Multi-sensor Data Fusion And The Application In The Process Monitoring

As the fast development of computer technology and sensor technology,multisensor data fusion technology has been used widely in the military andcivil fields. However, the key of data fusion is the design for modeling andthe algorithm for fusing. The traditional data fusion algorithms based on synchronous data can'tget the holistic 'Best' in the computation and estimation accuracy. And theexisting asynchronous fusion algorithms have been mostly several questions,for instance time delay, the large computation and so forth. These questionsall will impact the application of the algorithms in the practical systems. According to the existing questions of traditional algorithms, we take upwith the design of the new fusion algorithm based on the previous works, andapply multisensor data fusion theory into the process monitoring. The mainworks are as follows: ⑴ Analyze the sampling of multisensor dynamic system in detail. ⑵ Aiming at the large computation of exiting algorithms, a newsynchronous fusion algorithm based on Filtering step by step is presented. ⑶ Develop a multisensor fusion algorithm based on communication fault. ⑷ A asynchronous fusion algorithm frame based on transmission delay ispresented. ⑸ According to the large computation and time delay of existingasynchronous data fusion algorithms, we present a new data fusion algorithmof asynchronous sampling founding on constant increment modeling. ⑹ Combing Multiscale data fusion theory with process monitoring, wepresent a new process monitoring method based on (Multiscale) data fusion.