Due to the availability of wide bandwidths in the frequency range 2 (FR2) band, sixth-generation (6G) networks can provide both radar and communication services using shared spectrum and hardware, reducing costs and enhancing efficiency via integrated sensing and communication (ISAC) systems. Motivated by the advantages of orthogonal time frequency space (OTFS) modulation in high-mobility scenarios, this paper proposes an iterative hybrid precoder design method and a radar tracking algorithm for OTFS-ISAC systems with aerial targets, i.e., drones. Information is sent to the ground user through communication links using the subarray hybrid precoder to balance radar and communication performance through a weighted summinimization framework. The cubature Kalman filter (CKF) is employed for radar prediction and tracking. Simulation results indicate that the proposed hybrid precoder design algorithm effectively balances radar and communication performance, while the CKF-based radar tracking method outperforms other nonlinear Kalman filters.

Research on sixth-generation (6G) wireless networks has gained significant attention as wireless communications technologies advance. In the upcoming 6G era, artificial intelligence (AI) is expected to play a significant role in enhancing mobile communications. In particular, the application of AI techniques in channel estimation can enable accurate channel state information, even in dynamic scenarios. However, the limited computational resources in user equipment often prevent the deployment of complex algorithms, necessitating adaptive channel estimation solutions, balancing the accuracy and complexity dynamically. Conventionally, AI-based channel estimation algorithms rely on explicitly stacking deep learning (DL) layers/blocks, making adaptation challenging. This paper proposes an adaptive Implicit DL Channel Estimation Network (ICENet) that employs a lightweight, implicit network design to achieve dynamic adaptability. Numerical results show that our approach can achieve the trade-off between algorithm complexity and channel estimation accuracy by adapting based on channel quality. Additionally, it offers reduced memory cost compared to explicit layer/block-stacked networks while maintaining or surpassing their estimation accuracy. Furthermore, we analyze key factors influencing forward and backward propagations in ICENet and regularize the Jacobian matrix to ensure stable convergence during the training process.

This paper investigates intelligent predictive beamforming design for simultaneous wireless information and power transfer-integrated sensing and communication (SWIPT-ISAC) systems for low-altitude economy wireless networks. Considering the downlink scenario where the base station aims to localize the moving targets/communication users and also transfer power to them, we formulate a weighted sum optimization problem to balance the trade-off between achievable communication rate and harvested energy, subject to sensing accuracy constraints defined by the Cramér–Rao lower bound. To address the non-convexity of the problem, we propose the Time-Spatial Fusion Network (TSFusionNet), an unsupervised deep learning (DL) framework that leverages multi-step historical channel state information for predictive beamforming design. TSFusionNet integrates convolutional and recurrent layers with a differential attention mechanism to capture spatial-temporal dependencies and mitigate non-stationary channel dynamics. We introduce a dynamic penalty-based loss function to enforce sensing constraints during training. Simulation results show that by adjusting the weight factor, the proposed method achieves a trade-off between rate and energy while meeting sensing accuracy requirements. Moreover, it significantly reduces computational complexity by up to approximately 96.8% in parameters and 81.5% in FLOPs, compared to existing DL frameworks.

In the era of upcoming sixth-generation (6G) wireless systems, the intelligent integrated sensing and communication (ISAC) paradigm has emerged as a pivotal research domain, catalyzing advancement across a wide range of applications. In this paper, we investigate an ISAC-assisted anti-eavesdropping communication system, where an ISAC ground base station exploits its radar function to track potential aerial eavesdroppers and implements predictive beamforming to ensure secure communications with multiple ground users. We harness the powerful capability of the Transformer for time series prediction to establish a novel deep neural network, termed the ISACformer, for constructing predictive beamformers via exploiting previously estimated channel state information in an unsupervised manner. By eliminating the need for explicit channel prediction, our proposed framework effectively reduces signaling overhead and complexity. In addition, by formulating a weighted objective function, our design meticulously balances the trade-off between the ergodic achievable worst-case secrecy rate for ground users and the ergodic Cramér-Rao lower bound for the kinematic parameters of potential aerial eavesdroppers. Simulation results demonstrate that the proposed ISACformer can deliver the desired predictive beamforming for harmonizing radar and communication functionalities effectively. Moreover, our method achieves performance approaching the theoretical upper bound obtained by ignoring multi-user interference, thereby highlighting the robustness of the proposed approach.

