StockBot: An Autonomous Smart Shelf Management and Restock Robot

The demand for dark stores that maintain individual items on accessible shelves for quick fulfillment has increased due to the rapid growth of online shopping. However, maintaining optimal stock levels on these shelves remains a challenge. The labor-intensive and frequently delayed nature of traditional restocking procedures can interfere with order fulfilment and lower customer satisfaction. To address this problem, StockBot, an autonomous robot, is proposed to monitor and control stock levels in dark stores. StockBot works in tandem with an independent webcam system using computer vision and predefined locations for shelves and items to continuously monitor the inventory on each shelf. This enables it to pick and transfer objects from storage to shelves effectively and without the need for human assistance. The performance of StockBot is assessed according to metrics such as item and low-stock detection accuracy and navigation efficiency. Through stock management automation, StockBot reduces the chance of stockouts while saving time and effort in the restocking process. By increasing operational efficiency in dark stores and enabling real-time inventory management through vision-based object detection, our work has the potential to enhance how companies satisfy customer demand.