A Variational Bayes Approach to Adaptive Channel-gain Cartography

• Published in: Proceedings of the ... IEEE International Conference on Acoustics, Speech and Signal Processing (1998) pp. 8434 - 8438
• Main Authors: Lee, Donghoon, Giannakis, Georgios B.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.05.2019
• Subjects: active learning; Attenuation; Bayes methods; channel-gain cartography; Estimation; Fading channels; Hidden Markov models; radio tomography; Shadow mapping; Uncertainty; variational Bayes
• ISSN: 2379-190X
• DOI: 10.1109/ICASSP.2019.8683300

Channel-gain cartography relies on sensor measurements to construct maps providing the attenuation profile between arbitrary transmitter-receiver locations. State-of-the-art on this subject includes tomography-based approaches, where shadowing effects are modeled by the weighted integral of a spatial loss field (SLF) that captures the propagation environment. To learn SLFs exhibiting statistical heterogeneity induced by spatially diverse propagation environments, the present work develops a Bayesian approach comprising a piecewise homogeneous SLF with an underlying hidden Markov random field model. Built on a variational Bayes scheme, the novel approach yields efficient field estimators at affordable complexity. In addition, a data-adaptive sensor selection algorithm is developed to collect informative measurements for effective learning of the SLF. Numerical tests demonstrate the capabilities of the novel approach.