Sampling Rate Prediction of Biosensors in Wireless Body Area Networks using Deep-Learning Methods

Abstract

In this paper, we propose a scheme which aims at determining and forecasting sampling rate of active biosensors in Wireless Body Area Networks (WBANs). In this regard, from the first round until a certain round, the sampling rate of biosensors would be determined. Accordingly, we introduce our modified Fisher test, develop Spline interpolation method, introduce three main parameters namely information of patient's activity, patient's risk and pivot biosensor's value. Then, by employing these parameters plus introduced statistical and mathematical based strategies, the sampling rate of the active biosensors in the next round would be determined at the end of each entire round. After reaching a pre-denoted round the sampling rate of biosensors would be predicted through forecasting methods. In this regard, we develop two machine learning based techniques namely Adaptive Neuro Fuzzy Inference System (ANFIS) and Long Short Term Memory (LSTM) and compare them with four famous similar techniques. In addition to using forecasted sampling frequencies of the biosensors for controlling their energy expenditure, these forecasted values would also be used to forecast patient's status in the future. This is the first work in this domain that uses current information of the patient to determine adaptive sampling frequency and then employs the time series of determined sampling frequencies to forecast the patient's status and biosensors energy expenditure in the future. For estimating our schemes, we simulated them in MATLAB R2018b software and compared the results with a number of similar schemes. Based on the simulation results, the proposed schemes are capable to reduce data traffic by 81%, decrease energy consumption of the network by 73% while having the capability of predicting sampling rate of biosensors with 97% accuracy.