YLYitong Li

Yitong Li

*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


Yitong Li has a total of 47 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Priority-Aware Resource Allocation in AoI-Oriented UL-OFDMA Wi-Fi Networks Based on Multiagent Reinforcement Learning

    by Pengxue Liu, Dalong Zhang, Fasong Wang, Shijie Shi, Yanbin Zhang, Yitong Li

    The proliferation of time-sensitive Internet of Things (IoT) applications has significantly increased the demand for real-time communication in uplink orthogonal frequency division multiple access (UL-OFDMA) Wi-Fi networks. Despite extensive studies, how to meet the heterogeneous age of information (AoI) requirements across different stations (STAs) in Wi-Fi networks remains an open question. To tackle this issue, we propose a multi-agent reinforcement learning (MARL) resource allocation algorithm based on independent hybrid proximal policy optimization (IHPPO), aiming to minimize the AoI and power consumption for each STA while guaranteeing the heterogeneous AoI requirements among STAs within a resource-constrained environment. Specifically, the proposed strategy utilizes HPPO to directly optimize the original hybrid action space by combining discrete resource unit (RU) selection and continuous transmit power adjustment. Extensive simulations demonstrate the superior performance of the IHPPO mechanism in terms of convergence performance and the trade-off between AoI and power consumption, relative to the decomposed multi-agent deep deterministic policy gradient (DE-MADDPG) algorithm, the fully decentralized MADDPG (FD-MADDPG) algorithm, and the random method in different access scenarios.


  • EAS Anti-Interference Technique Based on Time-Frequency 2-D Randomized Waveform Design

    by Xin Xi, Yanbin Zhang, Rui Li, Fasong Wang, Shijie Shi, Yitong Li

    Electronic Article Surveillance (EAS) is an anti-theft method that utilizes electronic technology to effectively reduce merchandise theft in open shopping environments.In large supermarkets and similar venues, the presence of numerous EAS devices can increase the likelihood of mutual interference between devices, resulting in a higher rate of false alarms. Interference-resistant waveform design is one of the key techniques for combating interference in EAS devices. This article analyzes the principle of interference generation based on EAS, proposes a waveform design using time-frequency two-dimensional randomization for interference resistance, derives and establishes a probability model, and conducts the simulation verification and the test of hardware platform. The results indicate that the time-frequency two-dimensional randomization method can effectively reduce interference between EAS devices.

  • Joint SSK Assisted by Mirror Array Grouping and LEDs in Indoor VLC System

    by Yicong Sun, Fasong Wang, Rui Li, Shijie Shi, Yitong Li, Xingwang Li, Wei Xiang

    In the mirror array-assisted indoor visible light communication (VLC) system, activating all mirrors simultaneously can result in a complex system design and excessive consumption of resources. To address this issue and enhance the spectral efficiency, this paper proposes a joint space shift keying (SSK) technique, which is achieved by jointly utilizing the mirror array grouping and LED activation procedures in indoor VLC systems. By doing so, the SSK scheme utilizes the activation status of mirror array blocks to transmit information and can avoid excessive utilization of mirror resources while achieving efficient and reliable communication. The fairness of the considered system is improved by introducing a mirror array grouping technique that relies on the channel gains. Concurrently, a novel joint SSK modulation scheme is proposed, whereby the mirror array block is activated in conjunction with the LED. This scheme employs a information bit-based activation technique to enhance spectral efficiency. Additionally, the low complexity orthogonal matching pursuit (OMP) detection and the optimal maximum likelihood (ML) detection algorithms are employed. The effectiveness of the proposed schemes has been confirmed through extensive simulations.


