Using Machine Learning to Estimate the Optimal Transmission Range for RPL Networks

Nowadays, IoT low power networks have been largely adopted in various scenarios such as the factory of the future, smart agriculture, smart cities, and so forth. These networks allow the interconnection of resource-constrained devices (e.g. actuators and sensors) using wireless communication links. In such networks, having energy-efficient routing protocols is necessary in order to maximize the network lifetime. However, many factors may influence the energy consumption in IoT low power networks. These factors include the transmission power of devices, the data routing strategy, etc. In this paper, we address the effect of the transmission power on energy consumption in RPL networks (RPL : IPV6 Routing Protocol for Low-Power and Lossy Networks) and propose a new solution based on a multi-layer perceptron (MLP) model, to estimate the optimal transmission range of dynamic IoT low power networks deployed following a 3D topology. The results of our experiments show that our MLP model perfoms well compared to other machine learning models.