Multi-kernel polar codes, which are constructed by different sizes of kernels make the length of the traditional polar code more flexible. In this paper, we extend the multi-kernel polar codes to 5×5 kernels including kernel design, code design, and fast decoding. First, we construct a new 5×5 kernel with the maximum polarization rate. The first row of this kernel is an all-one vector, which enable us to find the relation between node modification and distance property of multi-kernel polar codes. Second, we modify the repetition (REP) nodes to the new 5×5 kernel, which can increase the distance property of the constructed polar codes, improve the decoding performance and reduce memory requirements during the fast decoding. In addition, we find our REP node modification can be generalized to multi-kernel polar codes when the first row of the used kernels is all-one vector. Third, we prove the four kinds of basic nodes (Rate-1 node, Rate-0 nodes, Repetition (REP) nodes, and Single-Parity check (SPC) nodes) of fast decoding are suitable for the new kernel. Moreover, we generalize the Generalized Parity-Check (G-PC) nodes and the Generalized Repetition (G-REP) nodes, which are suitable for multi-kernel polar codes constructed by any size of kernels and any order of kernels. By simulations, we show that our modification improves the error-rate performance of the fast decoding. Meanwhile, the fast decoding of our multi-kernel polar codes has at least 68.4% latency reduction compared with the traditional SC decoder in all cases considered where code lengths are 80, 135, 400, and 675.

Network Intrusion Detection Systems (NIDS), as critical tools for defending against malicious attacks, still have several limitations, such as false positives and false negatives, limited adaptability to new attacks, and scalability issues. To enhance the adaptability and real-time detection capabilities of NIDS for novel attacks, this paper proposes a network intrusion detection method based on semi-supervised learning and random forest. Firstly, an autoencoder is used to reduce the dimensionality of the NSL-KDD dataset, increasing robustness against high-dimensional complex data. Sencondly, a semi-supervised learning framework is established through the DBSCAN clustering method, improving the model’s performance and stability. Finally, multi-layer perceptron (MLP) is employed to select data features and extract an optimal feature subset, and random forest (RF) classifier is constructed to reduce computational complexity and effectively boost model performance. Experimental results show that the proposed method improves accuracy on the NSL-KDD and UNSW-NB15 datasets by 8.92% and 6.1%, respectively, which demonstrates strong generalization ability and robustness, achieving precise and efficient detection and classification of network intrusions and attack types.

In this paper, a new single carrier frequency domain equalization (SC-FDE) communication scheme with code index modulation (CIM), called CIM-SC-FDE, is proposed for UAV payload communication link with high spectral and energy efficiency requirements. In this scheme, part of the information bits is used to determine the index of the spread spectrum code, which can convey additional bits. The transceiver structure of the proposed scheme is constructed, and its bit error rate (BER) is derived. Numerical results show the superiority of the CIM-SC-FDE scheme at various values of spectral efficiency (SE). And, the approximate theoretical BER curve agrees with their simulation counterparts well.

We investigated the overfitting characteristics of a reservoir-computing (RC)-based nonlinear equalizer, which is used to compensate for optical nonlinear waveform distortion in optical fiber communications systems. The RC-based nonlinear equalizer is gaining attention due to its high-speed training capability and feasibility for hardware implementation. However, its overfitting characteristics have not been thoroughly explored yet. We evaluated the overfitting of the RC-based nonlinear equalizer by comparing it with that of a well-studied three-layer artificial neural network (ANN)-based nonlinear equalizer. The number of reservoir units was 50, 100, and 200. The spectral radius of the RC was also varied to control the performance of the RC-based nonlinear equalizer. We employed pseudo-random binary sequences (PRBSs) and repeated random bit sequences (RRBSs) to train both equalizers and evaluated the overfitting on the training signals. Our results showed that when the equalizers were trained on PRBSs, the ANN exhibited stronger overfitting compared to the RC. This is because the ANN is more capable of efficiently learning the generation rules of the PRBSs and predicting the received PRBSs more easily than the RC. In contrast, when RRBSs were used to train the equalizers, the overfitting was weaker compared to the case with PRBSs in both the ANN and RC.

