BROWSE
Detail
Journal Article
Prediction of LTE Spectrum Saturation Using Quantiles of Busy Hour RB Usage Rates
Cited - time in 
Share 
Share
- Authors
- Yunbae Kim, Hyeyeon Kwon, Seungkeun Park
- Issue Date
- 2019-08
- Citation
- IEEE Communications Letters, v.23, no.8, pp.1427-1431
- ISSN
- 1089-7798
- Publisher
- IEEE
- Language
- English
- Type
- Journal Article
- Abstract
- As long-term evolution (LTE) leads the mobile communications ecosystem, LTE operators have focused on securing the spectrum resources to accommodate the mobile traffic demand generated by their subscribers. Since the spectrum resources are used in the form of resource blocks (RBs) in LTE, analyzing the variations of RB usage rate can be a direct way to estimate the spectrum saturation predetermined in terms of percentage of RBs. Through a stochastic modeling on busy hour RB usage rates from sampled LTE cells in a metropolitan area, we derive Gaussian distributions for their quantiles. Also, devising a concept of virtual co-site, we define an effective RB (ERB) usage rate that explains the utilization of total available spectrum consisting of multiple frequency bands for an LTE operator. Finally, we predict future ERB with the help of recurrent neural networks to examine whether an LTE operator's spectrum becomes saturated before the upcoming spectrum supply.
- KSP Keywords
- Gaussian Distribution, Long Term Evolution(LTE), Metropolitan area, Mobile traffic, Resource Block, Stochastic modeling, Traffic Demand, co-site, mobile communication, multiple frequency bands, neural network(NN)