A Context-Aware, Decentralized Learning Approach for Fog-based Smart and Connected Community

Smart and Connected Community (SCC) used to provide the essential network platform to transform the physical communities into connected communities and to perform the synchronized tasks in an advanced manner. It is also instrumental in building and managing new communities to achieve socioeconomic and environmental sustainability to the connected groups. Context-aware computing can be leveraged in SCC for a better understanding of the surroundings particularly the ones that are of high relevance. Fog networking can be an enabler of context-aware, decentralized, and collaborative learning in a highly distributed, hierarchical network and to intelligently execute localized, modular tasks at the edge of the network. The Fog-based Radio Access Network (F-RAN) emerges as a better alternative of Centralized RAN (C-RAN) for the 5G systems to support latency-critical applications. For handling an ever-increasing, highly distributed community structure and learning from intra-community and inter-community interactions of devices connected to Fog RAN in the 5G system (5GS), we propose a novel multi-level decentralized learning approach. The proposed approach can be implemented with the help of dynamically selected organizer nodes in the fog network. This will enable context-aware tasks for the benefit of the local communities under the SCC framework and ultimately enhance the quality of life in smart cities.