  • A Retransmission Framework for Over-the-Air Computation Under Time-Varying Channel Fading

    by Shijie Shi, Guanzhong Lei, Fasong Wang, Yitong Li, Wei Ni, Abbas Jamalipour

    In computing-oriented communication scenarios, over-the-air computation (AirComp) directly computes results by leveraging the waveform superposition of wireless multiple access channels (MAC), eliminating the need to recover individual edge device (ED) signals. This approach enhances efficiency but faces challenges in maintaining low function distortion under deep fading. In this paper, we propose a new retransmission-based AirComp framework that incorporates a channel prediction model to mitigate function distortion. Our framework introduces a performance analysis and optimization strategy involving cross-timeslot receiver combining. The optimal transmission coefficients and denoising factors are derived to minimize computing mean squared error (MSE) and reduce sum-power consumption. We design and analyze AirComp retransmission schemes that account for channel state information (CSI) estimation errors in the initial timeslot. Simulations validate the effectiveness of our framework, which markedly decreases the MSE and power consumption, e.g., by about 61% and 55%, compared to existing benchmarks, demonstrating its superior noise resilience and significantly alleviated the impact of imperfect CSI. These results highlight the potential of our framework to enable efficient and reliable AirComp in time-varying fading channels, particularly for resource-constrained Internet of Things (IoT) environments.


  • Open Access

    Jailbreak attack of large language model based on scene construction

    by Zhicong Hu, Yitong Li, Zhiyong Shen

    The jailbreaking of large language models (LLMs) has recently garnered significant attention. Before releasing the model, extensive fine-tuning was conducted using RLHF (i.e., optimizing language models through reinforcement learning from human feedback) and other methods to ensure its behavior aligns with human values. However, even aligned LLMs can be maliciously manipulated, resulting in unintended behavior, termed "jailbreaking." In this study, inspired by manually crafted jailbreaking prompts, we introduce the Scenario Construction Attack (SC-Attack) and develop three seed templates for jailbreaking based on scenario construction, generating additional templates using the GPT-4o context. Additionally, to enhance the jailbreaking of aligned large models, we utilize Energy-constrained Decoding with Langevin Dynamics (COLD) to regulate the generation of adversarial suffixes for the attack. We define this combined attack as SC-COLD. Extensive experiments revealed that SC-Attack alone can achieve a 100% success rate on closed-source models like GPT-3.5 and GPT-4o, with the output content being more toxic. It also achieved a success rate of over 90% on aligned open-source models such as LLM (Llama2, Gemma). The extensive experiments demonstrate that SC-COLD possesses broad applicability, robust controllability, and a high success rate.


  • Novel Micro-UAV Identification Approach Assisted by Combined RF Fingerprint

    by Qilong Li, Fasong Wang, Yanbin Zhang, Rui Li, Shijie Shi, Yitong Li, Xingwang Li

    In response to the threat to societal security posed by unauthorized unmanned aerial vehicles (UAVs), there is an urgent need for a fast and accurate method of identifying UAVs. The detection and identification of UAVs based on Radio Frequency (RF) fingerprint features are considered crucial for addressing this issue. This paper presents a model for identifying micro-UAV targets based on RF fingerprint information. The system aims to effectively address the challenges of UAV safety management by providing reliable support for UAV control and management. In this model, based on the collected UAV RF signals in the environment, the six statistical RF features of instantaneous amplitude and fractal dimension in the time domain information are then extracted, resulting in a combined seven-dimensional RF fingerprints of the UAV. The seven-dimensional RF fingerprints are then reduced to three dimensions using the t-distributed Stochastic Neighbour Embedding (t-SNE) and Principal Component Analysis (PCA) methods. Finally, the reduced dimensional data is fed into three machine learning algorithms: Support Vector Machine (SVM), Weighted K-Nearest Neighbour (WKNN), and Random Forest (RandF). The identification results are as follows: in the self-constructed UAV RF dataset, the median identification accuracy using SVM, WKNN, and RandF is up to 98.06%, 99.64%, and 99.63%, respectively. In the open-source UAV RF dataset, the median identification accuracy for SVM, WKNN, and RandF is up to 95.33%, 98.91%, and 98.91%, respectively. All the above experimental results confirm the reliability and effectiveness of the proposed model described in this paper.


