ZGZhenzhen Gao

Zhenzhen Gao

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Zhenzhen Gao has a total of 48 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Neural-Assisted Detector with Low False-Alarm Rate for Dense Multi-Target Sensing in OTFS-ISAC Systems

    by Imran Younas, Zhenzhen Gao, Kejia Ma

    Dense multi-target sensing in orthogonal time–frequency space (OTFS)-based integrated sensing and communication (ISAC) is challenged by sub-resolution target spacing, unknown target cardinality, and inter-target radar-cross-section (RCS) variation. This paper develops a neural-assisted detector (NAD) that combines convolutional-neural-network candidate generation with continuous delay–Doppler refinement, physics-based validation, active splitting, and explicit model-order control. The CNN generates candidates from the nominal delay–Doppler grid while predicting fractional offsets, so that the final target positions are continuous-valued rather than confined to grid centers. Isolated candidates are refined individually, whereas closely spaced candidates are jointly optimized through alternating position and complex-gain updates, with active splitting used to resolve dense target pairs. For unknown target number, NAD-Hybrid combines conservative candidate generation with information-criterion-based weak-target rescue and backward pruning to control false alarms under RCS mismatch. On a frozen 8000-frame blind test covering two scenes and four RCS standard deviations, NAD-Hybrid achieves a macro-averaged F1 score of 0.9338 and reduces false alarms relative to the unknown-order OMPFR-BIC baseline, at the cost of lower detection probability under severe RCS mismatch. The proposed detector therefore provides a conservative operating point that prioritizes low false alarms and reliable target-count estimation. The sensing framework is further demonstrated to coexist with an index-modulated OTFS communication link using a shared frame.

  • A DRSN-Based Wireless Channel Propagation Scenario Identification

    by Yunzhe Zhu, Xuewen Liao, Ang Li, Zhenzhen Gao

    Channel scenario identification plays an important auxiliary role in wireless channel modeling as well as in the design of physical layer and network layer algorithms for wireless communications. Traditional approaches rely on exhaustive channel parameter estimation—including path-specific delays, angles and Doppler shift—a process demanding prohibitive computational resources. To enable real-time and accurate identification of the current propagation scenario, this letter proposes a scenario identification method based on least squares channel estimation and a deep residual shrinkage network (LSE-DRSN). The channel impulse response (CIR) is obtained via the low-complexity least squares (LS) method, while the deep residual shrinkage network (DRSN) ensures the scenario identification accuracy in propagation environments where significant estimation errors exist in LS-based CIR acquisition. The simulation results demonstrate that the identification accuracy of the proposed method outperforms existing scenario identification methods, especially under low signal-to-noise ratio (SNR) conditions.


  • Distributed Noncoherent Joint Transmission Based on Multi-Agent Reinforcement Learning for Dense Small Cell MISO Systems

    by Shaozhuang Bai, Zhenzhen Gao, Jinchi Wang, Xuewen Liao

    This paper investigates the sum rate maximization problem in dense small cell (DSC) networks, where multi-antenna small cell base stations (SBSs) serve single-antenna users over a shared frequency band using noncoherent joint transmission (JT). While noncoherent JT enhances capacity by enabling users to receive signals from multiple coordinated SBSs, its practical deployment is hindered by the nonconvex and NP-hard nature of the optimization problem, along with the high-dimensional continuous optimization variables. Existing optimization-based approaches, despite achieving near-optimal performance, require global channel state information (CSI) and multiple iterations, making them impractical for DSC networks. To overcome these challenges, we first prove that the optimal beamforming structure is the same for both the power minimization problem and the sum rate maximization problem, and then mathematically derive the optimal beamforming structure for both problems by solving the power minimization problem. The optimal beamforming structure can effectively reduces the variable dimensions. By exploiting the optimal beamforming structure, we propose a deep deterministic policy gradient-based distributed noncoherent JT scheme to maximize the system sum rate. In the proposed scheme, each SBS utilizes global information for training and uses local CSI to determine beamforming vectors. Simulation results demonstrate that the proposed scheme achieves comparable performance with considerably lower computational complexity and information overhead compared to centralized iterative optimization-based techniques, making it more attractive for practical deployment.


