Study Of Multi-scale Optimal Filtering Fusion Algorithms For Systems With Multiplicative Noises

Signal estimation theory and multi-sensor data fusion technology for discrete stochastic systems with multiplicative noises (SMN) are very important in many applications such as oil seismic exploration, underwater remote targets detection and speech signal processing. The optimal filtering fusion algorithms for systems with multiplicative noises under multi-sensor observation are mainly researched in this dissertation.In the past, for uniform scale multi-sensor systems with multiplicative noises, the optimal filtering fusion algorithms have been presented. However, a large number of phenomena or processes possess features and physically significant effects at multiple scales. At the same time, the available data are collected at several different scales (resolutions). Therefore, it is important to solve the date fusion for systems with multiplicative noises under multi-sensor observation at the different scales.Recently interest in wavelets has grown at an explosive rate. One of the more recent areas of investigation in multi-scale analysis has been the emerging theory of multi-scale representations of signals and the development of multi-scale signal processing algorithms.By use of the multi-scale analysis, this dissertation combines the model-based dynamic analysis method with the multi-scale transformation method based on the statistical characteristics. The multi-scale decentralized optimal filtering fusion algorithms are proposed for mono-channel and multi-channel systems with multiplicative noises. Finally, the final fusion estimation of the finest scale based on the entire information is achieved. These fusion algorithms in the dissertation are optimal in the sense of linear minimum-variance.The main researching contents of this dissertation are as follows:1. The development and status quo of signal estimation for SMN is recalled. Moreover, data fusion technology for multi-sensor systems and theory of wavelet transformation are introduced simply.2. By using the wavelet transform as a linking between signals at different scales, multi-scale decentralized optimal filtering fusion algorithms are developed for mono-channel and multi-channel systems with multiplicative noises which are observed by different sensors at different scales. Moreover, a proposal to reduce delay time is put...