Adaptive Data Sampling with Dual-Scale Prediction in Deterministic Edge Network

With the rapid increase of data traffic and real-time requirements for control-based services in the Electricity-Energy Internet, the information processing technologies of deterministic edge network has broad application prospects for real-time and high reliable transmission. However, the high dimensionality of sensing data, complex processing algorithms, and difficulty in learning in deterministic edge networks pose high challenges to sampling detection. Therefore, a sampling period adaptive adjustment algorithm with dual-scale prediction is proposed in this paper. Dual-scale prediction is utilized, to depict the short term trend of power equipment’s operation data and the long term fluctuation of them. The proposed algorithm realizes data reconstruction for the missing time points, which can greatly reduce the amount of transmitted data, thus satisfying high-quality transmission. Extensive simulation results validate that this proposed algorithm can effectively reduce the data reconstruction error through edge sensing.