The implementation of indoor localization is being facilitated through commercially available low-cost mm-wave sensors. These sensors generate point cloud outputs containing noisy estimates of detected targets due to hardware noise and multipath reflections. In contrast to previously studied regression approaches, this study introduces a classification-based approach to predict the angle-of-arrival (AoA) and range of a human target from point clouds obtained from an mm-wave sensor. Our proposed methodology achieves a 7% and 26% improvement in AoA and range prediction, respectively, compared to the baseline models. All experiments have been conducted and validated using real data recorded by the mm-wave sensor.

Avstract-The primary concern in 5G and beyond wireless technologies revolve around the exponentially growing number of network users and subsequently increasing demand for frequency spectrum utilization. Traditional static spectrum allocation schemes have proven to be highly inefficient in using the available spectrum. In recent years, Dynamic Spectrum Sharing (DSS) solutions based on Cognitive Radio (CR) and Artificial Intelligence (AI)/Machine Learning (ML) have been proposed. CR involves spectrum sensing, spectrum sharing and decision-making paradigms essential for efficient use of radio frequency spectrum. AI/ML techniques, with their autonomous classification, learning, and decision-making capabilities, offer an improved approach. This paper proposes a Support Vector Machine (SVM) based ML technique for spectrum sensing in beyond 5G networks. A theoretical examination of detection and false alarm probabilities is first conducted. It is shown that the SVM algorithm achieves a high probability of detection at 99.66%, ensuring reliable spectrum sensing. Further, it is verified through simulations that the experimental outcomes align closely with the theoretical results concerning both detection and false alarm probabilities.

Simultaneous refracting/transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) has emerged as a potential technology for future-generation wireless networks to support extremely high data rates with a broader coverage area. In this work, with an aim to provide a novel analytical framework, we investigate the performance of a STAR-RIS assisted full duplex (FD) wireless communication system under finite block length (FBL) transmission. In particular, we first derive the probability density function and cumulative distribution function of the signal-to-interference-plus-noise ratio (SINR) for the uplink and downlink users. We then analyze the system performance by deriving closed form expressions for their block error rate (BLER) and goodput. Finally, we validate the accuracy of the derived analytical expressions using Monte-Carlo simulations and show that as the number of elements in the STAR-RIS is increased the system performance also improves. Furthermore, we graphically demonstrate the impact of imperfect channel state information and compare the performance of STAR-RIS in mode switching (MS) and energy splitting (ES) protocol.

In this paper, we investigate a full duplex (FD) multi-user non-orthogonal multiple access (NoMA) communication system based on the optimization of received signal-to-interference-plus-noise ratio (SINR) per unit power. Since the communication system operates in the FD mode, co-channel interference (CCI) and self-interference (SI) dominate the system's performance. Accordingly, to combat the CCI, we adopt a game-theoretic approach and propose users' clustering algorithms and to suppress the SI, we formulate an optimization problem to maximize the power-normalized SINR (PN-SINR). While the user clustering optimization problem is constrained by: 1) the successive interference cancellation (SIC) constraint and 2) two binary constraints for the allocations of uplink (UL) and downlink (DL) users, the PN-SINR problem is constrained by: 1) total transmit power budget at the base station and UL users; 2) the fundamental condition for the implementation of successive interference cancellation in the NoMA; and 3) the minimum fairness condition for the UL users. The original PN-SINR problem is non-convex and hence is converted into an equivalent subtractive-form problem, after which we propose an iterative algorithm to find the optimal power allocation policy. Properties of all the proposed algorithms are thoroughly investigated and the numerical results are provided. Based on the channel conditions and suppression level of SI and CCI, the superiority of the proposed FD-NoMA system over half-duplex NoMA and FD orthogonal multiple access systems is verified.

