A cooperative pattern division multiple access with power splitting enabled incremental relaying (PSEIR-CPDMA) is proposed for downlink satellite-terrestrial mobile networks. The strongest user decides whether to operate cooperative communication according to the feedback of the weak users. When in CPDMA mode, the strongest user as a power splitting enabled incremental relay node uses the harvested energy to relay signals. Under delay-limited transmission condition, the outage probability and throughput of the proposed system are investigated in depth, and the theoretical analysis is verified by simulation for performance evaluation. The results demonstrate that the PSEIR-CPDMA scheme can achieve better performance than the conventional CPDMA with PS (PS-CPDMA).

The growth of real-time streaming services has significantly increased network bandwidth demands. In this context, multicast technology has garnered widespread attention as an effective means to alleviate bandwidth pressure. However, the inherent complexity of multicast routing poses significant challenges for failure recovery, necessitating swift and efficient responses to ensure reliability. Moreover, if link states are not considered during the failure recovery process, it may increase the load on existing links, leading to congestion and overloading of the network. In this paper, we propose a bandwidth-aware multicast fast rerouting mechanism based on Information-Centric Networking. The proposed mechanism precomputes a set of backup paths and dynamically assigns them to multicast groups based on network conditions, enabling fast rerouting and load balancing. To implement this mechanism, we designed a backup path construction algorithm based on routing tree transformations to compute the set of backup paths. Additionally, we propose a backup path selection algorithm that selects paths for multicast groups based on link load, aiming to achieve load balancing. Furthermore, we developed a bit-index-based forwarding mechanism to ensure packet transmission along backup paths without requiring routers to maintain backup state information for each group. Extensive simulation results demonstrate that the proposed mechanism outperforms baseline algorithms in terms of overload control and forwarding efficiency.

A G-band waveguide diplexer, designed and fabricated by SU8 micromachining to obtain high precision dimensions and measured, is presented in this paper. The designed diplexer is composed of a common non-resonant (NRN) cavity junction and rectangular dual-mode resonators, which have two cross-coupled dual-mode cavities employing the TE₁₀₂ and TE₃₀₁ modes. It operates at 215-218 GHz and 221-224 GHz respectively. The entire device in the SU8 chips consists of the designed diplexer and three bends for measuring. The entire device has a total of three layers, and his entire circuit can be completed by two SU8 chips spliced together, in which SU8 chip I contains only one layer structure of the input standard waveguide port I, while SU8 chip II contains two layers structure for the designed diplexer at the top layer, the output waveguide port II and port III at the bottom layer. Between them, the use of copper as a masking layer solves the requirement of the SU8 micromachining that the top structure is larger than the bottom structure. In the two bands, the measured return loss (RL) is about 10.5 dB, the insert loss (IL) of the fabricated designed diplexer is about 1 dB, their center frequency shift are approximately 0.92% in the lower bandpass and 1.12% for the high bandpass, respectively.

For the requirements of methods for estimating radio propagation characteristics with wide estimable range, short calculation time, and high estimation accuracy in forthcoming 6G system, we proposed a novel method named color images method (CIM). Although initial performances of the CIM were evaluated, the detailed performances of the CIM such as estimable range, calculation time, and estimation accuracy have been not sufficiently studied yet. In this paper, we describe the details of the CIM first and then use the CIM to estimate path losses for different outdoor environments and various frequency bands and compare with not only measurement results but also with calculation results of a conventional ray tracing method (RTM). Comparison results show that, the CIM has the wider estimable range than the RTM. Moreover, the calculation time of the CIM is very short on the order of seconds and is 0.0002 times for the case of post-processing only and 0.0029 times for the case of including both pre-processing and post-processing, comparing with those of the RTM, respectively. In addition, root mean squared errors (RMSEs) between estimated path loss results of the CIM and measurement results are about 8 dB at 1.298, 2.2, 4.7, 26.4, and 66.5 GHz, and much smaller than those of the RTM with values of 15.1 18.4, 19.1, 22.7, and 22.8 dB, respectively.

