YXYuanjun Xia

Yuanjun Xia

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Yuanjun Xia has a total of 17 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Verifiable and Robust Privacy-Preserving Multidimensional Truth Discovery for IoT Crowdsensing

    by Shuang Chen, Jingxue Chen, Yuanjun Xia, Yangfan Liang, You Li, Yining Liu

    The rapid proliferation of IoT devices has popularized crowdsensing for distributed data collection, where Truth Discovery plays a critical role in inferring reliable information from heterogeneous observations. However, existing privacy-preserving truth discovery schemes face challenges including inefficient verifiability, limited weighting strategies, insufficient robustness, and excessive overhead. In this paper, we introduce VRPMTD, a verifiable, robust, and privacy-preserving multi-dimensional truth discovery framework. We achieve scalable verifiability via CRT based packing of multidimensional measurements together with commitments and a linear homomorphic hash. This design allows the data requester to batch verify aggregated results. To improve accuracy, we design a novel weighting mechanism using a Gaussian radial basis function residual and a sliding-window temporal loss, allowing workers’ weights to reflect both long-term reliability and recent behavior. Additionally, the framework improves robustness under realistic sensing and network failures. To optimize efficiency, we implement a lightweight dual-layer encryption mechanism and a difference-based uploading strategy. Formal security analysis indicates that VRPMTD preserves the input-level confidentiality of workers’ raw measurements and ensures verifiability of outsourced aggregation. Extensive experiments on real-world datasets and IoT devices demonstrate that VRPMTD achieves higher accuracy while incurring lower overhead.

  • Fully Anonymous Broadcast Signcryption for Secure Health Data Transmission in WBANs

    by Yangfan Liang, Gao Liu, Xianchao Zhang, Yiming Chen, Jingxue Chen, Yuanjun Xia, Yining Liu

    To support comprehensive and personalized healthcare delivery, sensitive personal health data must often be transmitted from wearable devices to multiple designated specialists via wireless body area networks. Although existing identity-based broadcast signcryption schemes provide secure one-to-many transmission, they typically fail to preserve recipient anonymity. In particular, each recipient is usually required to know the identities of all other recipients to unsigncrypt the message, which undermines privacy and enables inference attacks, especially problematic in telemedicine applications. In this paper, we propose a novel broadcast signcryption scheme that simultaneously achieves recipient anonymity, sender anonymity, and resilience to physical attacks, which are not realized in prior work. Our key innovation lies in an anonymity-preserving unsigncryption mechanism based on inner product, which enables each recipient to independently verify and decrypt the broadcast message without knowledge of others’ identities. In addition, we incorporate a physical unclonable function with a fuzzy extractor to resist physical attacks in the presence of noise. Furthermore, a comprehensive security analysis demonstrates that our scheme satisfies essential security properties, including authentication, integrity, confidentiality, and full anonymity. It is resilient against common attacks such as impersonation, modification, replay, and man-in-the-middle attacks. Finally, performance evaluations confirm the practical feasibility of our scheme.


  • TEVA: Training-Efficient and Verifiable Aggregation for Federated Learning for Consumer Electronics in Industry 5.0

    by Yuanjun Xia, Yining Liu, Jingxue Chen, Yangfan Liang, Fazlullah Khan, Ryan Alturki, Xiaopei Wang

    Federated learning (FL) has been widely used for privacy-preserving model updates in Industry 5.0, facilitated by 6G networks. Despite FL’s privacy-preserving advantages, it remains vulnerable to attacks where adversaries can infer private data from local models or manipulate the central server (CS) to deliver falsified global models. Current privacy-preserving approaches, primarily based on the FedAvg algorithm, fail to optimize training efficiency for non-independent and identically distributed (non-IID) data. This article proposes training-efficient and verifiable aggregation (TEVA) for FL to resolve these issues. This scheme combines threshold Paillier homomorphic encryption (TPHE), verifiable aggregation, and an optimized double momentum update mechanism (OdMum). TEVA not only leverages TPHE to protect the privacy of local models but also ensures the integrity of the global model through a verifiable aggregation mechanism. Additionally, TEVA integrates the OdMum algorithm to effectively address the challenges posed by non-IID data, promoting rapid model convergence and significantly enhancing overall training efficiency. Security analysis indicates that TEVA meets the requirements for privacy protection. Extensive experimental results demonstrate that TEVA can accelerate model convergence while incurring lower computational and communication overheads.