2020/03/20 by Paul Ferrand, Ferrand, Paul, Alexis Decurninge +3 · 5 citations
Engineering · #Advanced MIMO Systems Optimization #FOS: Computer and information sciences #FOS: Electrical engineering #Indoor and Outdoor Localization Technologies #Information Theory (cs.IT) #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Millimeter-Wave Propagation and Modeling #Networking and Internet Architecture (cs.NI) #Signal Processing (eess.SP) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2004.00363
openalex publication_date 2020/03/20 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
We consider the use of deep neural networks (DNNs) in the context of channel\nstate information (CSI)-based localization for Massive MIMO cellular systems.\nWe discuss the practical impairments that are likely to be present in practical\nCSI estimates, and introduce a principled approach to feature design for\nCSI-based DNN applications based on the objective of making the features\ninvariant to the considered impairments. We demonstrate the efficiency of this\napproach by applying it to a dataset constituted of geo-tagged CSI measured in\nan outdoors campus environment, and training a DNN to estimate the position of\nthe UE on the basis of the CSI. We provide an experimental evaluation of\nseveral aspects of that learning approach, including localization accuracy,\ngeneralization capability, and data aging.\n