Intelligent metasurfaces have gained significant importance in recent years due to their ability to dynamically manipulate electromagnetic (EM) waves. Their multifunctional characteristics, realized by incorporating active elements into the metasurface designs, have huge potential in numerous novel devices and exciting applications. In this article, recent progress in the field of intelligent metasurfaces are reviewed, focusing particularly on tuning mechanisms, hardware designs, and applications. Reconfigurable and programmable metasurfaces, classified as space gradient, time modulated, and space-time modulated metasurfaces, are discussed. Then, reconfigurable intelligent surfaces (RISs) that can alter their wireless environments, and are considered as a promising technology for sixth-generation communication networks, are explored. Next, the recent progress made in simultaneously transmitting and reflecting reconfigurable intelligent surfaces (STAR-RISs) that can achieve full-space EM wave control are summarized. Finally, the perspective on the challenges and future directions of intelligent metasurfaces are presented.
Intelligent metasurfaces have gained significant importance in recent years due to their ability to dynamically manipulate electromagnetic (EM) waves. Their multifunctional characteristics, realized by incorporating active elements into the metasurface designs, have huge potential in numerous novel devices and exciting applications. In this article, recent progress in the field of intelligent metasurfaces are reviewed, focusing particularly on tuning mechanisms, hardware designs, and applications. Reconfigurable and programmable metasurfaces, classified as space gradient, time modulated, and space-time modulated metasurfaces, are discussed. Then, reconfigurable intelligent surfaces (RISs) that can alter their wireless environments, and are considered as a promising technology for sixth-generation communication networks, are explored. Next, the recent progress made in simultaneously transmitting and reflecting reconfigurable intelligent surfaces (STAR-RISs) that can achieve full-space EM wave control are summarized. Finally, the perspective on the challenges and future directions of intelligent metasurfaces are presented.
“…The MLP-based NN addressing dynamic positioning of multiple RISs to overcome the storage and computational performance limitations [255] Optimized Energy Expenditure…”
Section: Mlp-position Basedmentioning
confidence: 99%
“…The fact that the considered scheme does not have a layered structure causes it to lag behind supervised learning performance. However, [255] has demonstrated that a design outperforming supervised learning in terms of both energy and time consumption can be created without the use of a layered structure. The proposed algorithm with trained NNs by the approach of deep unsupervised learning optimizes both energy expenditure and phase configuration to outperform even the Genetic algorithm, which is considered a groundbreaking development in optimization.…”
Deep learning (DL) has proven its unprecedented success in diverse fields such as computer vision, natural language processing, and speech recognition by its strong representation ability and ease of computation. As we move forward to a thoroughly intelligent society with 6G wireless networks, new applications and use cases have been emerging with stringent requirements for next-generation wireless communications. Therefore, recent studies have focused on the potential of DL approaches in satisfying these rigorous needs and overcoming the deficiencies of existing model-based techniques. The main objective of this article is to unveil the state-of-the-art advancements in the field of DL-based physical layer methods to pave the way for fascinating applications of 6G. In particular, we have focused our attention on four promising physical layer concepts foreseen to dominate next-generation communications, namely massive multiple-input multiple-output systems, sophisticated multi-carrier waveform designs, reconfigurable intelligent surface-empowered communications, and physical layer security. We examine up-to-date developments in DL-based techniques, provide comparisons with state-of-the-art methods, and introduce a comprehensive guide for future directions. We also present an overview of the underlying concepts of DL, along with the theoretical background of well-known DL techniques. Furthermore, this article provides programming examples for a number of DL techniques and the implementation of a DLbased multiple-input multiple-output by sharing user-friendly code snippets, which might be useful for interested readers.
“…The relay selection optimization to reduce propagation loss over distance by the proposed Deep Reinforcement Learning model that can learn from the environment [241] Deep Reinforcement Learning-DNN A method that addresses the challenge in path loss optimization for RIS-aided terahertz communication having high molecular absorption and attenuation [242] Deep Reinforcement Learning-NN Resource allocation for D2D networks and phase shift configuration optimization in terms of achievable rate and computational time performance [219] Deep Reinforcement Learning-DNN Deep Reinforcement Learning approach presenting a fully energy-efficient method by optimizing the ON/OFF state of RIS elements besides transmit power [243] MLP-Position Trained NN The MLP-based NN addressing dynamic positioning of multiple RISs to overcome the storage and computational performance limitations [244] Unsupervised Learning-NN A NN trained by the approach of deep unsupervised learning optimizing both energy expenditure and phase configuration [245] ML/DL-DNN The ML-based approach leveraged by DL techniques has no need for CSI for direct mapping [218] Unsupervised Learning-DNN Proposed algorithm bypassing channel prediction process, requires fewer pilots compared to prior studies with the channel estimation [246] Unsupervised Learning-DNN Proposed deep-transfer learning-based algorithm requesting less sampled data for training process resulting in reduced hardware complexity and training load [247] Deep Reinforcement Learning Deep Reinforcement Learning-based novel architecture capable of learning channel behaviour performance in every cycle thanks to the evaluation of previous rewards for actions, so that enhance optimization of phase matrix. The channel behaviour can be interpreted by evaluating historical line-of-sight path channels to determine optimal phase configuration for further actions.…”
Section: Referencementioning
confidence: 99%
“…The fact that the considered scheme does not have a layered structure causes it to lag behind supervised learning performance. However, [244] has demonstrated that a design outperforming supervised learning in terms of both energy and time consumption can be created without the use of a layered structure. The proposed algorithm with trained NNs by the approach of deep unsupervised learning optimizes both energy expenditure and phase configuration to outperform even the Genetic algorithm, which is considered a groundbreaking development in optimization.…”
Deep learning (DL) has proven its unprecedented success in diverse fields such as computer vision, natural language processing, and speech recognition by its strong representation ability and ease of computation. As we move forward to a thoroughly intelligent society with 6G wireless networks, new applications and use-cases have been emerging with stringent requirements for next-generation wireless communications. Therefore, recent studies have focused on the potential of DL approaches in satisfying these rigorous needs and overcoming the deficiencies of existing model-based techniques. The main objective of this article is to unveil the state-of-the-art advancements in the field of DL-based physical layer (PHY) methods to pave the way for fascinating applications of 6G. In particular, we have focused our attention on four promising PHY concepts foreseen to dominate next-generation communications, namely massive multiple-input multiple-output (MIMO) systems, sophisticated multi-carrier (MC) waveform designs, reconfigurable intelligent surface (RIS)-empowered communications, and PHY security. We examine up-to-date developments in DL-based techniques, provide comparisons with state-of-the-art methods, and introduce a comprehensive guide for future directions. We also present an overview of the underlying concepts of DL, along with the theoretical background of well-known DL techniques. Furthermore, this article provides programming examples for a number of DL techniques and the implementation of a DL-based MIMO by sharing user-friendly code snippets, which might be useful for interested readers.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.