Abstract
<div class="section abstract"><div class="htmlview paragraph">In intelligent surveillance and reconnaissance (ISR) missions, multiple autonomous vehicles, such as unmanned ground vehicles (UGVs) or unmanned aerial vehicles (UAVs), coordinate with each other for efficient information gathering. These vehicles are usually battery-powered and require periodic charging when deployed for continuous monitoring that spans multiple hours or days. In this paper, we consider a mobile host charging vehicle that carries distributed sources, such as a generator, solar PV and battery, and is deployed in the area where the UAVs and UGVs operate. However, due to uncertainties, the state of charge of UAV and UGV batteries, their arrival time at the charging location and the charging duration cannot be predicted accurately. We propose a stochastic modeling approach to deal with these uncertainties based on certain physical assumptions such as the flight time for a UAV, distance travelled for a UGV, and the final state of charge of the battery before they leave the host charging vehicle. Based on the stochastic model, an aggregated charging power demand is forecasted. A model predictive control-based operation is then used for the operation of the distributed sources on the host vehicle to meet the forecasted charging power demand. The host vehicle battery works as a buffer during abrupt changes in the charging power demand. The operational scenario is simulated with ten UAVs, ten UGVs and a host vehicle carrying a diesel generator, a battery pack and a PV system. The result of this work is applicable to energy-aware charging management for a fleet of vehicles.</div></div>