Intelligent Multi-connectivity Based Energy-Efficient Framework for Smart City

Smart city enhances the intelligence and sustainability of the city assets through advanced and diversified applications characterised by specific Quality of Service requirements. To realise the stringent requirements of these applications, multi-connectivity (MC) emerged as a potential solution that ensures seamless mobility, high reliability and enhanced data rates. However, the smart city verticals driven by the energy-constrained IoT devices experience an energy optimisation challenge in the MC configured environment. Recently, significant research has been reported in the direction of MC but they are incapacitated in terms of enhancing energy efficiency. To address the aforementioned issue, an intelligent framework based on software-defined wireless networking and edge computing has been proposed. The proposed framework leverages the synergetic integration of Double Deep Reinforcement Learning and Matching Game Theory to attain energy-efficient multi-connectivity association policy. In addition to this, the proposed approach defines the preference functions to guarantee service provisioning while respecting the radio access technologies constraints. The analytical results validated through the rigorous simulation exhibited an improvement of 45% in the overall energy efficiency. Furthermore, the proposed association scheme outperformed the other existing schemes by 12%, 27%, and 82% in terms of fairness, robustness and system satisfaction degree respectively.