Table of Content

    21 September 2023, Volume 21 Issue 3
    Download the whole issue (PDF)
    The whole issue of ZTE Communications September 2023, Vol. 21 No. 3
    2023, 21(3):  0. 
    Asbtract ( )   PDF (33784KB) ( )  
    Related Articles | Metrics
    Special Topic
    Special Topic on Reinforcement Learning and Intelligent Decision
    GAO Yang
    2023, 21(3):  1-2.  doi:10.12142/ZTECOM.202303001
    Asbtract ( )   HTML ( )   PDF (334KB) ( )  
    References | Related Articles | Metrics
    Double Deep Q-Network Decoder Based on EEG Brain-Computer Interface
    REN Min, XU Renyu, ZHU Ting
    2023, 21(3):  3-10.  doi:10.12142/ZTECOM.202303002
    Asbtract ( )   HTML ( )   PDF (1551KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    Brain-computer interfaces (BCI) use neural activity as a control signal to enable direct communication between the human brain and external devices. The electrical signals generated by the brain are captured through electroencephalogram (EEG) and translated into neural intentions reflecting the user's behavior. Correct decoding of the neural intentions then facilitates the control of external devices. Reinforcement learning-based BCIs enhance decoders to complete tasks based only on feedback signals (rewards) from the environment, building a general framework for dynamic mapping from neural intentions to actions that adapt to changing environments. However, using traditional reinforcement learning methods can have challenges such as the curse of dimensionality and poor generalization. Therefore, in this paper, we use deep reinforcement learning to construct decoders for the correct decoding of EEG signals, demonstrate its feasibility through experiments, and demonstrate its stronger generalization on motion imaging (MI) EEG data signals with high dynamic characteristics.

    Multi-Agent Hierarchical Graph Attention Reinforcement Learning for Grid-Aware Energy Management
    FENG Bingyi, FENG Mingxiao, WANG Minrui, ZHOU Wengang, LI Houqiang
    2023, 21(3):  11-21.  doi:10.12142/ZTECOM.202303003
    Asbtract ( )   HTML ( )   PDF (1238KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    The increasing adoption of renewable energy has posed challenges for voltage regulation in power distribution networks. Grid-aware energy management, which includes the control of smart inverters and energy management systems, is a trending way to mitigate this problem. However, existing multi-agent reinforcement learning methods for grid-aware energy management have not sufficiently considered the importance of agent cooperation and the unique characteristics of the grid, which leads to limited performance. In this study, we propose a new approach named multi-agent hierarchical graph attention reinforcement learning framework (MAHGA) to stabilize the voltage. Specifically, under the paradigm of centralized training and decentralized execution, we model the power distribution network as a novel hierarchical graph containing the agent-level topology and the bus-level topology. Then a hierarchical graph attention model is devised to capture the complex correlation between agents. Moreover, we incorporate graph contrastive learning as an auxiliary task in the reinforcement learning process to improve representation learning from graphs. Experiments on several real-world scenarios reveal that our approach achieves the best performance and can reduce the number of voltage violations remarkably.

    A Practical Reinforcement Learning Framework for Automatic Radar Detection
    YU Junpeng, CHEN Yiyu
    2023, 21(3):  22-28.  doi:10.12142/ZTECOM.202303004
    Asbtract ( )   HTML ( )   PDF (456KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    At present, the parameters of radar detection rely heavily on manual adjustment and empirical knowledge, resulting in low automation. Traditional manual adjustment methods cannot meet the requirements of modern radars for high efficiency, high precision, and high automation. Therefore, it is necessary to explore a new intelligent radar control learning framework and technology to improve the capability and automation of radar detection. Reinforcement learning is popular in decision task learning, but the shortage of samples in radar control tasks makes it difficult to meet the requirements of reinforcement learning. To address the above issues, we propose a practical radar operation reinforcement learning framework, and integrate offline reinforcement learning and meta-reinforcement learning methods to alleviate the sample requirements of reinforcement learning. Experimental results show that our method can automatically perform as humans in radar detection with real-world settings, thereby promoting the practical application of reinforcement learning in radar operation.