Field-Programmable Gate Arrays (FPGAs), with their parallelism, programmability and Partial Reconfiguration (PR), offer great potential for accelerating heterogeneous data center networks and Virtual Network Functions (VNFs). Virtual Extensible LAN (VXLAN) enables scalable virtualized networks, chaining for multi-tenant VNFs by encapsulating packets in UDP packets. However, software-based solutions like Open vSwitch (OVS) incur CPU overhead, limiting performance. This paper presents an FPGA-accelerated VXLAN chaining prototype for PR VNFs in heterogeneous data centers with a hardware-software co-design based on our previous work, RosebudVirt, which is a high-performance and PR virtualization framework designed for virtualized networks. The hardware implements a line-rate VXLAN packet-processing pipeline, with exact matching (EM) using Cuckoo hashing, longest prefix matching (LPM) based on Content-Addressable Memory (CAM) or Ternary Content-Addressable Memory (TCAM), and linear search, achieving 100 Gbps line rate. The software agent optimizes storage utilization, supports over 1 million tenants, and achieves 99.97% utilization in a 4-slot Cuckoo hashing EM table for VXLAN Tunnel Endpoint (VTEP) mapping. Implemented on the Xilinx Alveo U200 platform, our prototype achieves a 193.67-times throughput improvement and a 99.85% latency reduction compared to OVS, without additional CPU overhead, offering a scalable solution for FPGA-accelerated VXLAN chaining for PR VNFs in multi-tenant heterogeneous data centers.

Vehicle-to-Everything (V2X) networking is poised to further integrate vehicular communication with Mobile Edge Computing (MEC) technologies, enhancing autonomous vehicle driving and intelligent applications. However, the rapid growth in computational demands and vehicle mobility often results in inefficient task processing, particularly in resource-constrained environments. Additionally, the broadcasting nature of wireless channels poses significant risks to user privacy and network confidentiality. This paper addresses the challenges of task offloading and resource allocation in secure V2X networking, to improve the computational efficiency of tasks. To achieve this, we propose a Deep Reinforcement Learning (DRL)-based algorithm to obtain the optimal task offloading and resource allocation strategy by employing the Actor-Critic (AC) framework. In this algorithm, a Convolutional Neural Network (CNN) is employed as the actor network to output offloading decisions, efficiently extracting features from high-dimensional state spaces. The critic network evaluates the offloading decisions using a subcarrier allocation strategy based on a greedy algorithm and a power allocation strategy derived from an improved water-filling (IWF) algorithm. The proposed power allocation strategy offers a closed-form solution for subcarrier transmit power. Simulation results demonstrate that the proposed approach significantly enhances the computational efficiency of task processing while ensuring secure communication.

Long-term rain attenuation statistics and variability are discussed using Ka and Ku band satellite signal observations conducted at Osaka Electro-Communication University in Neyagawa, Japan, during 37 years from 1986 to 2023. Before 2006, the long-term rain attenuation statistics agree well with the ITU-R predictions. Yearly 0.01% values of the attenuation indicate fairly large variations which amount to about 20% around the mean value. Besides the yearly rainfall rate statistics, these variations seem to be caused by difference in equivalent path length in each year. However, the increase in the equivalent path length is not fully explained by that of rain height determined by the ground temperature, but rather related to the types of rainfall which frequently appear in summer time, such as shower and typhoon. The occurrence time of shower and typhoon is newly found to be correlated with the sea surface temperature anomalies off the coast of Japan in Pacific Ocean, possibly caused by the tropical climate change in the equatorial region. After 2007, the long-term rain attenuation statistics are considerably larger than the ITU-R predictions. This is found to be even caused by the increase in the occurrence time of warm, cold, and stationary fronts rather than shower and typhoon due to the recent rise of the sea surface temperature due to the climate change.

Cell-free massive multi-input multi-output (CF mMIMO) is one of the candidate elemental technologies for the sixth generation (6G) mobile communication system to provide excellent wireless quality everywhere in a service area. In addition, utilization of frequency range (FR) 3, corresponding to around 7 to 24 GHz, has been expected as a candidate new spectrum in 6G. In the such FR, a hybrid beamforming (BF) architecture, which combines digital and analog BF, has been employed considering the power consumption and implementation cost of base stations. In this paper, we reveal the impact of the antenna configuration in sub-array architecture on the spectral efficiency of CF mMIMO deploying access points (APs) with hybrid BF functionality in the FR3 band through computer simulations. Concretely, we assume that the central processing unit performs digital BF, and each AP, which is equipped with multiple radio frequency chains and array antennas, utilizes analog BF using a connected sub-array, and clarify a suitable sub-array configuration in CF mMIMO deployed mainly towards terrestrial UEs through comparative study with simulations.

