Human-multi-robot team collaboration for efficent warehouse operation

Multi-robot systems have been successfully deployed in modern warehouses where they are required to move merchandise and equipment from one place to another. In such systems, the part of the human worker is often overlooked in the system's design. In this paper we use a novel approach of utilizing automated advising agents for assisting a human worker to better manage her time and workload while supervising a team of multiple robots in a warehouse environment. We introduce the k-Myopic Advice Optimization (k-MYAO) Heuristic which prioritizes the human's tasks in human-multi-robot team collaboration environments. This heuristic is an extension of the 1-MYAO heuristic that was already found to be successful in the Search And Rescue (SAR) task in previous work. We exemplify the k-MYAO's benefit in designing 2 automated advising agents AID1 and AID2 which are deployed in a simulated warehouse. Our proposed intelligent advising agents were evaluated through an extensive empirical study with 60 human operators, and showed significant improvements in the human-multi-robot team performance.