Algorithms Research On Multisensor Data Fusion

Multisensor data fusion is a process of multi-level and multi-aspects, and during this process, multi-source data are examined, combined, correlated, estimated and assembled to get an accurate estimation of state and identity, and a complete, timely situation and threat assessment. In this paper, algorithms of multisensor data estimation fusion and threat assessment were lucubrated in detail.Theories of data estimation fusion are detailedly introduced in this paper firstly.And diverse data fusion structures in the multi-sensor system are summarized.The Kalman filter-based model is constructed for the data fusion estimation, and the Kalman filter algorithms for different data fusion structures, including centralized, sequential, parallel, joint and so on, are enumerated and simulated. The final simulation results could lead to the proposed conclusion that the estimation accuracy can be considerably improved by the multi-sensor data fusion. Then the Kalman filter under colored noise or information related situation are investigated,and the corresponding processing methods are proposed.And then the unified linear data model is constructed. In addition, the optimal rules for linear fusion in diverse scenario are introduced, especially for the cases with both linear and non-linear data. Finally threat assessment algorithm is investigated. The combination of analytic hierarchy process and fuzzy theory is adopted for the threat assessment, and in consequent the advantages existing in both methods can be fully utilized. The combination algorithm is validated by its application in the air-defense system, and a more scientific and rational method for the operational command decision-making is provided...