    Boundary Data Augmentation for Offline Reinforcement Learning
    SHEN Jiahao, JIANG Ke, TAN Xiaoyang
    2023, 21(3):  29-36.  doi:10.12142/ZTECOM.202303005
    Asbtract ( )   HTML ( )   PDF (1865KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    Offline reinforcement learning (ORL) aims to learn a rational agent purely from behavior data without any online interaction. One of the major challenges encountered in ORL is the problem of distribution shift, i.e., the mismatch between the knowledge of the learned policy and the reality of the underlying environment. Recent works usually handle this in a too pessimistic manner to avoid out-of-distribution (OOD) queries as much as possible, but this can influence the robustness of the agents at unseen states. In this paper, we propose a simple but effective method to address this issue. The key idea of our method is to enhance the robustness of the new policy learned offline by weakening its confidence in highly uncertain regions, and we propose to find those regions by simulating them with modified Generative Adversarial Nets (GAN) such that the generated data not only follow the same distribution with the old experience but are very difficult to deal with by themselves, with regard to the behavior policy or some other reference policy. We then use this information to regularize the ORL algorithm to penalize the overconfidence behavior in these regions. Extensive experiments on several publicly available offline RL benchmarks demonstrate the feasibility and effectiveness of the proposed method.

    Research Papers
    Differential Quasi-Yagi Antenna and Array
    ZHU Zhihao, ZHANG Yueping
    2023, 21(3):  37-44.  doi:10.12142/ZTECOM.202303006
    Asbtract ( )   HTML ( )   PDF (2723KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    A novel differential quasi-Yagi antenna is first presented and compared with a normal single-ended counterpart. The simulated and measured results show that the differential quasi-Yagi antenna outperforms the conventional single-ended one. The differential quasi-Yagi antenna is then used as an element for linear arrays. A study of the coupling mechanism between the two differential and the two single-ended quasi-Yagi antennas is conducted, which reveals that the TE0 mode is the dominant mode, and the driver is the decisive part to account for the mutual coupling. Next, the effects of four decoupling structures are respectively evaluated between the two differential quasi-Yagi antennas. Finally, the arrays with simple but effective decoupling structures are fabricated and measured. The measured results demonstrate that the simple slit or air-hole decoupling structure can reduce the coupling level from -18 dB to -25 dB and meanwhile maintain the impedance matching and radiation patterns of the array over the broad bandwidth. The differential quasi-Yagi antenna should be a promising antenna candidate for many applications.

    Massive Unsourced Random Access Under Carrier Frequency Offset
    XIE Xinyu, WU Yongpeng, YUAN Zhifeng, MA Yihua
    2023, 21(3):  45-53.  doi:10.12142/ZTECOM.202303007
    Asbtract ( )   HTML ( )   PDF (1438KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    Unsourced random access (URA) is a new perspective of massive access which aims at supporting numerous machine-type users. With the appearance of carrier frequency offset (CFO), joint activity detection and channel estimation, which is vital for multiple-input and multiple-output URA, is a challenging task. To handle the phase corruption of channel measurements under CFO, a novel compressed sensing algorithm is proposed, leveraging the parametric bilinear generalized approximate message passing framework with a Markov chain support model that captures the block sparsity structure of the considered angular domain channel. An uncoupled transmission scheme is proposed to reduce system complexity, where slot-emitted messages are reorganized relying on clustering unique user channels. Simulation results reveal that the proposed transmission design for URA under CFO outperforms other potential methods.