  • Multi-Agent Reinforcement Learning Based Distributed Multi-User Scheduling and Beamforming Design in Multi-Cell Systems

    by Shaozhuang Bai, Zhenzhen Gao, Xuewen Liao

    In multi-cell systems, joint multi-user scheduling and beamforming (MUS-BF) design are two fundamental problems that are usually studied separately in the existing literature. In this work, we focus on the joint MUS-BF design with the goal of minimizing the network experience delay. Since the objective function is long-term and non-convex, we first equate the network experience delay with the long-term average network experience delay cost, which can be optimized by reinforcement learning techniques. Then, we propose a distributed MUS-BF scheme based on multi-agent reinforcement learning to minimize the long-term average network experience delay cost, where the adaptive multi-deep Q network (DQN) architecture is designed at each base station (BS) to find suitable MUS-BF with low computational complexity. Each BS is able to train its own multi-DQN and implement appropriate MUS-BF based on the local information. Simulation results validate that the adaptive multi-DQN architecture decreases 70% computational complexity of the proposed algorithm without compromising performance compared to the non-adaptive design. Compared to existing schemes, the proposed scheme achieves superior performance with the lowest information overhead, making it a more appealing solution for real-time MUS-BF in delay-sensitive scenarios.


  • Open Access

    Game-Based Intelligent Jamming Strategy without Prior Information in Wireless Communications

    by Yongcheng Li, Jinchi Wang, Zhenzhen Gao, Gangming Lv

    Traditional jamming technologies have become less effective with the development of anti-jamming technologies, especially with the appearance of intelligent transmitters, which can adaptively adjust their transmission strategies. To deal with intelligent transmitters, in this paper, a game-based intelligent jamming scheme is proposed. Considering that the intelligent transmitter has multiple transmission strategy sets whose prior probabilities are unknown to the jammer, we first model the interaction between the transmitter and the jammer as a dynamic game with incomplete information. Then the perfect Bayesian equilibrium is derived based on assumptions of some prior information. For more practical applications when no prior information about the transmitter is available at the jammer, a Q-learning-based method is proposed to find an intelligent jamming strategy by exploiting the sensing results of the wireless communications. The design of the jammer’s reward function is guided by the game utility and the reward is calculated based on the Acknowledgement/Negative Acknowledgement feedback of the receiver. Simulation results show that the proposed scheme has only 0.5% loss in jamming utility compared to that of the perfect Bayesian equilibrium strategy. Compared to existing jamming schemes, a higher packet error rate can be achieved by the proposed scheme by consuming less jamming power.


  • WiADN: Asymmetrical Dual-Task Attention Network for WiFi Sensing

    by Jinggan Zhou, Xuewen Liao, Zhenzhen Gao, Qiao Li, Chunlei Zheng

    WiFi-sensing systems that can accomplish multiple relative tasks simultaneously are attracting significant attention due to their potential for large-scale commercial applications. However, different WiFi sensing scenarios may often rely on various task-specific features, posing a challenge in balancing these different, or asymmetrical, characteristics across tasks. In this article, we propose a system called WiADN which aims to address the asymmetrical problems in the joint recognition of users’ locations and activities. Our system is composed of two critical parts: 1) an asymmetrical network architecture and 2) an adaptive weight loss (AWL) module employed during the training phase. First, we define activity recognition as a high-level task and location recognition as a low-level task based on their respective difficulty levels. Then, the proposed architecture leverages the cascading attention-based modules to transfer the prior knowledge between different tasks. The key insight of the proposed architecture is to mimic the skilled learners in similar situations, who often tackle easier problems first to enable them to solve more challenging problems later on. Based on this behavioral strategy, the proposed attention-based modules are designed to generate masks to select specific characteristics from the low-level task to help the high-level task to learn respective features more effectively. Additionally, the AWL module based on the task uncertainty theory is employed to balance two tasks’ asymmetry from the perspective of loss optimization. Furthermore, extensive experiment results based on two open data sets demonstrate the superiority of our system in the accuracy with other state-of-the-art methods. At last, the effectiveness and the robustness of our system are also verified through the comparative studies and ablation experiments. Our source codes are available at https://github.com/jzhoujg/WiADN .


  • Open Access

    Intelligent Jamming Strategy for Wireless Communications Based on Game Theory

    by Yongcheng Li, Weijian Miao, Zhenzhen Gao, Gangming Lv

    This paper explores the scenario that a jammer attacks an intelligent transmitter which can sense and adapt to the jamming environment. Given the non-cooperative relationship between the transmitter and the jammer, two main challenges are addressed in this paper: how to model their interactions and how to devise a jamming strategy without prior knowledge of the transmitter. A non-zero-sum game is used to model and analyze such non-cooperative interactions. An approximate mixed-strategy Nash equilibrium (NE) under complete information is derived to serve as a benchmark for comparison. According to the non-zero-sum game model, a Deep Q-Network (DQN) approach is proposed to determine jamming strategies by exploiting the detection results of the legitimate signals, such as Acknowledgements (ACKs) feedback and the modulation recognition results obtained by the jammer. Simulation results demonstrate that, without requiring complete information about the transmitter, the proposed DQN approach can achieve a similar utility as the benchmark strategy using complete information. Compared to other learning-based jamming schemes and random jamming strategy, the proposed DQN approach achieves a higher packet error rate for the communication transceiver with reduced jamming power consumption.


