Location Privacy-Preserving Task Offloading in LEO Satellite Mobile Edge Computing
TL;DR
This work proposes a location privacy-aware task offloading scheme, and develops a privacy-preserving Lyapunov Transformer-based Soft Actor–Critic framework to solve the energy efficiency optimization problem under location privacy constraint in this system.
Abstract
Low Earth Orbit (LEO) satellite mobile edge computing system enables the flexible and wide-area task offloading from ground users to nearby satellites, but the accumulated signal observations at different satellites pose a severe threat to user location privacy. Based on a proper design of satellite selection policy and task offloading decision to restrain the accumulation of signal observations across satellites and time slots, we first propose a location privacy-aware task offloading scheme for the concerned system. We then apply Fisher Information Matrix (FIM) to conduct theoretical modeling for localization performance and system energy efficiency under the proposed scheme. We further explore the energy efficiency optimization problem under location privacy constraint in this system, and develop a privacy-preserving Lyapunov Transformer-based Soft Actor–Critic (PLTSAC) framework to solve this problem. Finally, extensive numerical results are provided to demonstrate the efficiency the proposed scheme in ensuring both user location privacy and system energy efficiency.
