Dynamic Wireless sensor networks (DWSNs) as an important means of industrial data collection are a key part of IIoT, where security and reliability are important characteristics of trustworthiness. However, due to dynamics, the security of key management is caused by a nontrusted base station (BS) that is easily targeted. The distribution key management schemes that avianize the role of BS also lead to additional and heavy overhead on sensors. To tackle these issues, in this paper, we propose a blockchain-based secure key management scheme (BC-EKM). First, stake blockchain is constructed based on the hybrid sensor network. In addition, we design a secure cluster formation algorithm and a secure node movement algorithm to implement key management, where stake blockchain as a trust machine replaces the majority functions of the BS. Finally, we conduct security analysis and ample simulations. The results indicate that the BC-EKM scheme is effective and efficient.

The mobile crowdsensing technology with a large number of IoT devices provides an economic and efficient solution to participation in coordinated large-scale sensing tasks. Edge computing powers mobile crowdsensing to form the mobile edge crowdsensing (MECS) framework. Privacy disclosure of sensing data in multiple stages is a significant challenge in the MECS. To tackle this issue, combining machine learning with game theory, we propose an AI-enabled three-party game (ATG) framework for guaranteed data privacy in the MECS. Specifically, based on the random forest classifier and the k-anonymity algorithm, we propose a classification-anonymity model that effectively guarantees the privacy of sensitive data. Moreover, we construct a three-party game model for analyzing the data privacy leakage in different phases in the MECS. Finally, we conduct numerical and theoretical analysis and ample simulations. The results indicate that the ATG framework is effective and efficient, and better suited to the MECS of IoT.

Metaverse is a vast virtual environment parallel to the physical world in which users enjoy a variety of services acting as an avatar. To build a secure living habitat, it’s vital to ensure the virtual-physical traceability that tracking a malicious player in the physical world via his avatars in virtual space. In this paper, we propose a two-factor authentication framework based on biometric-based authentication and chameleon signature. First, aiming at disguise in virtual space, we design an avatar’s two-factor identity model to ensure the verifiability of avatar’s virtual identity and physical identity. Second, facing at authentication efficiency and keys holding cost, we propose a chameleon collision signature algorithm to efficiently ensure that the avatar’s virtual identity is associated with its physical identity. Finally, aiming at impersonation in the physical world, we design two decentralized authentication protocols based on the avatar’s identity model and the chameleon collision signature to achieve real-time authentication on the avatar’s identity. Security analysis indicates that the proposed authentication framework guarantees the consistency and traceability of the avatar’s identity. Simulation experiments show that the framework not only completes the decentralized authentication between avatars but also achieves virtual-physical tracking.

Federated learning of deep neural networks has emerged as an evolving paradigm for distributed machine learning, gaining widespread attention due to its ability to update parameters without collecting raw data from users, especially in digital healthcare applications. However, the traditional centralized architecture of federated learning suffers from several problems (e.g., single point of failure, communication bottlenecks, etc.), especially malicious servers inferring gradients and causing gradient leakage. To tackle the above issues, we propose a robust and privacy-preserving decentralized deep federated learning (RPDFL) training scheme. Specifically, we design a novel ring FL structure and a Ring-Allreduce-based data sharing scheme to improve the communication efficiency in RPDFL training. Furthermore, we improve the process of distributing parameters of the Chinese residual theorem to update the execution process of the threshold secret sharing, supporting healthcare edge to drop out during the training process without causing data leakage, and ensuring the robustness of the RPDFL training under the Ring-Allreduce-based data sharing scheme. Security analysis indicates that RPDFL is provable secure. Experiment results show that RPDFL is significantly superior to standard FL methods in terms of model accuracy and convergence, and is suitable for digital healthcare applications.

Edge services provide an effective and superior means of real-time transmissions and rapid processing of information in the Industrial Internet of Things (IIoT). However, the continuous increase of the number of smart devices results in privacy leakage and insufficient model accuracy of edge services. To tackle these challenges, in this article, we propose a blockchain-based machine learning framework for edge services (BML-ES) in IIoT. Specifically, we construct novel smart contracts to encourage multiparty participation of edge services to improve the efficiency of data processing. Moreover, we propose an aggregation strategy to verify and aggregate model parameters to ensure the accuracy of decision tree models. Finally, based on the SM2 public key cryptosystem, we protect data security and prevent data privacy leakage in edge services. Theoretical analysis and simulation experiments indicate that the BML-ES framework is secure, effective, and efficient, and is better suitable to improve the accuracy of edge services in IIoT.