Multiscale Information Fusion Algorithm

In the field of target tracking, the observations of the target have multi-scale features, it is difficult to demonstrate the time-frequency information by traditional single scale information fusion. However, the essential characteristics of the target will be reflected better by multi-scale information fusion. So the algorithm research on multi-scale information fusion has important theoretical and practical value. This paper mainly focuses on the study of multi-scale information fusion based on wavelet transformation. The specific contents are as follows:First of all, some basics the paper needs such as wavelet transformation, multi-scale theory and state estimation theory are introduced. These basics laid the theoretical foundation for the study of information fusion algorithm.Then, two information sampling methods are analyzed. Because the Nyquist sampling theorem causes a lot of useless information, this paper studies data collection and recovery algorithm based on compressed sensing. The simulation results show that the algorithm can use small amount of data to regain the original signal which reduces the pressure of storing and transmitting the data.The paper studies a muti-scale information fusion algorithm based on harr wavelet. In order to meet the requirement of real time capability, we improve the estimation algorithm of projection operator, and then get a multi-scale model for information fusion. An algorithm is proposed solving the problem about real-time fusion of information of different scales from different sensors and getting the optimal estimation of the system's state. The simulation results show the validity of the algorithm and its superiority compared with the algorithm on a single scale.Finally, consider the problem that the practical dynamic system is offen nonlinear, we focus on a multi-scale system of which the state equation is nonlinear and the output equation is linear, build its fusion model, based on which we get an algorithm. The simulation results show that the algorithm is an effective method for the kind of nonlinear dynamic system.