| With the rapid development of modern information technology, CPS (Cyber-Physical System), especially the systems combining self-organization network have being gained widely attention because of its flexibility and practicability. However, the structure of these systems is quite complex, and that character makes the application of CPS in industrial control face great challenges.Given the complexity of CPSs, how to carry on the effective analysis of large amounts of heterogeneous information is a significant problem we should focus on. Most of previous information aggregation algorithms mainly considered the same kind of data fusion. Although they can reduce the network load effectively, yet they are far from enough for decision-making in complex systems like CPSs.In this paper, the research applied data fusion technology to self-organized CPS. After summarizing the characteristics of existing data fusion algorithms, and analyzing several typical algorithms in detail, we propose a new algorithm called STAC (Spatial-Temporal-Attribute Correlation), which combines the temporal correlation, spatial correlation and attribute correlation. Based on the concept lattice theory and weighted Bayesian theory, STAC turns a large amount of raw data to general information in the first step, considering the temporal correlation and spatial correlation. Then STAC produces more comprehensive information according to the attribute correlation.At last, series of simulation scenario are designed, and corresponding simulations on STAC algorithm are made based on NS2platform The simulation results indicate that the STAC algorithm can reduce the network load, and is able to maintain the closed-loop control automatically in CPS network at the same time. The most important thing is STAC guarantees the accuracy of event prediction. To sum up, as a multi-source heterogeneous data fusion algorithm, STAC has made a contribution to enhancing the reliability and effectiveness of self-organized CPS. |
PDF Full Text Request