UAV-Enabled Covert Cross-Technology Communication in Heterogeneous IoT Networks

Heterogeneous Internet of Things (IoT) networks enabled by Unmanned Aerial Vehicles (UAVs) operate in various protocols and spectrum bands (e.g., WiFi, LoRa, Zigbee) to collect and offload data generated from heterogeneous sensors. However, achieving timely and secure communications is challenging due to uncertain data generation, mobility, and location of wardens. This paper presents a collaborative framework that exploits cross-technology communications to achieve covertness constraints. The aim is to minimize the age of covert information (AoCI) and energy consumption by jointly optimizing data scheduling, power allocation, and offloading decisions by collaborating with UAVs. A multi-agent actor-critic algorithm that incorporates federated learning and attention mechanisms (cluster-MAAC-attention) is presented to solve the previous problem. Our simulation results show that our algorithm reduces the worst AoCI by 4 times, and reduces the penalty by 20 times compared to existing schemes.