    Learning-Based Admission Control for Low-Earth-Orbit Satellite Communication Networks
    CHENG Lei, QIN Shuang, FENG Gang
    2023, 21(3):  54-62.  doi:10.12142/ZTECOM.202303008
    Asbtract ( )   HTML ( )   PDF (1188KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    Satellite communications has been regarded as an indispensable technology for future mobile networks to provide extremely high data rates, ultra-reliability, and ubiquitous coverage. However, the high dynamics caused by the fast movement of low-earth-orbit (LEO) satellites bring huge challenges in designing and optimizing satellite communication systems. Especially, admission control, deciding which users with diversified service requirements are allowed to access the network with limited resources, is of paramount importance to improve network resource utilization and meet the service quality requirements of users. In this paper, we propose a dynamic channel reservation strategy based on the Actor-Critic algorithm (AC-DCRS) to perform intelligent admission control in satellite networks. By carefully designing the long-term reward function and dynamically adjusting the reserved channel threshold, AC-DCRS reaches a long-run optimal access policy for both new calls and handover calls with different service priorities. Numerical results show that our proposed AC-DCRS outperforms traditional channel reservation strategies in terms of overall access failure probability, the average call success rate, and channel utilization under various dynamic traffic conditions.

    A 220 GHz Frequency-Division Multiplexing Wireless Link with High Data Rate
    ZHANG Bo, WANG Yihui, FENG Yinian, YANG Yonghui, PENG Lin
    2023, 21(3):  63-69.  doi:10.12142/ZTECOM.202303009
    Asbtract ( )   HTML ( )   PDF (2704KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    With the development of wireless communication, the 6G mobile communication technology has received wide attention. As one of the key technologies of 6G, terahertz (THz) communication technology has the characteristics of ultra-high bandwidth, high security and low environmental noise. In this paper, a THz duplexer with a half-wavelength coupling structure and a sub-harmonic mixer operating at 216 GHz and 204 GHz are designed and measured. Based on these key devices, a 220 GHz frequency-division multiplexing communication system is proposed, with a real-time data rate of 10.4 Gbit/s for one channel and a transmission distance of 15 m. The measured constellation diagram of two receivers is clearly visible, the signal-to-noise ratio (SNR) is higher than 22 dB, and the bit error ratio (BER) is less than 10-8. Furthermore, the high definition (HD) 4K video can also be transmitted in real time without stutter.

    Log Anomaly Detection Through GPT-2 for Large Scale Systems
    JI Yuhe, HAN Jing, ZHAO Yongxin, ZHANG Shenglin, GONG Zican
    2023, 21(3):  70-76.  doi:10.12142/ZTECOM.202303010
    Asbtract ( )   HTML ( )   PDF (537KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    As the scale of software systems expands, maintaining their stable operation has become an extraordinary challenge. System logs are semi-structured text generated by the recording function in the source code and have important research significance in software service anomaly detection. Existing log anomaly detection methods mainly focus on the statistical characteristics of logs, making it difficult to distinguish the semantic differences between normal and abnormal logs, and performing poorly on real-world industrial log data. In this paper, we propose an unsupervised framework for log anomaly detection based on generative pre-training-2 (GPT-2). We apply our approach to two industrial systems. The experimental results on two datasets show that our approach outperforms state-of-the-art approaches for log anomaly detection.

    Robust Beamforming Under Channel Prediction Errors for Time-Varying MIMO System
    ZHU Yuting, LI Zeng, ZHANG Hongtao
    2023, 21(3):  77-85.  doi:10.12142/ZTECOM.202303011
    Asbtract ( )   HTML ( )   PDF (1578KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    The accuracy of acquired channel state information (CSI) for beamforming design is essential for achievable performance in multiple-input multiple-output (MIMO) systems. However, in a high-speed moving scene with time-division duplex (TDD) mode, the acquired CSI depending on the channel reciprocity is inevitably outdated, leading to outdated beamforming design and then performance degradation. In this paper, a robust beamforming design under channel prediction errors is proposed for a time-varying MIMO system to combat the degradation further, based on the channel prediction technique. Specifically, the statistical characteristics of historical channel prediction errors are exploited and modeled. Moreover, to deal with random error terms, deterministic equivalents are adopted to further explore potential beamforming gain through the statistical information and ultimately derive the robust design aiming at maximizing weighted sum-rate performance. Simulation results show that the proposed beamforming design can maintain outperformance during the downlink transmission time even when channels vary fast, compared with the traditional beamforming design.