In peak power reduction of orthogonal frequency-division multiplexing (OFDM) signals, discrete Fourier transform spread OFDM (DFT-s-OFDM) has the advantages of data independence, no side information, and no error degradation. A redundant version of DFT-s-OFDM with frequency-domain pulse shaping and phase rotation can further reduce the peak power. The optimal phase rotation, however, needs many complex multiplications in DFT-s-OFDM systems, limiting its practicality in resource-constrained devices like mobile terminals. This work proposes a novel DFT-s-OFDM scheme to achieve close-optimal peak power reduction with minimal computational complexity, comparable to practical DFT-s-OFDM using plain BPSK or QPSK without phase rotation. The proposed scheme outperforms conventional phase-rotated methods, such as π/2-BPSK or π/4-QPSK adopted in current systems in both peak power reduction and implementation efficiency.

The 5th-generation mobile communication system (5G) has already launched commercially since 2019 and has been introduced worldwide. 5G attracts much attention as a promising technology that can solve future industrial and social issues. Since industrial 5G use cases need not only high-speed and large-capacity communications, but also high-reliability and low-latency communications, higher requirements and performances are required compared to communication services for consumers. This paper introduces 5G experimental trials using millimeter-wave and sub-6 GHz bands in an actual manufacturing factory. The experimental trials evaluate and compare some performances of 5G wireless transmission in 28 GHz and 3.7 GHz bands using an autonomous mobile robot (AMR) in the factory. In addition, this paper describes a fixed-position experimental trial in a more realistic factory environment where humans and robots move. This experimental trial visualizes more realistic 5G performances and evaluates them from the perspective of differences in location and height in the factory. Experimental results demonstrate both advantages and disadvantages in 28 GHz and 3.7 GHz bands and the effectiveness of 28 GHz compared to 3.7 GHz in the manufacturing factory. Furthermore, from results of the fixed-position experimental trial, some issues and solutions for promoting industrial 5G in the factory are considered.

Rain attenuation characteristics were investigated by using simultaneous observation data of rain attenuation conducted on the same propagation path using Ka-band and Ku-band beacon signals of Superbird-B3 at NTT Yokosuka Research and Development Center (Yokosuka City) from June to December 2023. The ratio of rain attenuation in both frequency bands shows slightly larger values than the ITU-R frequency scaling method, suggesting that the ration is underestimated by the ITU-R predictions. The deviation in the ratio can be explained almost within the range indicated by the theoretical values of three typical kinds of raindrop size distribution: drizzle type (Jd type), standard type (MP type), and thunderstorm type (Jt type). Compared with these theoretical values, the ratio is found to be influenced by the Jd type when stationary fronts pass through, while it is influenced by the Jt type on the south of the stationary fronts. Also, it is influenced by the Jt type in convective rainfall such as summer time showers. As a result, the instantaneous values of Ka-band rain attenuation can be estimated from those of Ku-band rain attenuation with an accuracy of about 15% or less considering the raindrop size distribution. On the other hand, the two-year rain attenuation statistics were presented for BS and JCSAT-5A signal observations from April 2022 to March 2024. Then, the frequency scaling methods from Ku to Ka band are discussed at Neyagawa, Yokosuka, and Matsuyama stations. At Neyagawa and Yokosuka stations, the inferred Ka-band attenuation statistics are found to be larger than the ITU-R predictions using the 0.01% rainfall rate. At Matsuyama station, however, the rain attenuation statistics are slightly smaller the ITU-R predictions.

Geolocation of interference sources using Time Difference Of Arrival (TDOA) and Frequency Difference of Arrival (FDOA) observed via two geostationary satellites requires precise orbit information of the satellites, as the geolocation accuracy depends on the precision of orbit determination. This paper proposes a method for estimating satellite orbits using TDOA data among multiple known ground stations and derives theoretical error equations for both orbit determination and geolocation of interference source. Detailed accuracy analyses reveal that, with three ground stations and 24 hours of TDOA measurements, satellite position and velocity accuracies are approximately 0.4 km and 0.03 m/s, respectively. Furthermore, the geolocation of interference source accuracy using the determined orbit achieves approximately 8.0 km. The results also demonstrate that orbit determination accuracy heavily depends on the spatial distribution of ground stations. Notably, the northeast-southwest elongated geography of Japan necessitates careful selection of station locations to ensure high orbit determination accuracy. The theoretical formulas derived in this paper for orbit determination and subsequent geolocation accuracy enable comprehensive accuracy evaluations under various conditions without relying on Monte Carlo simulations.