Mobility-Aware Offloading with SINR Prediction and Selective Replication in Fog-IoT Networks
The rise of mobile Internet of Things (IoT) applications has made reliable, low-latency task offloading to nearby fog nodes essential. However, user mobility, time-varying wireless links, short-packet transmission errors, and fog-queue congestion make the execution of delay-sensitive tasks challenging. Existing schemes typically address reliability, replication, or offloading separately, without jointly considering mobility-aware link variation. To address this gap, this paper proposes a mobility-aware reliability-driven offloading framework for fog-enabled IoT networks. The proposed scheme combines a lightweight multilayer perceptron (MLP)-based SINR predictor with a Lyapunov driftplus-penalty (LDPP) controller to select among local execution, single-fog offloading, and replicated-fog execution. Simulation results show that the proposed method reduces task delay, reliability violations, and energy consumption by $17 \%, 13 \%$, and 41%, respectively as compared to existing scheme.