Issue link: https://iconnect007.uberflip.com/i/1379105
74 SMT007 MAGAZINE I JUNE 2021 sition (including edge mobile app), inspection checklist management, and inspection data. Figure 3 shows the hybrid quality inspec- tion framework proposed. In this framework, AI-assisted inspection and human cogni- tion inspection share common edge integra- tion and communication as well as common data backup and report strategy. e proposed framework offers three key advantages: • Great operational flexibility to switch between AI-assisted inspection and human cognition inspection when there is a disruption in AI service • A common data layer for the consumption and retrieval of quality inspection data • As companies move towards more AI-assisted inspection, minimal investment would be required to upgrade the inspection stations at the edge. • It would enable smoother transition into AI-assisted inspection by integrating inspection with human cognition into the framework in early stage. Conclusion is paper discussed the prospect of deploy- ing large-scale AI-assisted inspection onto manufacturing floors, both within a plant and across plants in different geographies. An edge computing architecture was presented as a via- ble solution to achieve the operational require- ments in availability, scalability, performance and security. A hybrid inspection frame- work was proposed to address the operational needs of running both AI-assisted inspec- tion and human cognition inspection concur- rently on the manufacturing floor. A full solu- tion based on the edge computing architecture and inspection framework are being deployed into IBM manufacturing facilities to transform the quality inspection process with higher effi- ciency and accuracy. Figure 3: Hybrid inspection framework illustration.