Research Of Data Fusion Algorithm For Multi-Source Heterogeneous Data In Cyber-Physical System

Cyber physical system consists of computing, communication and controlability, the network computer system interaction on different entities in different ways.The cyber physical system has broad application space:intelligent transportation, precision agriculture, aviation and so on. It has great economic value and practical significance for the research on information system of physics.Interactive mode in information fusion system information of the physical world and the physical world determines the information facing the world is a huge amount of heterogeneous data, because the monitoring of the physical world is characterized by the time of the implementation of different object monitoring. The massive heterogeneous data fusion is the key to realize the cyber and the physical world. But at present the data fusion algorithm and information fusion system can’t meet the physical requirements of data processing.Aiming at the problem of fusion processing of multi-source heterogeneous data,th e work of this paper is as follows:Firstlya data fusion algorithm based on matrix tri-factorization was proposed after introducing the development of matrix factorization data fusion algorithm and furtherly research the advantages and disadvantages of those algorithms. This algorithm can still work even if some relation matrix were missed.Secondly, a data fusion algorithm of node selection based on the properties of traverse type was proposed. This algorithm reduce power consumption and network latency by deciding the way of data fusion and the roles that nodes should be.Finally, aiming at solve the problem of catastrophic failure in Cyber-Physical syst em, a KID combination optimal algorithm was proposed. This algorithm improves the stability of the system, ensure effective use of the algorithm and reduce the impact generated by the ripple effect by optimizing the combination of K and ID and makin g the value of threshold value smallest, where K is the number of two-way connection outside edge, ID refers to the cohesion within the network.Finally, using NS2 and MATLAB respectively to simulate the above three algorithm s,it can be seen from the simulation results, the algorithm proposed in this paper, has more superiority than the previous algorithm in consumption, energy, network delay an d the fusion speed, precision.