2020/12/12 by Yung-Lin Hsu, Hsu, Yung-Lin, Chen–Feng Liu +7
Computer Science · #Age of Information Optimization #Distributed Sensor Networks and Detection Algorithms #FOS: Computer and information sciences #Networking and Internet Architecture (cs.NI) #Stochastic Gradient Optimization Techniques
paper · pdf · doi:10.48550/arxiv.2012.06860
openalex publication_date 2020/12/12 · openalex created_date 2022/10/01 · openalex updated_date 2026/07/28
This work studies a real-time environment monitoring scenario in the\nindustrial Internet of things, where wireless sensors proactively collect\nenvironmental data and transmit it to the controller. We adopt the notion of\nrisk-sensitivity in financial mathematics as the objective to jointly minimize\nthe mean, variance, and other higher-order statistics of the network energy\nconsumption subject to the constraints on the age of information (AoI)\nthreshold violation probability and the AoI exceedances over a pre-defined\nthreshold. We characterize the extreme AoI staleness using results in extreme\nvalue theory and propose a distributed power allocation approach by weaving in\ntogether principles of Lyapunov optimization and federated learning (FL).\nSimulation results demonstrate that the proposed FL-based distributed solution\nis on par with the centralized baseline while consuming 28.50% less system\nenergy and outperforms the other baselines.\n