Intelligent Reflecting Surface Configurations for Smart Radio Using Deep Reinforcement Learning

Wang, Wei; Zhang, Wei

doi:10.1109/jsac.2022.3180787

Cited by 34 publications

(29 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…where K r represents the ratio of power attenuation in the LOS component to that in the NLOS component; H UD,LOS represents the LOS component, the value of which depends on the relative positions of UAV and data center; H UD,NLOS represents the NLOS component that associated with multipath effect and is thus fast time varying. The NLOS component of the channel obeys circularly symmetric complex Gaussian distribution [23].…”

Section: Dynamic Channel Modelmentioning

confidence: 99%

“…To tackle the issues in the first area, we construct the cascaded channel model by following the paradigm of passive IRS channel estimation [22]. In IRS-assisted communication scenarios, common channel models include the cascaded channel model [23] and the spatial scattering channel model [24]. When using the cascaded channel model, the IRS is considered as an integral part of the entire wireless communication channel.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Intelligent reflecting surface‐assisted UAV inspection system based on transfer learning

Du,

Qi,

Wang

et al. 2024

IET Communications

View full text Add to dashboard Cite

Intelligent reflective surface (IRS) provides an effective solution for reconfiguring air‐to‐ground wireless channels, and intelligent agents based on reinforcement learning can dynamically adjust the reflection coefficient of IRS to adapt to changing channels. However, most exiting IRS configuration schemes based on reinforcement learning require long training time and are difficult to be industrially deployed. This paper, proposes a model‐free IRS control scheme based on reinforcement learning and adopts transfer learning to accelerate the training process. A knowledge base of the source tasks has been constructed for transfer learning, allowing accumulation of experience from different source tasks. To mitigate potential negative effects of transfer learning, quantitative analysis of task similarity through unmanned aerial vehicle (UAV) flight path is conducted. After identifying the most similar source task to the target task, parameters of the source task model are used as the initial values for the target task model to accelerate the convergence process of reinforcement learning. Simulation results demonstrate that the proposed method can increase the convergence speed of the traditional DDQN algorithm by up to 60%.

show abstract

Section: Dynamic Channel Modelmentioning

confidence: 99%

mentioning

confidence: 99%

Intelligent reflecting surface‐assisted UAV inspection system based on transfer learning

Du,

Qi,

Wang

et al. 2024

IET Communications

View full text Add to dashboard Cite

show abstract

“…Furthermore, there is an additional issue. Currently, DLbased approaches have overlooked the fact that the communication domain involves complex-value computations [34][35][36]. Yang et al trained the agent based on deep reinforcement learning (DRL), post decisional state in dynamic time varying environments to get more related information [35].…”

Section: Prior Workmentioning

confidence: 99%

“…The authors in ref. [36] also designed a DRL framework to control Controls the increment of the reflection matrix. Their work is very useful, but it does not take into account that conventional neural networks are not suitable for the field of communication security.…”

Section: Prior Workmentioning

confidence: 99%

See 1 more Smart Citation

Secure transmission for reconfigurable intelligent surface assisted communication with CV‐DDPG

Gao,

Xu,

Zhang

et al. 2023

IET Communications

View full text Add to dashboard Cite

Reconfigurable intelligent surface (RIS) has great potential in securing wireless transmission, and it can flexibly change the wireless communication environment. However, the employment of RIS increases the difficulty in beamforming of base station (BS), and existing schemes cannot be directly utilized to enhance the security of system. Therefore, a reinforcement learning framework to jointly control BS and RIS for the purpose of enhancing the secure performance is proposed. Specifically, the beamforming ia transformed with artificial noise and reflection control as a Markov decision problem (MDP) in complex domain. Then, we develop the deep deterministic policy gradient based on complex‐valued neural networks (CV‐DDPG) to imitate the rules of complex computation. Additionally, the CV‐DDPG is performed as the agent of reinforcement learning to control the BS and RIS, which enables better utilization of the implied phase information of complex values in the system. The simulation results show that the proposed model promotes the performance of the control and greatly improves the secrecy rate of the legitimate user.

show abstract

Performance investigation of channel estimation for intelligent reflecting surface assisted wireless communications using neural network

Ahmadinejad,

Talebi

2024

Engineering Applications of Artificial Intelligence

View full text Add to dashboard Cite

Intelligent Reflecting Surface Configurations for Smart Radio Using Deep Reinforcement Learning

Cited by 34 publications

References 42 publications

Intelligent reflecting surface‐assisted UAV inspection system based on transfer learning

Intelligent reflecting surface‐assisted UAV inspection system based on transfer learning

Secure transmission for reconfigurable intelligent surface assisted communication with CV‐DDPG

Performance investigation of channel estimation for intelligent reflecting surface assisted wireless communications using neural network

Contact Info

Product

Resources

About