Analog and Multi-modal Manufacturing Datasets

本数据集由南卡罗来纳大学机械工程系创建，旨在通过模拟和多模态数据推动制造业智能化的研究。数据集包含两个部分：模拟数据集记录了30小时的装配线操作数据，包括传感器数据和操作异常；多模态数据集则同步记录了装配线的图像和模拟数据。数据集的创建过程中，研究人员手动引入了装配缺陷以模拟实际生产中的异常情况。这些数据集特别适用于训练和验证用于制造业的人工智能模型，尤其是质量控制和过程优化领域。

This dataset was developed by the Department of Mechanical Engineering at the University of South Carolina to advance research on intelligent manufacturing using simulated and multimodal data. It comprises two core components: the simulated dataset, which records 30 hours of assembly line operation data including sensor readings and operational anomalies, and the multimodal dataset, which simultaneously captures both assembly line images and corresponding simulated operational data. During the dataset creation process, researchers manually introduced assembly defects to simulate anomalous scenarios in actual production environments. These datasets are particularly well-suited for training and validating artificial intelligence models for manufacturing applications, especially in the domains of quality control and process optimization.