DScPA: A Dynamic Subcluster Privacy-Preserving Aggregation Scheme for Mobile Crowdsourcing in Industrial IoT

Mobile crowdsourcing (MCS) is a promising new paradigm for intelligent data perception in large-scale sensor applications such as the Industrial Internet of Things (IIoT). This approach assigns industrial perception tasks to mobile devices for data collection and sharing, creating a bright outlook for building strong industrial systems and improving industrial services. However, the particular IIoT network environment is vulnerable to a range of malicious attacks, including the manipulation or deletion of data. Moreover, industry-aware nodes, which have limited energy resources, are susceptible to various failures that can result in distorted data and inaccurate trend analysis. To tackle the above issues, we propose a dynamic subcluster privacy-preserving aggregation (DScPA) scheme for the crowdsourced industrial virtual areas, by exploring the equilibrium between privacy security and data benefits for industrial users. Specifically, we propose joining and exiting the virtual area aggregation system protocol, and design low-cost privacy-preserving aggregation algorithm to achieve flexible and dynamic construction of virtual areas. Additionally, we propose a supervised mechanism-based subset aggregation protocol that takes into account the remaining energy of the nodes in the virtual areas aggregation system and their distance from the base station. We employ the relevant characteristics of polynomial functions and binary data to achieve data privacy protection and integrity verification at a lower computational cost. Furthermore, we develop an asymmetric information iterative static noncooperative game model to verify the soundness of DScPA. Finally, security analysis demonstrates that the DScPA scheme meets the security objectives. Simulations show that implementing DScPA in the industrial virtual area can improve the network lifetime by about 16.7% compared to existing solutions, while also increasing the average remaining energy of industry-aware nodes.