In high-speed optical communication systems employing advanced modulation formats, such as 16-QAM, the nonlinear distortion caused by the analog components, including electrical drivers and transimpedance amplifiers, has become a key factor that significantly limits the overall system performance. Accurate modeling of such nonlinear behavior is critical for system design and performance improvement. On the one hand, Volterra and memory polynomial models are widely used due to their high flexibility. On the other hand, the Wiener-Hammerstein (WH) model is often preferred for its simpler structure and clearer physical interpretation. Previous studies have investigated the identification of the WH model based on either amplitude-to-amplitude modulation (AM-AM) and amplitude-to-phase modulation (AM-PM) characteristics or total harmonic distortion (THD) measurements. These methods typically rely on adaptive algorithms for identification and offer limited insight into frequency-dependent behavior. In this paper, we propose a novel approach to WH model identification based on a frequency-dependent THD from single-tone measurements. This approach enables the systematic derivation of both the frequency responses of linear filters and nonlinear polynomial coefficients. We extend the method to a parallel WH structure to incorporate frequency-dependent AM-AM/AM-PM characteristics, enabling the linear filters to reflect frequency-dependent gain compression effects. Although the proposed method uses only conventional single-tone measurements, it provides a simple and intuitive model that can accurately represent nonlinear behavior. We demonstrate that this modeling method is effective for high-speed analog components in coherent optical transceivers, where nonlinear distortion depends on frequency.

In this paper, we propose a hierarchical clustering-based trajectory planning method for multiple Unmanned Aerial Vehicles (UAVs) in UAV-assisted edge computing systems. Edge computing systems have been developed to efficiently process data collected from client nodes at edge servers deployed near the network edge. Recently, increasing attention has been focused on UAV-assisted edge computing, where lightweight edge servers are mounted on UAVs to actively accommodate requests from client nodes, i.e., the UAVs approach the client nodes to perform data collection and processing tasks. In the UAV-assisted edge computing systems, trajectory optimization of UAVs with respect to data processing time and the number of deployed UAVs is a key challenge. To address this, the proposed method applies hierarchical clustering to select the optimal waypoints for the UAVs, utilizing minimum bounding circles as the clustering criterion. It then solves the Multiple Traveling Salesman Problem to determine the trajectories of UAVs, while reducing the number of deployed UAVs. Numerical experiments demonstrate the effectiveness of the proposed trajectory planning method.

In high-frequency circuits, transmission lines operate as distributed elements. Consequently, design variables must include not only the placement and impedance of mounted components but also the characteristic impedance of the interconnecting transmission lines. Furthermore, in circuits containing nonlinear elements, the operating impedance varies according to input conditions, making analytical design difficult. To address these challenges, circuit design techniques using Genetic Algorithms (GAs) have been widely adopted. Common coding schemes include the Real-Coded Genetic Algorithm (RCGA) and the Binary Genetic Algorithm (Binary GA). Conventional RCGA typically requires the circuit topology to be predetermined. In contrast, while Binary GA offers a high degree of freedom, it suffers from the generation of problematic point-contact structures. This paper proposes a new structural representation method to resolve these issues. The circuit geometry is represented as a set of rectangular patches allowed to overlap. Optimization is performed using the vertex coordinates of each patch as parameters. This approach significantly reduces the number of design variables compared to Binary GA while offering greater geometric flexibility than conventional RCGA. The effectiveness of the proposed method is verified through the optimization of an RF-DC conversion circuit. The designed circuit was fabricated and measured, confirming its operation with high RF-DC conversion efficiency.

This paper presents a theoretical and simulation-based study of a two-element Electronically Steerable Passive Array Radiator (ESPAR) antenna designed for antenna pattern multiplexing (APM) in single-RF-chain receiver systems. We derive a closed-form analytical expression for the time-varying reactance required at the parasitic element to generate a continuously evolving radiation pattern, modeling the array factor as a rotating phasor in the complex plane. Two reactance profiles are analyzed: a practically realizable bounded case and an idealized unbounded case as its special form. The bounded formulation constrains the phase variation to a finite range, resulting in multiple spectral components, with the two strongest components exhibiting low correlation (≈ 0.20), suitable for diversity reception. The unbounded formulation produces exactly two dominant spectral components at DC and + f_s, but yields a higher correlation coefficient (≈ 0.48), reducing its diversity advantage despite its spectral compactness. A Fourier-based analysis quantifies the harmonic content, and correlation coefficients between dominant components are used to assess diversity potential. The estimated diversity gains are 7.6 dB for the bounded case and 9.0 dB for the unbounded case. The results demonstrate that bounded reactance control offers a viable and effective implementation for compact, low-complexity diversity antennas, while the unbounded case serves as a useful theoretical reference.

