Name
Abstract | ||
|---|---|---|
The amount of connected devices has been growing tremendously over the past decade. These connected devices range from the traditional smart phones to electrical appliances, solar panels, converters, electric vehicles and wearables. Satisfying their connectivity demand is adding pressure to the wireless networks which are already pressed with serving their bandwidth-hungry mobile users. LTE Unlicensed (LTE-U) aims to exploit the unlicensed spectrum to offload mobile user (LTE users) traffic, increase capacity, and hence improve the user/device experience in an era of inflated demand. Meanwhile, WiFi is the dominant technology operating at the unlicensed spectrum. Therefore, LTE-U needs to ensure the performance of WiFi users do not degrade as LTE users offload their traffic. In this paper, we propose a Q-learning based medium access approach to enhance the Listen Before Talk (LBT) mechanism of LTE-U. Q-learning based LBT helps with the co-existence issue by enhancing the performance of WiFi users at times when LTE-U users try to access the unlicensed bands. Our results show that the proposed Q-learning based LBT reduces the end-to-end delay of WiFi users in the order of several tens of seconds in comparison to the standard LBT implementation. It also increases the delivery success rate of WiFi traffic by up to 71%. |
Year | DOI | Venue |
|---|---|---|
2018 | 10.1109/ccece.2018.8447665 | 2018 IEEE CANADIAN CONFERENCE ON ELECTRICAL & COMPUTER ENGINEERING (CCECE) |
Keywords | Field | DocType |
Medium access, Listen before talk (LBT), LTE-U, Q-learning | Base station,Wireless network,Microsoft Windows,Spectrum management,Wearable computer,Computer science,Computer network,Electronic engineering,Exploit,Converters,Reinforcement learning | Conference |
ISSN | Citations | PageRank |
0840-7789 | 0 | 0.34 |
References | Authors | |
0 | 2 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Meesam Haider | 1 | 0 | 0.68 |
| Melike Erol-Kantarci | 2 | 638 | 50.55 |