Research Area: Vehicular Ad Hoc Networks
Federated learning (FL) can empower Internet-of-Vehicles (IoV) to help the vehicular service provider (VSP) improve the global model accuracy for road safety and better profits for both VSP and participating smart vehicles (SVs). Nonetheless, there exist major challenges when implementing FL in IoV including dynamic activities and diverse quality-of-information (QoI) from a large number of SVs, VSPs limited payment budget, and profit competition among SVs. In this paper, we propose a novel dynamic FL-based economic framework for an IoV network to address these challenges. Specifically, the VSP first implements an SV selection method to determine a set of the best SVs for the FL process according to the significance of their current locations and information at each learning round. Then, each selected SV can collect on-road information and offer a payment contract to the VSP based on its collected QoI. For that, we develop a multi-principal one-agent contract-based policy to maximize the profits of the VSP and learning SVs under the asymmetric information between them. Through experimental results using real-world on-road datasets, we show that our framework can converge 57% faster and obtain social welfare of the network up to 27.2 times compared with those of other baseline FL methods.
Keywords:
Author(s) Name: Yuris Mulya Saputra; Hoang Thai Dinh; Diep Nguyen; Le-Nam Tran; Shimin Gong; Eryk Dutkiewicz
Journal name: IEEE Transactions on Mobile Computing ( Early Access )
Conferrence name:
Publisher name: IEEE
DOI: 10.1109/TMC.2021.3122436
Volume Information: Page(s): 1 - 1
