In the sixth-generation (6G) communications, how to deploy and manage massively connected Internet-of-Things (IoT) nodes will be one of the key technical challenges, because 6G is expected to provide 10 times higher connection, compared to 5G. At the same time, due to the sharp growth in connected devices and newly adopted technologies, learning-based attacks and big data breaches are expected to occur more frequently. With the advances of quantum computing in the future, conventional cryptography-based security protocols may be obsolete in the future wireless networks, which makes physical layer security (PLS) an attractive alternative or complement for secure communications. In this context, cooperative beamforming (CB)-based PLS schemes are known to be effective solutions to guarantee high secrecy rate with IoT devices, which have limited power and hardware complexity. However, the existing CB-based PLS algorithms suffer from extremely low secrecy rate, in case that eavesdroppers are close to the intended receiver. To overcome such critical issue, in this paper, we propose a CB-based PLS with artificial noise (AN) injection, which can be realized in a fully distributed manner to minimize the overhead in the IoT networks with a large number of devices. We analyze the array factor using the virtual antenna array (VAA) created by the proposed PLS algorithm. Then, the secrecy rate is derived in a closed-form expression, which can be used to optimize the performance for given system parameters in both the absence and presence of channel state information (CSI) error. The proposed scheme provides considerably higher secrecy rate compared to the conventional CB-based PLS schemes, when an eavesdropper exists near to the intended receiver. Furthermore, through simulation and numerical results, we show that the secrecy rate of the proposed scheme can be maximized by adjusting the ratio between the data beamformer and AN injection beamformer components. As a result, the proposed method shows a performance improvement of up to two times compared to the conventional CB-based PLS schemes, in terms of the secrecy rate. Such performance gain increases as the angular location of Eve becomes closer to that of Bob, which corresponds to the most vulnerable situation of the conventional CB-based PLS algorithms.

Multibeam low-Earth orbit (LEO) satellite communication is a promising solution for providing high-data rate and wide area coverage. Therefore, satellite communication is introduced into Internet of Things (IoT) networks to support large-scale connectivity. In the satellite communication system, multibeam management and power control are challenging issues because interbeam interference severely affects system performance and power consumption influences the battery life of the satellite. Thus, traditional LEO satellite systems mainly focus on minimizing a capacity-demand gap to develop an effective power reduction algorithm. In contrast to this approach, in this article, we present a theoretical analysis of the optimal conditions for minimizing the transmit power of the satellite while satisfying the traffic demands of users in multibeam LEO satellite-assisted IoT networks. Based on this analysis, we propose algorithms for user-beam association, beam pattern selection, timeslot scheduling, and power allocation to minimize the transmit power while satisfying the traffic demands. In addition, we provide low-complexity algorithms for power minimization to reduce the computational complexity. Simulation results demonstrate that the proposed methods outperform the conventional schemes in terms of power consumption, capacity-demand gap, and computational complexity.

In this paper, we consider downlink power-domain non-orthogonal multiple access (NOMA) in heterogeneous networks (HetNets) and propose resource allocation algorithms for subchannels and transmit powers to improve the sum rate performance while satisfying a minimum data-rate requirement. The proposed subchannel allocation scheme is an iterative algorithm to achieve NOMA gain by selecting the best subchannel from the viewpoint of each user, without the constraint of the number of NOMA users on each subchannel. The proposed power allocation scheme for NOMA is a deep neural network (DNN)-based unsupervised learning algorithm, where the output of the subchannel allocation scheme is used, and unsupervised learning is adopted to reduce the training complexity, as compared to supervised learning. Through simulation, we show that the proposed subchannel allocation scheme provides better sum rates compared to the conventional two-sided matching scheme, and the proposed power allocation scheme achieves a comparable sum rate to the interior point method (IPM).

In non-orthogonal multiple access (NOMA) systems, transmit power allocation plays a crucial role in maximizing the sum rate, while satisfying quality of service requirements of users. In prior study, an interior point method (IPM) has been presented as an iterative algorithm for power allocation, which can maximize the sum rate and minimize the outage rate. However, such iterative algorithms typically suffer from low computational efficiency and excessively high latency. To overcome these limitations, machine learning-based power allocation schemes have been widely studied. Along the same line, in this paper, focusing on unsupervised learning to improve the computational complexity for power allocation as well as user selection, we propose a deep neural network (DNN)-based joint user selection and power allocation technique for the cellular downlink multiple-input single-output(MISO) NOMA systems, where the transmit antenna selection and multi-cell cooperation are considered. Especially, an optimization problem for user selection and power allocation is formulated to maximize the sum rate under the constraint of minimum data-rate requirement. We propose a new loss function and training model for user selection and power allocation to solve the optimization problem, where the training and testing models are based on the signal-to-interference-plus-noise-ratio (SINR) information instead of channel state information, which can significantly reduce the input data size in DNN. Through simulations, we show that our proposed scheme provides comparable performances in terms of sum rate and outage rate to the optimal solutions, while requiring lower computational complexity compared to the optimal ones.

In this work, we propose a novel algorithm for subchannel and power allocation for Internet of Things (IoT) networks using downlink multicarrier nonorthogonal multiple access (MC-NOMA). Unlike single-carrier NOMA, MC-NOMA can utilize multiple subchannels, which is more suitable for supporting the massive connectivity of IoT users. However, in MC-NOMA, the joint subchannel and power allocation problem leads to a mixed-integer nonlinear programming problem, which is challenging to find an optimal solution. Therefore, in this article, we reformulate the joint subchannel and power allocation problem into a binary decision problem for subchannel allocation with a mathematical analysis of power allocation. Then, using the transformed problem, we propose a subchannel allocation scheme for MC-NOMA to improve the sum rate and outage performances compared with the conventional approaches. Even though many prior studies on the power allocation for MC-NOMA focused on deep learning-based methods to achieve an optimal solution with low complexity, it is difficult to jointly optimize the maximum power of each subchannel and the power for each NOMA user. Thus, we propose a deep learning-based training algorithm to optimize the maximum per-subchannel power with a mathematical analysis of power allocation for NOMA users. In addition, we introduce the user selection algorithm to avoid performance loss due to an outage user, where the presented algorithm can select users satisfying the data rate requirement. Through simulations, we show that the proposed subchannel and power allocation schemes have outstanding sum rate and outage performances compared with the existing schemes.

In the overlay device-to-device (D2D) communication systems, transmit power control is critical to better manage interference, so that the sum rate is maximized. Such power control for sum-rate optimization is NP-hard, which is typically tackled by iterative algorithms such as weighted minimum mean square error (WMMSE) method. However, the iterative power control schemes inherently incur high complexity and excessive latency. To overcome the limitations, we propose a deep learning-based power control scheme with reduced complexity and latency, where partial and outdated channel state information (CSI) is considered. Using a deep neural network (DNN)-based approach, we formulate an optimization problem to maximize the spectral efficiency under the constraints of user fairness and energy efficiency, where the DNN-based method is based on unsupervised learning with no label data generation process. In addition, a CSI reporting method based on the channel-to-interference power ratio is proposed for partial CSI feedback, which considerably reduces the feedback overhead. Through simulations, we show the results of the spectral efficiency, energy efficiency, and fairness performance for various topographical sizes and channel correlation coefficients. Also, it is shown that the proposed scheme achieves better spectral efficiency and energy efficiency than the WMMSE scheme even when it uses a small amount of CSI feedback.