Publication Details
Abstract
This article develops a plan for integrating artificial intelligence (AI) into the Quality Management System (QMS) of an enterprise for the production of automobiles and their components. It describes the research and a practical deployment package across the continuous quality lifecycle from body geometry and welding to painting, final assembly, and production end-of-line (EOL) testing that combines computer vision, machine learning (ML), and predictive analytics with the plant’s existing measurement and control infrastructure.
The method is based on an extensive catalogue of automotive control points body geometry, panel gap symmetry, weld integrity, paint parameters, torque discipline, paint film thickness measurement, torque discipline, chassis alignment, electronics, and EOL verification mapped to measurable tolerances and operational targets. A VIN‑oriented data structure, model selection guidelines, decision-making logic, and governance tools for traceability and corrective actions are proposed. Evaluation metrics (precision/recall, mAP, Cp/Cpk, false‑escape rate) and expected benefits (reduction of defect escapes, faster root‑cause analysis, higher overall equipment effectiveness, OEE) are explained. The contribution is a reusable scheme a stack of processes, data, and models plus a metrics suite that maps defect modes to measurable Key Quality Indicators (KQIs).