  • Joint Caching and Transmission Design for Delay Minimization in Small Cell Networks

    by Shaozhuang Bai, Zhenzhen Gao, Xuewen Liao

    Caching and coordinated multi-point noncoherent joint transmission (JT) has been considered on top of small cell networks (SCNs) to greatly decrease network delay. Unlike existing studies that focus solely on optimizing caching based on specific transmission policies, we investigate the joint design of caching and JT strategies by considering their interactions. Due to the nonconvex nature of the joint optimization problem, we decouple it into a caching optimization subproblem and a transmission optimization subproblem. In SCNs, SBSs are selectively activated according to the transmission strategy to efficiently manage interference and enhance signal quality. In solving the caching optimization subproblem, the activation probabilities of all SBSs are modeled and a closed-form suboptimal caching algorithm is derived based the activation probabilities of all SBSs. Based on the closed-form caching algorithm, a distributed transmission scheme is designed by using the multi-agent reinforcement learning (MARL) method. Simulation results demonstrate that the closed-form caching algorithm achieves nearly optimal performance with significantly reduced complexity. Compared to heuristic schemes, optimization schemes, and MARL schemes, the proposed joint scheme achieves at least 80% reduction of the network delay.

  • Open Access

    Learning-Based Multi-Domain Anti-Jamming Communication with Unknown Information

    by Yongcheng Li, Jinchi Wang, Zhenzhen Gao

    Due to the open nature of the wireless channel, wireless networks are vulnerable to jamming attacks. In this paper, we try to solve the anti-jamming problem caused by smart jammers, which can adaptively adjust the jamming channel and the jamming power. The interaction between the legitimate transmitter and the jammers is modeled as a non-zero-sum game. Considering that it is challenging for the transmitter and the jammers to acquire each other’s information, we propose two anti-jamming communication schemes based on the Deep Q-Network (DQN) algorithm and hierarchical learning (HL) algorithm to solve the non-zero-sum game. Specifically, the DQN-based scheme aims to solve the anti-jamming strategies in the frequency domain and the power domain directly, while the HL-based scheme tries to find the optimal mixed strategies for the Nash equilibrium. Simulation results are presented to validate the effectiveness of the proposed schemes. It is shown that the HL-based scheme has a better convergence performance and the DQN-based scheme has a higher converged utility of the transmitter. In the case of a single jammer, the DQN-based scheme achieves 80% of the transmitter’s utility of the no-jamming case, while the HL-based scheme achieves 63%.


  • Open Access

    Physical Layer Security Design for FDD IM-OTFS Transmissions Based on Secure Mapping

    by Kejia Ma, Zhenzhen Gao, Jinchi Wang, Linling Cheng

    In this paper, a physical layer security scheme is proposed for index modulation aided orthogonal time frequency space (IM-OTFS) modulation systems working in Frequency Division Duplex (FDD) mode. In FDD systems, the channel reciprocity, which is the basic assumption for most physical layer security (PLS) techniques, is not applicable. To deal with the security challenge in FDD IM-OTFS systems, Chaos sequence is generated by exploiting the angular reciprocity of the legitimate link to construct the secure mapping rule between the information bits and the modulation symbols as well as the symbol indices. A PLS scheme based on the proposed secure mapping is introduced. The secrecy performance of the proposed PLS scheme is analyzed in terms of Bit Error Rate (BER) and the ergodic secrecy rate. Furthermore, a close-form approximated ergodic secrecy rate is derived to reduce the computational complexity. A truncation method is proposed to obtain robust secure transmissions when the angles of the uplink and downlink legitimate channels are estimated with errors. Simulation results show that the theoretical BER curve approaches the simulated BER curve closely for medium and high signal to noise ratio (SNR) regions, and the eavesdropper’s BER is always around 0.5 for all SNRs. The closed-form approximated ergodic secrecy rate fits the ergodic secrecy rate exactly in low and high SNR regions, and can work as a lower bound of the ergodic secrecy rate in medium SNR region. The secure mapping at the legitimate users is not affected by the angle estimation errors by using the truncation method, while the eavesdropper’s BER is influenced and the influence is analyzed.