    Design of Raptor-Like LDPC Codes and High Throughput Decoder Towards 100 Gbit/s Throughput
    LI Hanwen, BI Ningjing, SHA Jin
    2023, 21(3):  86-92.  doi:10.12142/ZTECOM.202303012
    Asbtract ( )   HTML ( )   PDF (2033KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    This paper proposes a raptor-like low-density parity-check (RL-LDPC) code design together with the corresponding decoder hardware architecture aiming at next-generation mobile communication. A new kind of protograph different from the 5G new radio (NR) LDPC basic matrix is presented, and a code construction algorithm is proposed to improve the error-correcting performance. A multi-core layered decoder architecture that supports up to 100 Gbit/s throughput is designed based on the special protograph structure.

    Hybrid Architecture and Beamforming Optimization for Millimeter Wave Systems
    TANG Yuanqi, ZHANG Huimin, ZHENG Zheng, LI Ping, ZHU Yu
    2023, 21(3):  93-104.  doi:10.12142/ZTECOM.202303013
    Asbtract ( )   HTML ( )   PDF (4586KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    Hybrid beamforming (HBF) has become an attractive and important technology in massive multiple-input multiple-output (MIMO) millimeter-wave (mmWave) systems. There are different hybrid architectures in HBF depending on different connection strategies of the phase shifter network between antennas and radio frequency chains. This paper investigates HBF optimization with different hybrid architectures in broadband point-to-point mmWave MIMO systems. The joint hybrid architecture and beamforming optimization problem is divided into two sub-problems. First, we transform the spectral efficiency maximization problem into an equivalent weighted mean squared error minimization problem, and propose an algorithm based on the manifold optimization method for the hybrid beamformer with a fixed hybrid architecture. The overlapped subarray architecture which balances well between hardware costs and system performance is investigated. We further propose an algorithm to dynamically partition antenna subarrays and combine it with the HBF optimization algorithm. Simulation results are presented to demonstrate the performance improvement of our proposed algorithms.

    Simulation and Modeling of Common Mode EMI Noise in Planar Transformers
    LI Wei, JI Jingkang, LIU Yuanlong, SUN Jiawei, LIN Subin
    2023, 21(3):  105-116.  doi:10.12142/ZTECOM.202303014
    Asbtract ( )   HTML ( )   PDF (3135KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    The transformer is the key circuit component of the common-mode noise current when an isolated converter is working. The high-frequency characteristics of the transformer have an important influence on the common-mode noise of the converter. Traditionally, the measurement method is used for transformer modeling, and a single lumped device is used to establish the transformer model, which cannot be predicted in the transformer design stage. Based on the transformer common-mode noise transmission mechanism, this paper derives the transformer common-mode equivalent capacitance under ideal conditions. According to the principle of experimental measurement of the network analyzer, the electromagnetic field finite element simulation software three-dimensional (3D) modeling and simulation method is used to obtain the two-port parameters of the transformer, extract the high-frequency parameters of the transformer, and establish its electromagnetic compatibility equivalent circuit model. Finally, an experimental prototype is used to verify the correctness of the model by comparing the experimental measurement results with the simulation prediction results.

    Statistical Model of Path Loss for Railway 5G Marshalling Yard Scenario
    DING Jianwen, LIU Yao, LIAO Hongjian, SUN Bin, WANG Wei
    2023, 21(3):  117-122.  doi:10.12142/ZTECOM.202303015
    Asbtract ( )   HTML ( )   PDF (1448KB) ( )  
    Figures and Tables | References | Related Articles | Metrics

    The railway mobile communication system is undergoing a smooth transition from the Global System for Mobile Communications-Railway (GSM-R) to the Railway 5G. In this paper, an empirical path loss model based on a large amount of measured data is established to predict the path loss in the Railway 5G marshalling yard scenario. According to the different characteristics of base station directional antennas, the antenna gain is verified. Then we propose the position of the breakpoint in the antenna propagation area, and based on the breakpoint segmentation, a large-scale statistical model for marshalling yards is established.