Charlie Street, Bruno Lacerda, Manuel Mühlig, Nick Hawes, "Proactive Multi-Robot Task Allocation Under Spatiotemporal Uncertainty", Autonomous Robots and Multirobot Systems (ARMS) 2022, 2022.

Multi-robot task allocation methods should be robust to task announcements during execution, where task announcement times and locations are uncertain. In this paper, we model task announcement us- ing continuous-time Markov chains which can be learned from empirical data. We then evaluate announcement time and location distributions through model checking techniques. To service uncertain tasks efficiently, allocation should occur proactively, such that robots are near or at the task location upon its announcement. We present a framework which extends sequential single-item auctioning to minimise the expected total waiting time, i.e. the time between task announcement and a robot be- ginning to service the task. To execute tasks proactively, robots navigate to intermediate waiting points which are near potential task locations, and wait for announcement. We demonstrate the efficacy of our approach in simulation, where we outperform baselines that do not consider task models and intermediate waiting points.