Research Project: Kablosuz Kontrol Sistemleri İçin Makina Öğrenme ve Uç Değer Teorisi Tabanlı Ultra Güvenilir Haberleşme
Loading...
Contributors
Funders
ID
TB.00698
Authors
Ergen, Sinem Çöleri
Faculty Member
Publications
Teacher-student learning based low complexity relay selection in wireless powered communications☆
(ELSEVIER, 2025) Köprü, Berkay; Ergen, Sinem Çöleri; Köprü, Berkay; Onalan, Aysun Gurur; Department of Electrical and Electronics Engineering; Graduate School of Sciences and Engineering; Yes; College of Engineering; GRADUATE SCHOOL OF SCIENCES AND ENGINEERING
Radio Frequency Energy Harvesting (RF-EH) networks are pivotal in enabling massive Internet-of-Things by facilitating controlled, long-distance energy transfer to energy-constrained devices. Relays, which assist in either energy or information transfer, significantly enhance the performance of such networks. However, the relay selection problem in multiple-source-multiple-relay RF-EH networks poses substantial computational challenges. To address these, this paper proposes a novel deep-learning-based relay selection framework that integrates convolutional neural networks (CNNs) and teacher-student learning. Specifically, the joint relay selection, time allocation, and power control problem are studied under non-linear EH conditions. First, the optimal solution to the time and power allocation problem for a given relay selection is derived. Then, the relay selection problem is formulated as a classification task, and two CNN-based architectures are proposed. To further improve computational efficiency without compromising accuracy, the teacher-student learning paradigm is employed, wherein a smaller student network is trained with the distilled knowledge of a larger teacher network. A novel dichotomous search-based algorithm is introduced to determine the optimal architecture of the student network. Simulation results demonstrate that the proposed solutions achieve lower complexity compared to state-of-the-art iterative approaches while maintaining optimality.
Wireless 6G connectivity for massive number of devices and critical services
(Institute of Electrical and Electronics Engineers Inc., 2024) Ergen, Sinem Çöleri; Kalor, Anders E.; Durisi, Giuseppe; Parkvall, Stefan; Yu, Wei; Mueller, Andreas; Popovski, Petar; Department of Electrical and Electronics Engineering; Yes; College of Engineering
Compared to the generations up to 4G, whose main focus was on broadband and coverage aspects, 5G has expanded the scope of wireless cellular systems toward embracing two new types of connectivity: massive machine-type communications (mMTCs) and ultrareliable low-latency communications (URLLCs). This article discusses the possible evolution of these two types of connectivity within the umbrella of 6G wireless systems. This article consists of three parts. The first part deals with the connectivity for a massive number of devices. While mMTC research in 5G predominantly focuses on the problem of uncoordinated access in the uplink for a large number of devices, the traffic patterns in 6G may become more symmetric, leading to closed-loop massive connectivity. One of the drivers for this type of traffic pattern is distributed/decentralized learning and inference. The second part of this article discusses the evolution of wireless connectivity for critical services. While latency and reliability are tightly coupled in 5G, 6G will support a variety of safety-critical control applications with different types of timing requirements, as evidenced by the emergence of metrics related to information freshness and information value. In addition, ensuring ultrahigh reliability for safety-critical control applications requires modeling and estimation of the tail statistics of the wireless channel, queue length, and delay. The fulfillment of these stringent requirements calls for the development of novel artificial intelligence (AI)-based techniques, incorporating optimization theory, explainable AI (XAI), generative AI, and digital twins (DTs). The third part analyzes the coexistence of massive connectivity and critical services. Specifically, we consider scenarios in which a massive number of devices need to support traffic patterns of mixed criticality. This is followed by a discussion about the management of wireless resources shared by services with different criticality.
Optimization theory and deep learning based resource allocation in net-zero-energy networks with short packets
(IEEE-Inst Electrical Electronics Engineers Inc, 2023) Önalan, Aysun Gurur; Ergen, Sinem Çöleri; Department of Electrical and Electronics Engineering; Graduate School of Sciences and Engineering; Yes; College of Engineering; GRADUATE SCHOOL OF SCIENCES AND ENGINEERING
Net-zero-energy networks enable IoT applications by balancing harvested and consumed energy when the connection to the electric grid or changing batteries is not feasible for the information sources. This letter studies the optimization problem for minimizing the schedule length for net-zero-energy networks with short packets where the schedule length is defined as the total time duration required for the RF energy harvesting (EH) in the downlink and information transmission by exhausting the harvested energy in the uplink. The problem is nonlinear and non-convex, so hard to solve. To obtain a near-optimal solution, a bi-level optimization-based framework is proposed with the master problem searching for optimal EH duration iteratively and subproblems of calculating the schedule length for a given EH time in each iteration. Then, we propose a low complexity optimization theory based deep learning framework based on the simplification of the deep learning architecture by using optimality conditions. The proposed approaches outperform state-of-art algorithms in terms of schedule length. The optimization theory based deep learning approach further decreases the complexity.
Channel prediction using deep recurrent neural network with EVT-Based adaptive quantile loss function
(Ieee-Inst Electrical Electronics Engineers Inc, 2025) Valiahdi, Parmida Sadat; Ergen, Sinem Çöleri; Valiahdi, Parmida Sadat; Mehrnia, Niloofar; Gross, James; Department of Electrical and Electronics Engineering; Yes; College of Engineering
Ultra-reliable low latency communication (URLLC) systems are pivotal for applications demanding high reliability and low latency, such as autonomous vehicles. In such contexts, channel prediction becomes essential to maintaining communication quality, as it allows the system to anticipate and mitigate the effects of fast-fading channels, thereby reducing the risk of packet loss and latency spikes. This letter presents a novel framework that integrates neural networks with extreme value theory (EVT) to enhance channel prediction, focusing on predicting extreme channel events that challenge URLLC performance. We propose an EVT-based adaptive quantile loss function that integrates EVT into the loss function of the deep recurrent neural networks (DRNNs) with gated recurrent units (GRUs) to predict extreme channel conditions efficiently. The numerical results indicate that the proposed GRU model, utilizing the EVT-based adaptive quantile loss function, significantly outperforms the traditional GRU. It predicts a tail portion of 7.26%, which closely aligns with the empirical 7.49%, while the traditional GRU model only predicts 2.4%. This demonstrates the superior capability of the proposed model in capturing tail values that are critical for URLLC systems.
