Targeted Learning for the Dynamic Selection of Channel Estimation Methodology
The explosive expansion of collected data-in terms of dimensionality, diversity, volume-increases more rapidly than we can analyze to draw useful conclusions, make informed decisions, and provide specific recommendations. Various fields such as medical, healthcare, aviation, telecommunication require new tools to process the data which they collect to process effectively and economically and benefit from the estimated quantities that were learned from the data itself. In particular, there are different methodologies proposed and used in telecommunications to estimate the channel coefficients of different types of channels. All these methodologies are grounded based on the assumption of the statistical property of the channel. However, a flexible solution that can dynamically deploy different methods based on the received signal yields higher performance and maintained over time. In this paper, we propose to apply targeted learning and explore the suitable parameters for a communication system. The initial results demonstrate it is possible to distinguish and identify the best methodology to fit the current channel conditions.
A. M. Chandran et al., "Targeted Learning for the Dynamic Selection of Channel Estimation Methodology," Proceedings of the 2020 IEEE International Conference on Smart Computing (2020, Bologna, Italy), pp. 250 - 252, Institute of Electrical and Electronics Engineers (IEEE), Nov 2020.
The definitive version is available at https://doi.org/10.1109/SMARTCOMP50058.2020.00055
2020 IEEE International Conference on Smart Computing, SMARTCOMP 2020 (2020: Sep. 14-17, Bologna, Italy)
Electrical and Computer Engineering
Mathematics and Statistics
Keywords and Phrases
Channel Estimation; Target Parameters; Targeted Learning; Targeted Maximum Likelihood Estimator (TMLE)
International Standard Book Number (ISBN)
Article - Conference proceedings
© 2020 Institute of Electrical and Electronics Engineers (IEEE), All rights reserved.
06 Nov 2020