Protecting Privacy in Community-Driven Sensing with Secure Aggregation

Abstract

Mobile  Crowd  Sensing  (MCS)  leverages  ubiquitous  mobile  devices  to  collect  large-scale environmental,  health,  and  urban  data.  While  enabling  powerful  analytics,  MCS  faces significant  privacy  risks as  user-contributed  data  often  contains  sensitive  personal information such as location, health status, or behavior patterns. To ensure both utility and confidentiality, this work explores the integration of Differential Privacy (DP) and Secure Aggregation mechanisms for MCS platforms. We analyze privacy–utility trade-offs, present algorithms for local and global differential privacy, and design lightweight secure aggregation protocols  suitable  for  resource-constrained  mobile  devices.  The  proposed  framework ensures individual-level privacy, defends against inference attacks, and maintains high data utility for real-time analytics.
Index Terms