Reliable IEEE 802.15.4g based smart utility networks via adaptive modulation selection and re-transmission shaping

In this thesis work, we propose and evaluate a strategy to improve transmission in IEEE 802.15.4g SUN networks. This kind of network is at the basis of many promising IoT applications that require high reliability while maintaining low power consumption. The proposed strategy consists of two distinct parts: re-transmission shaping and modulation selection. The re-transmission shaping mechanism keeps track of unused packet re-transmissions and allocates additional re-transmissions when the instantaneous link quality decreases due to channel impairments. The modulation selection strategies apply Multi-Armed Bandits algorithms to dynamically choose the best transmission modulation. The combined effect of these two mechanisms aims to maximize link reliability while minimizing energy consumption and meeting radio-frequency regulation constraints. To evaluate the proposed methods we use trace-based simulations using an IEEE 802.15.4g SUN data-set and two widely used metrics, the PDR (Packet Delivery Ratio) and the RNP (Required Number of Packets). The obtained results show that re-transmission shaping and modulation selection are useful mechanisms to improve link reliability of low-power wireless communications. Their combined use can increase PDR from 77.9% to 98.7% while sustaining an RNP of 1.7 re-transmissions per packet when compared to using a single re-transmission per packet.