Autonomous stock counting based on a stigmergic algorithm for multi-robot systems

• Published in: Computers in industry Vol. 122; p. 103259
• Main Authors: Casamayor-Pujol, Victor, Morenza-Cinos, Marc, Gastón, Bernat, Pous, Rafael
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.11.2020
• Subjects: Inventory; Multi-robot system; Real environment experiments; RFID; Stigmergy; Stock counting; RFID; Multi-robot system; Stigmergy; Inventory; Real environment experiments; Stock counting
• ISSN: 0166-3615; 1872-6194
• DOI: 10.1016/j.compind.2020.103259

Maintaining an accurate and close to real time inventory of items is crucial for an efficient Supply Chain Management (SCM), which is one of the main pillars of successful business decisions in the retail market. Due to theft and misplacement, perpetual inventory systems are not enough for having an accurate picture of the current inventory. However, even if the retailer has implemented an RFID-based solution, manual inventories using handheld RFID readers tend to be tedious, expensive and inaccurate. Therefore, a solution that can autonomously take inventories with high accuracy is expected to have a great impact in the market. One of the most promising possibilities of automatic inventories are inventory RFID-based robots. However, current inventory robots are not yet fully autonomous. This article proposes a fully autonomous solution for an inventory robot that, in addition, can be implemented in very simple robots reducing its cost and therefore its entrance barrier. The article first defines the problem of stock counting and a solution based on a multi-robot system is proposed. The algorithm developed determines the state of the problem using the same RFID tags that retailers add to their items, so they can guide the robot through a complete stock counting task. Simulation and tests in a real environment, a university library, validate the developed algorithm and its application for multi-robot systems obtaining accuracy figures as high as 99.5% of accuracy.