The EnduRAI pilotsHuman-centric augmentation put to work
Different products, shared issues
The context
In most cases, manufacturing problems have nothing to do with machines. They follow a fatigued hand, an unstable torch angle, and a moment of strain that no one could anticipate and that becomes a defect by the time it is found in a post-process inspection.
The rationale
In EnduRAI’s view, it’s best to experience it first rather than recreate it later. This response is put to the test at two real industrial plants: GOSAN, a Spanish company manufacturing pulleys for lifting, transport, and machinery; and CEMRE, a Turkish shipyard constructing fishing, offshore and passenger boats. Regardless of the final product, the core problem remains the same: a tired worker, and a mistake nobody catches early. Skilled, physical work that unknowingly causes quality and safety problems at a much later stage than the time of inspection.
The approach
EnduRAI’s solutions (sEMG wristbands, adaptive interfaces, and multimodal AI) are tested directly on the shop floor: during real shifts, with real noise, under real production pressure. Two use cases and one overarching question: can the system anticipate worker’s fatigue and risk to take action, reduce defects, improve weld quality and support operators?
A deep-dive into our demos
In both of our pilots, the concept is straightforward: to prevent fatigue buildup and risk rather than just apply corrective measures when they are already present. EnduRAI feeds real-time information back to the worker in a non-intrusive, non-judgemental way to usher in a paradigm shift in manufacturing.
Inline defect detection in high-precision pulley manufacturing
The site
Spain
For more than 50 years, GOSAN in Spain has manufactured mechanical parts, components, and assemblies for lifting and transport machinery, which are shipped to 70 countries, representing more than 35,000 custom parts each year. It passes through cutting, bending and welding processes, even partially automatic, but the operators still need to set and adjust the parameters manually. A tired hand is an out-of-position weld or a poor weld angle, and today, no one finds out until the post-process inspection, which is time-consuming and costly, too.
The scenario
GOSAN will work on a wristband that, via sEMG, reads the operator’s fatigue, tremor and strain, and IR and visual cameras are used to inspect part geometry in real time. If both signals are in the same direction (e.g. shaking hand, excessive bead buildup, an irregular torch angle), the AI detects it instantaneously. It is sensed by the worker as a small vibration through the wristband, which alerts him or her to adjust before the weld is completed, before inspection.
The results
The goal at GOSAN is 10% fewer defects, 15% less rework and 10% more productivity. Safety and better coordination of welding is the aim at CEMRE in one of the most difficult environments on the shop floor.
Inline defect detection in high-precision pulley manufacturing
The site
Turkey
CEMRE Shipyard, in Turkey, builds fishing, offshore, and passenger vessels. Its most challenging work is welding blocks together: assembling huge steel components to form a hull, often in cramped, poorly ventilated environments and awkward positions. It is semi–manual work that relies heavily on teamwork, and errors are typically only identified by ultrasonic or radiographic testing, which can only be done after the weld has been completed, when fatigue-induced tremors, unstable torch angles and/or weld speed have caused an error.
The scenario
At CEMRE, it’s a crew concept that expands. Multiple wristbands run concurrently inside the welding bay, along with oxygen, carbon monoxide, and temperature sensors, and rugged cameras (IP65/67 rated) that monitor weld quality and torch stability, even through smoke and low light. It is not just the supervisor who notices a welder’s fatigue; the system detects it, and can stop work, share a task, or recalibrate a job before it becomes a safety problem.
The results
Safety and better coordination of welding is the aim at CEMRE in one of the most difficult environments on the shop floor.