6533b7dbfe1ef96bd1270934
RESEARCH PRODUCT
Low Latency Ambient Backscatter Communications with Deep Q-Learning for Beyond 5G Applications
Riku JanttiZheng ChangMuhammad Ali JamshedHaris PervaizFurqan Jameelsubject
BackscatterWireless networkComputer science05 social sciencesReal-time computing0202 electrical engineering electronic engineering information engineering0507 social and economic geographyQ-learning020206 networking & telecommunicationsNetwork performance02 engineering and technology050703 geography5Gdescription
Low latency is a critical requirement of beyond 5G services. Previously, the aspect of latency has been extensively analyzed in conventional and modern wireless networks. With the rapidly growing research interest in wireless-powered ambient backscatter communications, it has become ever more important to meet the delay constraints, while maximizing the achievable data rate. Therefore, to address the issue of latency in backscatter networks, this paper provides a deep Q-learning based framework for delay constrained ambient backscatter networks. To do so, a Q-learning model for ambient backscatter scenario has been developed. In addition, an algorithm has been proposed that employ deep neural networks to solve the complex Q-network. The simulation results show that the proposed approach not only improves the network performance but also meets the delay constraints for a dense backscatter network.
| year | journal | country | edition | language |
|---|---|---|---|---|
| 2020-05-01 | 2020 IEEE 91st Vehicular Technology Conference (VTC2020-Spring) |