Machine Learning based Energy Management at Internet of Things Network Nodes

Abstract

The Internet of Things networks comprising wireless sensors and controllers or IoT gateways offers extremely high functionalities. However, not much attention is paid towards energy optimization of these nodes and enabling lossless networks. The wireless sensor networks and its applications has industrialized and scaled up gradually with the development of artificial intelligence and popularization of machine learning. The uneven network node energy consumption and local optimum is reached by the algorithm protocol due to the high energy consumption issues relating to the routing strategy. The smart ant colony optimization algorithm is used for obtaining an energy balanced routing at required regions. A neighbor selection strategy is proposed by combining the wireless sensor network nodes and the energy factors based on the smart ant colony optimization algorithm. The termination conditions for the algorithm as well as adaptive perturbation strategy are established for improving the convergence speed as well as ant searchability. This enables obtaining the find the global optimal solution. The performance, network life cycle, energy distribution, node equilibrium, network delay and network energy consumption are improved using the proposed routing planning methodology. There has been around 10% energy saving compared to the existing state-of-the-art algorithms.