In the fifth-generation mobile network, massive machine-type communication (mMTC) enables a huge number of potential user devices to access one base station (BS) sporadically. To meet stringent requirements for mMTC, this paper proposes an unsourced random access scheme based on deep learning to handle the relative movement between user devices and BS. Such movements cause Doppler shift, which makes it difficult to track the time-varying channel state information (CSI) and thus, heavily damages the reliability of the symbol detection. Specifically, a DopplerFormer model based on the state-of-the-art Transformer architecture is proposed to facilitate the channel estimation and tracking of the receiver. For the channel estimation during the pilot period, the minimum mean square error, which is regarded as a typical and conventional channel estimation scheme, can hardly estimate CSI without knowledge of channel and noise statistics. On the contrary, the DopplerFormer can refine the CSI initially estimated by the least squares without the above knowledge, thereby improving the performance of subsequent channel tracking. For the channel tracking during the data period, the model can exploit the estimated CSI of two adjacent time slots during the pilot period to infer the average Doppler shift of each user device. Therefore, the temporal variation of the CSI can be compensated for. Computer simulations demonstrate that the trained DopplerFormer model can greatly enhance the accuracy of the channel estimation and tracking, while the proposed scheme is much superior to conventional approaches in terms of packet detection performance.

Cell-free massive multiple-input multiple-output (CF-mMIMO) provides service through multiple distributed access points (APs), improving the performance of traditional base station systems. However, in complex environments, such as densely populated urban areas, the presence of signal barriers and dead zones can reduce communication quality. To address the issue of communication quality degradation caused by signal obstructions, this paper constructs a reconfigurable intelligent surface assisted cell-free massive multiple-input multiple-output (RIS-CF-mMIMO) communication system that supports heterogeneous users. Different from existing research, this paper considers modeling user heterogeneity using different numbers of antennas in RIS-CF-mMIMO systems, aiming to maximize the energy efficiency (EE) of the RIS-CF-mMIMO system for heterogeneous users. To achieve this, we propose a scheme named heterogeneous deep deterministic policy gradient optimized beamforming and phase shift matrix (Het-DDPG-BF-PSM). This scheme utilizes the deep deterministic policy gradient (DDPG) algorithm to jointly optimize the receiving and transmitting beamforming matrices as well as the RIS phase shift matrix. The simulation results show that compared with existing methods, the proposed Het-DDPG-BF-PSM scheme significantly improves EE of CF-mMIMO communication systems and demonstrates its adaptability in complex and heterogeneous environments.

Reconfigurable intelligent surface (RIS) has recently emerged as a key enabling technology for enhancing physical layer security by reconfiguring wireless propagation environments. However, most existing studies rely on oversimplified channel models, thereby limiting their applicability in realistic scenarios. To overcome these limitations, this paper develops a comprehensive three-dimensional statistical channel model for RIS-assisted single-input single-output communication systems, incorporating both RIS-reflected and direct propagation links under near-field and rich-scattering propagation conditions, with respect to physical structure of the RIS and multipath fading modeling. Based on this model, the secrecy performance in terms of secrecy capacity and secrecy outage probability is evaluated with mainstream RIS configuration strategies. The impacts of various factors, including RIS deployment geometry, the physical structure and the RF characteristics of the RIS, and user locations, as well as practical non-ideal factors such as channel estimation errors and finite RIS control bits, are thoroughly investigated. Simulation results reveal that the secrecy performance critically depends on the size and the deployment position of the RIS. These findings provide valuable design guidelines for implementing RIS-assisted secure wireless communication in realistic scenarios.

This paper proposes an autonomous decentralized user association method for networks where a high-altitude platform station (HAPS) and conventional terrestrial base stations (BSs) are operated in the same coverage area. The throughput satisfaction rate for each user is defined as the ratio of the actual throughput to the target throughput. The proposed method aims to maximize the system-level throughput satisfaction rate, which is defined as the generalized mean of the throughput satisfaction rates of all users within the system coverage. The proposed method achieves optimal user association by iterating an autonomous decentralized process that does not require complex cooperation among HAPS and terrestrial BSs. In this process, each BS including HAPS first broadcasts to all users supplementary information regarding the bandwidth allocated to newly connected users. Based on this supplementary information, each user then calculates the metric for selecting the best BS and feeds it back to the BS with the highest metric. Finally, each BS determines the user to be newly connected based on the highest metric reported by multiple users. Computer simulation results show that the proposed method increases the system-level throughput satisfaction rate compared to the conventional cell range expansion method and that this increase is particularly significant when fairness among users is emphasized in terms of the throughput satisfaction rate.

The wavelength of the 300 GHz sub-THz wave is approximately 1 mm, which enables tomographic imaging at a desired distance with a high resolution of less than 1 cm in the near field, and is expected to be used in security applications, such as threat object detection. However, determining the depth of a threat object involves generating images focused at each depth, which presents a considerable computation burden. In this study, we propose a low-complexity threat object identification method to reduce the computational complexity of the depth location by roughly dividing the data into the elevation, azimuth, and distance directions and applying a clustering method based on the sensing data density called DBSCAN. The results of the analysis using the measured sub-THz data indicate that the proposed algorithm can reduce the computational overhead by more than 80% while accurately determining the areas of the threat objects.

Automatic Dependent Surveillance — Broadcast (ADS-B) is a position broadcast system for air traffic control. Although ADS-B outperforms traditional radars, positional verification — a technique used to check the validity of the position report — is necessary to detect anomaly information. A popular approach is using time difference of arrival (TDOA). However, many receiver sites are necessary to achieve higher detection performance. Therefore, performance enhancement by additional angle of arrival (AOA) measurement is proposed in this study. First, the conventional and proposed methods were mathematically modeled in a unified framework. Then, benefit of the proposed method compared with the conventional method was evaluated in numerical simulation. Finally, measurement-based analysis were conducted to consider practical implementation of the method and demonstrated its feasibility.