Title: A reinforcement learning-based link quality estimation strategy for RPL and its impact on topology management

Abstract: Over the last few years, standardisation efforts are consolidating the role of the Routing Protocol for LowPower and Lossy Networks (RPL) as the standard routing protocol for IPv6 based Wireless Sensor Networks (WSNs). Although many core functionalities are well defined, others are left implementation dependent. Among them, the definition of an efficient link quality estimation (LQE) strategy is of paramount importance, as it influences significantly both the quality of the selected network routes and nodes' energy consumption. In this paper, we present RLProbe, a novel strategy for link quality monitoring in RPL, which accurately measures link quality with minimal overhead and energy waste. To achieve this goal, RLProbe leverages both synchronous and asynchronous monitoring schemes to maintain up-to-date information on link quality and to promptly react to sudden topology changes, e.g. due to mobility. Our solution relies on a reinforcement learning model to drive the monitoring procedures in order to minimise the overhead caused by active probing operations. The performance of the proposed solution is assessed by means of simulations and real experiments. Results demonstrated that RLProbe helps in effectively improving packet loss rates, allowing nodes to promptly react to link quality variations as well as to link failures due to node mobility.