Rapid Non-invasive Blood Glucose Monitoring Based on Multi-Source Information Fusion

Metabolic heat integration method may realize effective non-invasive blood glucose monitoring. However, it is difficult to adopt the multi-source multi-parameter coupling, and the accuracy of predicting blood glucose for the method such as with multiple linear regression is low. For this reason, a rapid non-invasive blood glucose monitoring method based on the multi-source information fusion is proposed in this paper. On the basis of analyzing the mathematical relationship between the blood glucose concentration and the metabolic heat data, the data are normalized to overcome the influence of the metabolic parameters and the historical measurement data. Then, the metabolic parameter features are selected by using a correlation analysis algorithm. The metabolic parameter features and blood glucose concentration are combined to construct the blood glucose model based on the extreme learning machine tool. The metabolic heat parameter and the blood glucose data acquisition platform are constructed. Finally, experimental data validates the effectiveness of the proposed method..