Research on Data Fusion Scheme of WSNs Based on DGM Prediction Model

Data fusion method based on prediction is one of the research hotspots in information communication fields. It mainly was focus on reduce the data redundancy and energy consumption during transmission by calling the same prediction algorithm at sensor and sink node. However, there are still several drawbacks in the fusion process, such as lack of consideration for the association of data and the uncertainty of data. Moreover, the data transmission consumption is needs to be further reduce. The purpose of this paper is to propose a novel multisensor data fusion scheme using DGM prediction model (MS-DGM) to deal with the above problems. Firstly, the dynamic cluster strategy is used to partition the sensor nodes data. Next, the Grubbs Test Method is used to detect the outliers in same cluster. In order to eliminate the uncertainty of data, grey theory, which can effectively deal with the uncertain data with definite extension and vague intension, is introduced to construct the kernel and grey area to describe the correlation between the data. Finally, combined DGM prediction model and extended matrix to mine the development trends of sequences, the data fusion of WSNs is completed. Example verification and experiments analysis shows that the MAE of the MS-DGM is decreased by 33%-49 %, and the RMSE is decreased by 29%-55%, In general, this scheme can alleviate network pressure and prolong the life of whole network with low computational cost. The DGM has been introduced into WSNs field to reduce the number of dock nodes. The data fusion scheme based on DGM can effectively reduce the transmission distance of the sensor nodes. It greatly decreases the transmission energy consumption and alleviate the network congestion. With the help of this scheme, it can greatly expand the application field and scope of WSNs.