A Paillier Homomorphic Encryption-Based Lightweight Privacy Protection Model for Mobile Crowd Sensing Networks

The objective of this study is to enhance the privacy protection capability of mobile crowd perception networks and improve the detection accuracy of security threat data. To this end, a three-level collaborative distributed architecture is designed, which combines the concept of zero trust and proposes a lightweight security threat detection model based on the Paillier homomorphic encryption, deep neural networks, and box graph methods. Firstly, by optimizing the mobile crowd sensing networks framework and introducing three different network structures, a collaborative distributed system of terminal edge cloud was constructed. Secondly, the Paillier homomorphic encryption algorithm was designed to protect data privacy. The experimental results showed that the designed model achieved a detection accuracy of 98.84%, a mean square error of 0.03, and an average detection time of 0.15 seconds for A-class threats. In terms of processing efficiency, this model significantly improved data transmission efficiency, had lower computational overhead, and was suitable for various types of security threat detection. Therefore, the security threat detection model proposed in this study provides effective privacy protection technology for mobile crowd-sensing networks, significantly improving network security.