Blockchained Industry Information Handoff Based on Internet of Things Devices with Intelligent Customized Object Recognition

• Published in: Sensors (Basel, Switzerland) Vol. 22; no. 6; p. 2312
• Main Authors: Jian, Ming-Shen, Pan, Chin-Ju
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 17.03.2022; MDPI
• Subjects: Accuracy; Algorithms; Artificial intelligence; artificial intelligence (AI); Assembly lines; Blockchain; Cloud computing; Communication; Customization; Data encryption; Data exchange; Data transmission; Digital signatures; Internet of Things; Internet of Things (IoT); Manufacturing; object detection; object recognition; Privacy; Product lines; Production lines; Supply chains; object recognition; Internet of Things (IoT); object detection; blockchain; artificial intelligence (AI)
• ISSN: 1424-8220; 1424-8220
• DOI: 10.3390/s22062312

To determine the quality and safety of each product used in manufacturing, the exchange of measured data between machines, operators, production lines, and manufacturing companies is crucial. In this study, we developed a system with customized object recognition capability for the secure blockchain-based transfer of industry information through Internet of Things (IoT) devices. In the proposed system, product history data are transferred through blockchains through artificial intelligence (AI)-based object recognition. Individual objects are recognized and represented using a unique number sequence for use as a private key on a blockchain. The data history can be automatically secured, and all the data are traceable and trackable. The reliability and validity of the proposed system were verified using the Jetson Nano Developer Kit. The proposed AI-based system is a low-cost embedded system. Based on the open-source cloud computing platform, the required computing resources for blockchain computing and storage are available. In an experiment, the proposed system achieved >99% accuracy within 1 s. Furthermore, the computational cost of the proposed system was 10% that of traditional AI systems. The proposed device can be rapidly connected to IoT devices that require limited manual operation and can be adopted in manufacturing and production lines.