Search

Federated edge learning (FEEL) emerges as a privacy-preserving paradigm to effectively integrate edge computing for the implementation of deep learning-based vehicular applications. Nevertheless, the incentive mechanism for vehicles participating in varied learning tasks, has not been well explored yet. In this paper, software-defined network (SDN) technology is adopted for the training control among vehicles, and a novel FEEL framework, namely SDN-assisted semi-decentralized FEEL (SSD-FEEL) is investigated, where multiple edge servers collectively coordinate a large number of vehicular models from different learning tasks. By exploiting the low-cost and similar learning tasks among vehicles and edge servers, SSD-FEEL incorporates more training samples, while enjoying the flexibility of edge server assisted model aggregation. Aiming at motivating vehicles to actively participate in training while improving the model accuracy of multiple learning tasks, a joint contract design and task reorganization problem, combined with the evaluation of model convergence and contract performance, is formulated. Then, we propose a two-stage optimization algorithm incorporating iterative reward allocation and task matching, where the model parameters in different tasks are reconstructed according to the matching results with the mobility constraints. Extensive experiments conducted on multiple data sets validate that the proposed algorithm can achieve higher cluster utility and outperform the conventional multi-task FEEL schemes in terms of learning performance.

In current 6G study, space-aerial-ground integrated networks (SAGIN) are attracting the eye-catching attention. Traditional ground networks and facilities (e.g., open random access networks, Open-RAN) simply provide two-dimensional (2D) and limited-coverage services at low altitudes. SAGIN are designed to overcome these limitations and promise to provide 3D and seamless services and connectivity opportunities. Meanwhile, network softwarization (NetSoft) is highlighted as dominant 6G attribute. Through incorporating NetSoft, dedicated network appliance will be isolated as software blocks and general-purpose hardware. Initiated and customized services will be realized by chained software blocks and share the same underlying hardware. Thus, the resource usage and service diversity can be simultaneously achieved by adopting NetSoft. In current literature, SAGIN and NetSoft are well investigated. However, their joint research is at early stage. Therefore, we investigate the resource allocation in softwarized SAGIN for achieving open and seamless access services. We propose the efficient resource allocation and orchestration framework, abbreviated as Slice-Soft-SAGIN. Our framework approach is designed to achieve efficient resource allocation solution per slice service. Each time receiving one tailored slicing service, the first checking available resources will be conducted by our Slice-Soft-SAGIN. When passing the preliminary resource checking, the formal resource allocation of slicing service will be executed in order: first terrestrial node, second aerial node, third satellite node. In addition, our Slice-Soft-SAGIN executes wireless (spectrum) and wired (computing and storage) types of resources when serving this slicing service. This is one crucial highlight of our Slice-Soft-SAGIN framework. With the purpose of confirming merits of Slice-Soft-SAGIN, the comprehensive evaluation work is executed. Typical works and three extra counterparts constitute the whole evaluation. We analyze the evaluation results carefully.