Title: Learning Green Routes: It Pays to be Smart ( but not too Smart!)
Abstract: We will explore the impact of recent technologies on the performance of wireless networked systems that enable the Internet of Things (IoT), namely, green wireless networks. Specifically, we describe advances in wake-up radio technology with semantic addressing capabilities and its usage to obtain very short packet delivery delays and up to five orders of magnitude energy savings. We will then illustrate how combining energy harvesting from different sources (e.g., solar and wind) with wake-up radios enables perennial networked systems with performance similar to that of systems with battery-operated nodes. We demonstrate the positive impact of using smart (e.g., optimally-solved learning-based) forwarding techniques in green wireless networks, highlighting its non-negligible costs and advocating the implementation of smarter remedies (heuristic solutions of learning-based models).