This paper investigates performance of uplink multi-panel transmission for multi-TRP (transmission and reception point) operation in 5G-Advanced system to improve transmission performance by alleviating impacts on blockage under millimeter wave (mmWave) environments. In the third generation partnership project (3GPP), 5G NR (New radio) has been standardized and its enhancement called 5G-Advanced is being discussed, and radio interfaces and protocols for mmWave have been specified. In 3GPP release 18 (Rel-18), which is the first release of 5G-Advanced, multi-panel transmission was specified to improve the uplink transmission performance of mmWave communication by coordination of multiple TRPs and simultaneous transmission from multiple antenna panels equipped by each user equipment (UE). In this paper, we evaluate the transmission performances of uplink single frequency network (SFN) based 2-panel transmission under environment where blockage(s) exist in multi-TRP operation by computer simulation. Computer simulation confirms that the multi-panel transmission on the multi-TRP operation improves the degradation ratio on the 5%ile UE throughput from 25.5% to 11.9%.

Non-line-of-sight (NLOS) identification is a major challenge for reliable WiFi-based sensing. Existing NLOS identification methods commonly encounter limited statistical features, rely on pre-designed identification thresholds, and suffer from large computational complexity. To address these challenges, this paper presents a novel method called NLOS identification of channel state information system based on enriched features and improved random forest algorithm (NICEIR). A novel statistical feature is designed to enrich the feature set for recognition by incorporating amplitude variance and phase difference variance across all subcarriers. Furthermore, an improved random forest algorithm with C4.5 architecture is proposed to train the combined features and accurately discriminate NLOS conditions from incoming undetermined signals, eliminating the need for identification thresholds. Two pruning methods are developed for feature selection to reduce complexity and prevent overfitting. Experimental results across three different scenarios demonstrate the method’s high recognition accuracy, low computational complexity, and broad applicability.

This work analyzes the secrecy performance of a generalized selection combining (GSC)-based decode and forward (DF)-relay system over Nakagami fading channels. This study also investigates transmit antenna selection (TAS) for secrecy improvement in the GSC-based DF-relay system. This work derives the new cumulative distribution function (CDF) expressions of receive signal to noise ratio (SNR) for the proposed system. Also, this study derives the new expressions of the proposed system for the secrecy performance. Through asymptotic approach, we confirm that the sharpness of secrecy performance is proportional to the number of transmit and receive antennas as well as fading index for the proposed GSC-based DF-relay system with TAS.

This paper introduces movable antenna technology into the Reconfigurable Intelligent Surface (RIS)-assisted Non-Orthogonal Multiple Access (NOMA) communication system to explore the role of spatial domain freedom in enhancing communication system performance. It proposes a distributed irregular RIS-assisted NOMA system integrating user clustering with topology optimization. The irregular RIS topology is optimized to increase the communication freedom for NOMA users, thereby improving system performance based on channel differences. In this distributed RIS-assisted NOMA scenario, the improved spectral clustering algorithm (B-SPC) is first used to cluster the non-uniformly distributed NOMA users (UCs). Subsequently, irregular RIS-assisted communication is tailored for each UC based on fairness principles. The tabu search algorithm (TS) is then employed, utilizing the NOMA user grouping mechanism, to dynamically optimize the topology of the irregular RIS, with the goal of enhancing the channel differences between users within the group and improving decoding performance through successive interference cancellation. Additionally, the power allocation for users and the beamforming of the reflecting surface are jointly optimized to maximize the system’s total spectral efficiency. The simulation results demonstrate that combining the B-SPC algorithm with TS to optimize the irregular RIS topology enhances convergence and clustering performance, improves the uniformity of the UCs and ensures fairness in user service. Furthermore, it effectively manages user channel conditions, increasing channel diversity and significantly boosting the overall spectral efficiency of the distributed irregular RIS-assisted NOMA system.