  • SM and NOMA Joint Assisted Indoor Multi-User VLC Downlink

    by Fasong Wang, Ting Zuo, Jiankang Zhang, Shijie Shi, Yitong Li

    In this paper, a novel spatial modulation (SM) assisted successive interference cancellation (SIC) free non-orthogonal multiple access (NOMA) scheme is proposed for indoor multi-user visible light communications (VLC) downlink. In the proposed scheme, all users are grouped according to channel gains. The data information of users in each group is then mapped to the spatial domain and constellation symbol domain according to their bit error rate requirements. In addition, due to the inherent sparsity of SM modulated signals, the compressed sensing (CS) sparse reconstruction algorithm is revoked for demodulating signals at the receiver. In this process, the information carried by the activated light emitting diode index is demodulated using the CS aided sparsity reconstruction algorithm, while the constellation symbol information is demodulated using the maximum likelihood (ML) algorithm. Additionally, by combining the greedy algorithm with the ML algorithm, a new method for sparse signal reconstruction detection is proposed. Compared with traditional NOMA technology, at the receiver side, both the error propagation caused by SIC and intra-group interference can be eliminated. Furthermore, the computational complexity of demodulation is reduced by utilizing the proposed joint signal detection procedure. The effectiveness of the proposed indoor multi-user SM NOMA VLC architecture is validated through Monte Carlo simulations.


  • Orthogonal Chirp Division Multiplexing Assisted Dual-Function Radar Communication in IoT Networks

    by Shaohua Li, Fasong Wang, Yanbin Zhang, Rui Li, Shijie Shi, Yitong Li, Xingwang Li, Daniel Benevides da Costa

    The dual-function radar communication (DFRC) system utilizes a hardware platform to achieve both radar and communication functions. In comparison to independently exploiting radar and communication systems, the DFRC system can significantly reduce system redundancy, volume, weight, and energy consumption. This makes DFRC an important area of research with practical value in advanced Internet of Things (IoT) techniques. This article explores an exemplary DFRC system based on the orthogonal chirp division multiplexing (OCDM) methodology. In this system, OCDM achieves chirp multiplexing through the Fresnel transform and is identified as a potential replacement for orthogonal frequency-division multiplexing in high-speed communication systems. In the proposed system, we integrate index modulation (IM) into the communication subsystem and utilize the subchirp index of OCDM to convey additional communication information, thereby significantly enhancing the communication rate of the DFRC system. Furthermore, a radar processing algorithm utilizing the OCDM signal is developed. This algorithm integrates the sparsity-aided compressed sensing (CS) algorithm into the radar subsystem to enhance estimation precision and reduce the sampling rate and hardware complexity of the radar receiver. Based on the proposed OCDM-assisted DFRC scheme, the communication rate of the DFRC system and the ambiguity function of OCDM are evaluated. The feasibility and effectiveness of the proposed DFRC system are confirmed through numerical calculations and simulation results.


  • Open Access

    Joint Client Selection and CPU Frequency Control in Wireless Federated Learning Networks with Power Constraints

    by Zhaohui Zhou, Shijie Shi, Fasong Wang, Yanbin Zhang, Yitong Li

    Federated learning (FL) represents a distributed machine learning approach that eliminates the necessity of transmitting privacy-sensitive local training samples. However, within wireless FL networks, resource heterogeneity introduces straggler clients, thereby decelerating the learning process. Additionally, the learning process is further slowed due to the non-independent and identically distributed (non-IID) nature of local training samples. Coupled with resource constraints during the learning process, there arises an imperative need for optimizing client selection and resource allocation strategies to mitigate these challenges. While numerous studies have made strides in this regard, few have considered the joint optimization of client selection and computational power (i.e., CPU frequency) for both clients and the edge server during each global iteration. In this paper, we initially define a cost function encompassing learning latency and non-IID characteristics. Subsequently, we pose a joint client selection and CPU frequency control problem that minimizes the time-averaged cost function subject to long-term power constraints. By utilizing Lyapunov optimization theory, the long-term optimization problem is transformed into a sequence of short-term problems. Finally, an algorithm is proposed to determine the optimal client selection decision and corresponding optimal CPU frequency for both the selected clients and the server. Theoretical analysis provides performance guarantees and our simulation results substantiate that our proposed algorithm outperforms comparative algorithms in terms of test accuracy while maintaining low power consumption.