Personalized and sustainable federated learning in integrated Internet of Things
Xu Zheng, Yifu Zheng, Tingqi Wang, Chong Mu, Ke Yan, Ling Tian · Digital Communications and Networks · 2026
Integrated Internet of Things (IoTs) bring novel opportunities for pervasive smart services, as such systems allow for seamless information and resource sharing among IoT devices. Meanwhile, federated learning emerges as a new framework for distributed deployment of machine learning models and becomes a promising approach for the implementation of intelligent IoTs. However, integrated IoTs are usually composed of diverse IoT devices from different systems, such that their ownership, roles, data distribution, and capabilities are heterogeneous. Current federated learning algorithms mainly focus on handling Non-IID issues, and usually suffer reduced and unsustainable performance in integrated IoTs. Therefore, we investigate in this paper the problem of personalized and sustainable federated learning in integrated IoTs. First, we argue that different parties in integrated IoTs are heterogeneous and limited in available resources for federated learning, and these parties are also selfish and expect rational outcomes during cooperation, which guarantees the sustainability of integrated IoTs. Then, this paper provides a novel framework for device selection in federated learning. It first sets one instance of the model for each device, and iteratively selects devices to participate in model training based on the joint consideration of local model accuracy, similarity of parameters, and remaining resources per device. The proposed method guarantees the rational allocation of resources to balance the accuracy across all devices. In this way, the sustainability of the whole IoT system is improved such that no devices will suffer extreme resource exhaustion or severely poor performance. Finally, extensive evaluation is conducted to validate the advanced performance of the proposed method in integrated IoTs.