Autonomous Mobile Robot for Automatic out of Stock Detection in a Supermarket

Giuseppe De Simone, Pasquale Foggia, Alessia Saggese, Mario Vento; Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, 2023, pp. 1829-1838

Out of stock is among the main causes of sales losses for retailers. In order to face this issue, in this paper we propose ROSCH (the RObot for SCHelves analysis), an autonomous mobile robotic platform based on ROS framework whose aim is to inform the human operators in case of empty or partially empty shelves, so as to speed up the refilling process. ROSCH is able to autonomously move inside an environment, and to autonomously identify those shelves which are empty or partially empty, thanks to the use of a deep learning based detector, validated on a dataset composed by about 2000 manually annotated images, 900 of them acquired by our team in three different supermarkets in Italy. The proposed system has been tested in a supermarket in Salerno (Italy) at working time; the analysis conducted demonstrates that the proposed system is able to reliably support the supermarket staff, being 8 times faster than the human operator in its common manual out of stock detection activity.