In this paper, a 3-hop orthogonal frequency division multiple access (OFDMA) based variable- gain amplify and forward (AF) relaying network with selection combining receiver is considered, which is based on a scattered infrastructure known as shared user equipment (UE)-side distributed antenna system (SUDAS). The considered system model uses millimeter wave (mmWave) wireless links along with ultra high frequency (UHF) links, which can provide high data rate transmission (targeted for 5G systems). Wireless links are considered to be independent but non-identically distributed (i.n.i.d.), since Rician and Rayleigh distributions are preferred for mmWave and UHF links, respectively. In this work, closed-form expressions for the lower and upper bound of the outage probability are derived for the system under consideration. Derived analytical results are verified through Monte-Carlo simulations and impact of the Rician K-factor(s) of the considered links is also highlighted.

In this paper, the performance of a cloud radio access network (CRAN) is analysed, which consists of multiple randomly distributed remote radio heads (RRHs) and a macro base station (MBS). Different from previous works on CRAN where Poisson Point Process (PPP) is used to model spatial distribution of RRHs, a more realistic Matern Hard-core point process (MHCPP) model is adopted in this work. To compare system performance of CRAN when different transmission strategies are used, two RRH selection schemes are adopted including 1) the best RRH selection (BRS) and 2) all RRHs participation (ARP). Considering downlink transmission, the outage probability and system throughput of CRAN are analytically characterized. The presented results demonstrate that compared to PPP model, the presence of hard-core distance will increase outage probability. Furthermore, the BRS scheme is more energy-efficient than the ARP scheme. Moreover, it is shown that the hard-core distance has a more significant impact on systems with higher intensity of PPP distributed candidate points and in large hard-core distance regime increasing the intensity of candidate points can only provide a small improvement in outage performance.

Recently, the new concept of Licensed Shared Access/Authorized Shared Access (LSA/ASA) has emerged as a feasible commercial version of dynamic spectrum reuse based on Cognitive Radio (CR) technologies, e.g., via spectrum sensing or by exploiting geo-location information. This paper considers the problem of effective spectrum sharing between a colocated Multiple-Input-Multiple-Output (MIMO) radar that monitors the existence of a target and a wireless communications system. More specifically, the investigated scenario considers the downlink of a communications system represented by a Base Station (BS) trying to reuse the spectrum allocated for a colocated MIMO radar in order to communicate with an assigned terminal, in the vicinity of the radar system. We present an accurate model for the operation of the wireless system in the downlink, while the MIMO radar tries to maintain an acceptable detectability level of a target in the far field. The target detection problem is reformulated using a sensing approach based on energy detection, while the BS applies beamforming to null the interference created at the radar receiver. Based on the theory of Hermitian quadratic forms and with the aid of the Linearly Constrained Minimum Variance (LCMV) beamforming solution, the performance of target detection, when the MIMO radar coexists with the data transmission is quantified and numerical results show that spectral coexistence is feasible.

This paper studies the energy efficiency (EE) of the cloud radio access network (C-RAN), consisting of multiple remote radio heads (RRHs) equipped with electronically steerable parasitic array radiator (ESPAR) antennas, which provide multiple antenna functionality with a single radio frequency (RF) chain. An EE optimization problem is formulated to obtain the configuration of ESPAR and the closed-form expressions of the voltage feeding and the loadings are derived for signal transmission at each RRH. Specifically, we obtain the closed-form expressions for precoder and power allocation that are applicable not only for the ESPAR based system but also for standard MIMO antenna (SMA) system with multiple RF chains. It is shown that EA system can be configured with less complexity compared with SMA system in block fading channel. Furthermore, symbol error rate (SER) and EE performances are compared for EA and SMA based systems. It is shown that the system with EA provides better EE performance while providing similar SER performance. From our results, it is proved that EA system can provide better performance to satisfy the requirement of 5G wireless communication networks.

