Robot-based cable production quality management system and method

By using an improved variational autoencoder algorithm and a multimodal fusion network, the problems of data feature extraction bias and defect prediction in traditional cable production quality control are solved, enabling real-time dynamic control and parameter adjustment of the cable production process.

CN122114761AActive Publication Date: 2026-05-29FUZHOU YONGTONG WIRE & CABLE
7 Cites 0 Cited by

Patent Information

Application Number
CN202610580464.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-29
Publication Date
2026-05-29
Estimated Expiration
2046-04-29

AI Technical Summary

Technical Problem

Traditional cable production quality control methods do not take into account the actual physical rules and process standards of cable production. The feature extraction results of multi-source production data deviate from the actual working conditions, making it impossible to achieve three-dimensional reconstruction. Defect prediction cannot focus on key features, making it difficult to output the potential defect probability of future production links and generate real-time process parameter adjustment instructions adapted to robots.

Method used

An improved variational autoencoder algorithm is used to impose structured constraints on multi-source production data, generating deep quality feature representations. Combined with process state transition modeling and 3D defect reconstruction, the defect probability distribution is analyzed through a multimodal fusion network to generate real-time process parameter adjustment instructions.

Benefits of technology

It achieves precise correspondence between the feature extraction process of multi-source production data and the cable production process, fully presents the evolution path of the process state, restores the three-dimensional morphology of appearance defects, integrates various quality information through multi-modal data fusion, accurately outputs the probability distribution of defects in future production stages, generates real-time parameter adjustment instructions for robots on the production line, and realizes real-time dynamic control of the production process.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to the technical field of intelligent cable manufacturing, in particular to a robot-based cable production quality management system and method, and the method comprises the steps: collecting multi-source production data, such as a spectral feature sequence, a process parameter time sequence flow, a multi-view visual image and an online sensor measurement value, obtained by a robot; an improved variational auto-encoder integrated with cable production physical constraints and process rules is adopted to extract deep quality features, a process state evolution path is constructed, and appearance defect three-dimensional voxel reconstruction is completed. And inputting the multi-dimensional features into a multi-modal fusion network to obtain a comprehensive quality tensor, outputting a potential defect probability of a future link through an attention mechanism defect prediction network, and generating a robot real-time process adjustment instruction according to the potential defect probability. According to the mode, the feature extraction fitting degree can be improved, the defect form is completely presented, accurate defect pre-judgment and dynamic production parameter regulation and control are achieved, and the accuracy and real-time performance of cable production quality control are improved.
Need to check novelty before this filing date? Find Prior Art