Autonomous Electric Vehicles as Mobile Green Energy Sources

We envision a wide deployment of battery-operated edge devices such as data kiosk, to provide pervasive data collection and dissemination services. Although these data kiosks can be powered by green energy sources, their batteries may deplete overtime, and have to recharged frequently to prevent power outages and potential loss of critical data. In this paper, we propose an edge device recharging system using autonomous electric vehicles as Mobile Chargers (MCs). First, we determine the numbers and locations of Dispatching Centers (DCs) for these MCs, each of which will be responsible for recharging some edge devices in a surrounding area. Then, we plan optimal routes of the MCs in order to minimize the number of MCs needed to recharge all edge devices before a deadline. To reduce the time complexity involved in optimizing the routes, we cluster the edge devices and propose efficient algorithms to plan the route of MCs for each cluster based on the relaxation and rounding of a Mixed Integer Linear Programming (MILP) model. Extensive simulations show that our approach can reduce the number of MCs required to recharge all edge devices before a deadline by up to 71.93% compared with greedy-based heuristic algorithms.