In this paper, we analyze the performance of a cloud radio access network (CRAN), consisting of multiple randomly distributed remote radio heads (RRHs) and a macro base station (MBS), each equipped with multiple antennas. To model the spatial distribution of RRHs and analyze its performance, we use stochastic geometry tools. In contrast to previous works on CRAN that consider Poisson Point Process (PPP) model for the spatial distribution of RRHs, we consider a more realistic Matérn hard-core point process (MHCPP) model that imposes a certain minimal distance (referred to as hard-core distance) between the two RRHs so that the RRHs are not too close to each other. To compare system performance of CRAN when different transmission strategies are used, three RRH selection schemes are adopted including 1) the best RRH selection (BRS); 2) all RRHs participation (ARP); and 3) nearest RRH selection (NRS). Considering downlink transmission, the ergodic capacity, outage probability, and system throughput of CRAN are analytically characterized for different RRH selection schemes. The presented results demonstrate that compared to PPP model, the increase in hard-core distance will result in a higher outage probability and cause a negative impact on ergodic capacity. Furthermore, when the same total transmit power is consumed, BRS scheme provides the best outage performance while ARP scheme is the best RRH selection scheme when the same transmit SNR at each RRH is assumed. Moreover, it is shown that the hard-core distance has a more significant impact on systems with higher intensity of PPP distributed candidate points and in large hard-core distance regime increasing the intensity of candidate points can only provide a small improvement in outage performance. We extend our work to multiuser case with zero-forcing (ZF) precoding where it is proven that the results in multiuser case reduce to the derived results in this work by substituting K=1 for single-user.

In this paper, we use stochastic geometry to characterize spatially distributed multi-antenna users within a cell that consists of a single multiple-input multiple-output (MIMO) base station (BS) equipped with a large antenna array. We also use large dimensional random matrix theory (RMT) to achieve deterministic approximations of the sum rate of this system. In particular, we consider the users inside the cell to follow a Poisson point process (PPP). The sum rate of this system is analyzed with respect to (i) the different number of antennas at the BS as well as (ii) the intensity of the users within the coverage area of the cell. We obtained closed-form approximations for the deterministic rate at low signal-to-noise ratio (SNR) and high SNR regimes, which have very low computational complexity. We also derive the deterministic rate corresponding to a general user who is chosen from a set of users ordered in accordance with PPP.

This paper investigates the multiuser diversity introduced by opportunistic user selection in cognitive radio (CR) networks, where multiple cognitive users request to access the spectral resources of the licensed (primary) user. We investigate a simple cognitive user selection strategy aiming at maximizing the received signal-to-interference-plus-noise ratio (SINR) for a given power budget, under interference constraints to the primary. We study the statistics of the SINR at the cognitive receiver, and derive exact analytical expressions of its probability density function (PDF). We then analytically calculate the diversity gains introduced in the system due to the selection of one cognitive user amongst multiple candidates compared to the case when only one cognitive user exists and no selection occurs. Furthermore, we utilize the PDF of the SINR to predict the bit error rate (BER) of the selected cognitive user. Finally, the asymptotic behavior of the diversity gains for the low transmit power region of the primary and cognitive links, and as the number of candidate links becomes large is also investigated. All three multiaccess scenarios are investigated, namely multiple access channel (MAC), broadcast channel (BC) and parallel access channel (PAC), and the results show that the analytically derived expressions closely match simulated performance.

In this paper, we investigate the outage performance of interference aided adaptive linear precoding in the downlink of a multiuser multiple-input-single-output (MISO) overlay cognitive radio (CR) network. While previous research studies on linear precoding techniques in CR network aim to limit or completely cancel the interference to the primary users while achieving maximum downlink throughput of the CR system, we here make use of adaptive linear precoding at the cognitive base station (CBS) to exploit the interference to the primary and the secondary systems. We show by analysis and simulations that the outage performance of the proposed precoding technique is significantly enhanced in comparison to the conventional techniques.

This paper investigates the multiuser diversity introduced by opportunistic user selection in the cognitive radio parallel access channel (CR-PAC), where multiple cognitive users request to access the spectral resources of the licensed user. Assuming a simple cognitive user selection strategy based on maximizing the received signal-to-interference-plus- noise ratio (SINR), we study the statistics of the SINR at the cognitive receiver. We then use this to analytically calculate the diversity gains introduced in the system due to the selection of one cognitive user amongst multiple candidates, compared to the case when only one cognitive user exists in the network and no selection occurs. Finally, we investigate potential gains for the primary network from this user selection. The results show a close match between analytical expressions and simulation results, while a tight lower bound for the multiuser gain is derived in closed form.