AI inspection reduces kitchen appliance line check time by 25 percent
AI-Driven Collaborative Assembly Line Inspection: System Integration and Deployment Challenges
RoboticsComputer Vision and Pattern RecognitionHuman-Computer Interaction
Summary
Inspecting products on assembly lines by hand is tiring and can miss defects, especially after long shifts. The authors created a system combining two helpful robots and AI to spot issues faster and with less human fatigue. Their system used cameras, microphones, and sensors to check product quality and safety more quickly and reliably. They shared the challenges they faced and how their design made inspections 25% faster and much easier for workers.
What this means in practice
- •For industrial automation engineers: Deploy AI-enabled robot inspection cells to speed up quality checks and reduce human error in manufacturing lines.
- •For smart factory integrators: Integrate multi-sensor data streams in collaborative robotics to enhance real-time defect detection and operator assistance.
Authors
Asya Ünal, Amr Okasha, Ege Çırakman, Perin Ünal
Abstract
Manual visual inspection on assembly lines is a persistent manufacturing bottleneck: operator fatigue over extended shifts lowers defect-detection rates. This paper presents the design, integration, and field deployment of an AI-assisted collaborative inspection cell at the Silverline kitchen-appliance factory, developed within the AI-PRISM project. The cell couples a Universal Robots UR 10e cobot carrying a machine-vision defect-detection pipeline with a Comau Racer-5 cobot for functional tests, coordinated through ROS 2 Humble on an Ubuntu 22.04 LTS server. Multi-modal data (Basler camera imagery, TIA microphone acoustics, and SPS electrical-safety measurements) are logged locally and visualised in real time with Grafana. We report the practical deployment challenges (close-proximity safety, AI robustness under glare and reflections, ROS 2 namespace collisions across two cobots, and operating-system and dependency issues) together with the engineering solutions adopted, and structure the integration through a four-level Human-Robot Interaction analysis. The deployed cell cuts per-unit quality-check time from 82 s to 61 s (about 25%), raises final-control resource efficiency from 0.75 to 0.88, reduces operator visual-inspection viewing time by 82%, and significantly lowers operator mental demand (p = 0.005, NASA-TLX).