VISION SENSORS AS A COST-EFFECTIVE STRATEGY FOR OPTIMIZING QUALITY CONTROL IN AUTOMATED MANUFACTURING

Manual visual inspection in manufacturing is limited by human factors such as fatigue and variability, with error rates reaching up to 30% in repetitive tasks. In contrast, automated inspection based on machine vision offers greater precision, continuous operation, and traceability. This paper describes the implementation of a poka-yoke system with Cognex In-Sight vision sensors in a manufacturing company in Chihuahua, Mexico, integrated into an Allen-Bradley SoftLogix PLC architecture. The objective was to optimize the inspection and model identification process in an automotive assembly line to prevent customer complaints related to assembly errors. The methodology consisted of integrating the vision sensor as a validation condition in the final state of a control system based on a Moore state machine. The results showed that during 16 test shifts, no failures were recorded in the automated inspection, effectively differentiating between switch models and preventing assembly errors. It is concluded that the solution optimized the process, demonstrating increased productivity and robustness through a hybrid approach of human supervision and automation, with an imperceptible impact on the cycle time.