Many tasks require robots to operate in long-duration missions with minimal interruption for recharging or replenishing resources. However, robot power and resources are limited, and without a continuous supply of these resources, robots would have significant downtime for recharging or refreshing other resources, which could be severely disruptive to their mission.
In this project, we are developing a theoretical framework to deploy teams of robots (task robots) for exploration and surveillance while taking into account their energy and resource requirements. We envision a distribution center with replenishable resources (batteries, cameras etc.) that receives or predicts requests for fresh resources from deployed robots, and dispatches agile delivery robots (e.g., quadrotors) to deliver them in a timely manner.
We will address the scheduling and prediction problem underlying this distribution task and propose solutions which generate near-optimal schedules for resource redistribution with multiple incoming requests from deployed robots. The framework will incorporate priorities on task robots which can be changed over time, and allow a relaxed delivery schedule if there are not enough delivery robots available. Delivery robots can also be dynamically re-routed to make efficient use of time and resources to sustain long-duration missions for robotic teams operating in a given region.