Automated, AI-based Inspection of Drive Wheels on Overhead Hoist Transport Vehicles Zhu Hailong Rank Sebastian Schmidt Thorsten Overhead hoist transport systems are used to transport wafers in 300 mm semiconductor factories. These rail-based systems usually consist of hundreds of vehicles to ensure fast and safe transport of wafers between tools. Faults of individual vehicles can cause damage to the transferred goods and production downtimes. To minimize the risk of failure, extensive preventive maintenance of the vehicle's heavily stressed components is required. This includes the chassis and drive wheels. This article describes an automatic inspection approach that can drastically accelerate the inspection process. We have developed an automatic inspection approach for the drive wheels that can drastically speed up the inspection process. From the data obtained, we trained a deep convolutional autoencoder network to predict the growth of fractures on the surface of the wheels. With the help of our inspection approach, it is possible to carry out conditionbased predictive maintenance of the OHT vehicles. This approach promises cost savings compared to routine or time-based strategies for preventive maintenance, as we can carry out maintenance tasks only when they are justified. Fehlererkennung OHT Verschleißmodell Zustandsüberwachung autoencoder condition monitoring faults detection wear out model 620 periodical academic journal Logistics Journal : Proceedings 2021 17 2021 2192-9084 urn:nbn:de:0009-14-54509 10.2195/lj_Proc_zhu_en_202112_01 http://nbn-resolving.de/urn:nbn:de:0009-14-54509 zhu2021