source more by similar structure or heterogeneous multi-sensor data. It breaks through the limitations of the expression of a single sensor information, avoids the blind spot of a single sensor, improves the quality of the multi-source information processing results, is favorable the judgments and decisions. Information fusion technology includes a wide range, which includes the information fusion state estimation is an important area. There are many methods for state estimation methods, including the classical Kalman filter, Wiener filter. Then the Kalman filter estimation theory has become the main tool in recent decades.Although Kalman filtering applications has been very extensive, the standard Kalman filter can only handle the non-delay systems, delay systems does not apply. However, with the development of network technology, time-delay systems are more and more attention widely. The research is very little on this problem of linear delay systems information fusion estimates, and the issues of delay systems control. Especially in the multi-delays multi-system, uncertainties system, there are many issues need to been studied furtherly.First, we study the problem of linear discrete-time systems with multi-models information fusion filtering with the known noises. Standard Kalman filter is no longer applicable, based on re-organization innovation and the scalar weighting sense, the optimal information fusion Kalman filter is presented. The observational data with delay will been re-organized as observational data without delay from different observing systems, the time-delay systems is transformed into a non-delay system to solve this problem, and the new formula is proposed by Kalman filtering,Next, we study the problem for the linear discrete time-delay systems information fusion with the unknown statistical noise. By introducing the Krein space, the unknown noise statistics discrete time-delay system H∞estimation problem is transformed into for the H2 estimation problem in a Krein space. Based on re-organization of innovation and the projective theory, the paper gives a simple and effective solution for sub-system estimator. This method avoids the conservative of linear matrix inequalities and shortcomings of only being given sufficient condition. The method weighted by matrix sense improves the estimation accuracy.Another interest, based observation re-organization theory, we study the linear estimation problem for multipath time-delayed discrete systems. A uniform observation equation is given by observer reorganization and Kalman filter. The information fusion filter is given by integrate the subsystems estimator weighted by scalars or matrix, and fusion algorithm is given. Through the Matlab simulation, the simulation image is presented. The MATLAB simulation results show that the proposed algorithm is feasible and effective.