Currently, the number of wireless stations (STAs) connected to a wireless local area network (WLAN) is constantly increasing, and the quality of service (QoS) requirements of the transferred data are diverse. Even in such congested WLAN environments, it is necessary to assign sufficient frame transmission opportunities to each STA and base station (BS) according to the QoS requirements of the transferred data. IEEE 802.11ax introduces simultaneous transmission of multiple frames by orthogonal frequency division multiple access (OFDMA) for efficiently accommodating more STAs in a WLAN. For backward compatibility, OFDMA transmission in IEEE 802.11ax is used in cooperation with carrier sense multiple access with collision avoidance (CSMA/CA). However, when to trigger OFDMA transmission and which frames to transmit simultaneously are not specified in the standard. In this paper, we propose a frame transmission method that selectively uses OFDMA transmission and CSMA/CA-based transmission, taking into account both the characteristics of the two types of transmissions and the QoS requirements of the transferred data. In this method, to take full advantage of simultaneous transmission of multiple frames, OFDMA transmission is performed after more OFDMA-transmittable frames are buffered at the BS or STAs by applying transmission waiting processes to frames of data that requires minimizing the time for transferring the entire data. By contrast, frames of data that requires minimizing the transfer delay on a frame-by-frame basis are transmitted based on CSMA/CA without applying the transmission waiting process. Our extensive simulations show that the proposed method enhances data transfer throughput by effectively transmitting more frames by OFDMA and suppresses the increase of data frame transfer delay until more data transfer load.

This paper presents detection probabilities for the physical layer cell ID (PCID) in the presence of a large carrier frequency offset (CFO) and strong phase noise (PN) for the 3GPP New Radio (NR) downlink in sub-terahertz (sub-THz) bands. In the sub-THz bands, CFO due to the frequency error of a local oscillator of a set of user equipment is large. Moreover, the PN power becomes large because of the increasing uncorrelated Gaussian PN with high frequency components and correlated Wiener PN due to a significantly wide transmission bandwidth. We compare the PCID detection probabilities of two main PCID detection methods associated with different CFO estimation methods in the presence of PN. Computer simulation results show that the PCID detection method that estimates the primary synchronization signal (PSS) received timing and sequence after fractional frequency offset (FFO) estimation based on autocorrelation using a cyclic prefix achieves a higher PCID detection probability than that used to estimate the PSS received timing and sequence before the FFO estimation based on partial cross-correlation of the PSS in the presence of a large CFO and strong PN in the sub-THz bands.

Many quality-estimation models have been researched and developed for the purpose of quality design, quality monitoring and quality control of communication services. To achieve fair optimization from the viewpoint of Quality of Experience (QoE) in a network passing through various service traffic, it is desirable to execute optimization on the basis of the QoE on a common scale. However, conventional quality-estimation models that estimate QoE have been studied for each service individually, and the estimation results are not comparable among services because the scale differs for each model. Executing individual optimization without knowing the correspondence relationship between the outputs of these models may lead to a risk of unfairness in that users using one service are treated less favorably than users using the other. Therefore, we propose a comparison method for mapping the outputs of multiple quality-estimation models on a common scale. This method involves using a comparison-quality-estimation model that compares the outputs of two quality-estimation models and estimates which one is worse (or equal). We also propose three comparison-quality-estimation models: a power-function model, logistic-function model, and machine learning (random forest regression) model. We conducted subjective quality assessment experiments for web browsing and video streaming services to construct and validate the proposed comparison-quality-estimation models. We confirmed that the random forest regression model requires a large amount of data for training, and that the accuracy of quality estimation degrades when the amount of training data is insufficient. The power- and logistic-function models showed relatively stable performance even when the amount of training data was small.

With the advancement in Generative Adversarial Networks (GANs), forged face images cannot be distinguished by eyes. These forged faces threaten social stability and personal reputation, therefore effective detection methods are needed. However, existing methods often become ineffective for the faces forged by new techniques due to the inability to extract essential differences between real and fake. To solve this problem, we present a generalization-enhanced method for forged face detection based on high-frequency features. To construct the method, a forgery model is analyzed to extract the essential difference between real and fake face images. Our method depends on the spatial distribution characteristics of the high-frequency components of the input images. By studying the histogram probability distributions of high-frequency information, it is concluded that high-frequency information is essential for forged face detection. Besides, the optimal cutoff frequency of the high-pass filter is selected. Through this high-pass filter, high-frequency images of real and fake images are extracted. Finally, the extracted high-frequency images are fed into a lightweight classification network, Xception, for final detection. To verify the generalization ability of the proposed method, we use only ProGAN-generated fake faces for training. The proposed method achieved detection accuracies of 98.47%, 100%, 100%, 85.43%, and 98.85% on face images forged by DCGAN, ProGAN, StarGAN, StyleGAN2 and DeepFake, respectively. Experimental results demonstrate that our method can not only be used in detecting fake face images but also exhibits robust generalization ability across different forgery techniques.