Spatial And Temporal Alignment Of Multi-platform Multi-sensor Information Fusion Study

The temporal alignment and special alignment of Multi-sensor, Multiple-platform and Multiple-source data fusion is becoming an important research subject. It has important significance in theory and practice to studying the method and theory about sensor registration of Multi-sensor data fusion system for improve integrate performance. The temporal alignment algorithm and special alignment algorithm are introduced, special to special alignment algorithm in this paper. First, full-blown estimation theory and existing sensor registration theory are introduced, and gives an excellent overview of existing special alignment algorithm, studies the extended Kalman filter alignment algorithm, and studies the least squares technique and the maximum likelihood method. The existing special alignment algorithm is based on the stereographic projection implemented on a two-dimensional regional plane which distorts the data and the model in the regional plane is unable to represent the actual sensor model properly. Second, for removing the influence, we develop the least squares method based on the stereographic projection implemented on a three-dimensional plane. It is described by the expressions:(4.35挆4.39). This method resolves some problem in Multi-sensor alignment on the same platform when the attitude errors are relatively small. For Multiple-platform, we develop the Multi-sensor alignment in an Earth-centered Earth-fixed coordinate system. The algorithm registries Multi-sensor using s geodetic transformation and is described by the expressions:(5.l1?.20). It generalized the least squares method applied to sensor alignment from two-dimension plane to three-dimension plane. And an exact maximum likelihood registration algorithm is presented based on three-dimensional plane. This method contains two sets of partially separable variables (the actual target positions and the sensor registration errors) and provides a two-step recursive optimization algorithm to ensure fast convergence. Described by expressions:(5.26)?5.39). It resolves the problems on Multiple-platform. Finally, simulated data is used to evaluate the performance of the developedalgorithms and proposed algorithIn. The simulated results show that the algori thmis effective and rel iable.