Research on Multi-Source Data Feature Fusion Methods for Wireless Sensor Networks
In wireless sensor networks,different sensor nodes may collect overlapping or similar data,resulting in waste of computing,storage,and transmission resources.To this end,a multi-source data feature fusion method for wireless sensor networks is proposed.By combining complementary set empirical mode decomposition and wavelet threshold denoising methods,noise is removed while retaining the main features of the data.The first and second layer features of multi-source data are extracted by using principal component analy-sis,which are cascaded into the final extracted multi-source data features.The maximum and minimum closeness degree in fuzzy mathe-matics is used to describe the distance between different features,achieving multi-source data feature fusion.The simulation results show that the node mortality rate after the application of the proposed method is less than 5％,the fusion delay is less than 4 ms,the average node energy consumption remains below 4.5 J,and the feature fusion accuracy in different scenarios is higher than 86.3％,indicating that the proposed method has good multi-source data feature fusion performance and can effectively improve the data transmission per-formance of wireless sensor networks.
万方数据
维普
