[현장] 1700개 용접점 향한 길도 AI가 찾는다…현대차가 그리는 미래

At the <location>, a pathway of 1,700 welding points was filled with an AI-generated robotic arm. The arm paused and moved closer to the welding points, adjusting its path based on the virtual space created by the researchers. The camera sensor recognized the altered virtual factory layout and presented it on the screen. The AI then generated a smoother, uninterrupted movement of the robotic arm.

This demonstration took place on the 16th at Hyundai-Kia's Andong Research Center, showcasing a robot's ability to automatically generate movement paths to each welding point. Hyundai-Kia's manufacturing system integration technology team's manager, In-hwan Lee, explained that this automated path generation technology would reduce initial production line setup and production costs, cutting down on the time and effort required for engineers to manually set each welding point's location.

The plan is to complete and review the technology for mass production by June 2027. The three-day 'IpoForest TechDay 2026' opened on the 19th at the Andong Research Center, showcasing various future manufacturing technologies utilizing AI and software. Known as Hyundai-Kia's smart factory brand, 'IpoForest' hosts 200+ manufacturing technologies each year.

This event marks the seventh time Hyundai-Kia has participated. Their ultimate goal is to achieve an 'autonomous factory' where humans and robots collaborate, and AI learns and makes decisions based on data, continuously optimizing the production process. This includes risk-prone and repetitive tasks being handed over to robots, while humans focus on judgment and quality improvement.

The key to achieving this vision is data connection and integration. Previously, each factory developed and operated its production, quality, and equipment management systems separately, with data and development capabilities spread across multiple locations. Hyundai-Kia aims to address this issue by standardizing data collection, management, and development environments through their 'software-centric factory' (SDF) initiative, ensuring quick application of developed manufacturing technologies and software to other domestic and international factories.

The AI agent 'Polaris' demonstrated how data integration can be utilized in real-world scenarios. When asked about yesterday's primary defects, Polaris interpreted the request as 'welding, stamping, and assembly' processes and presented defect analysis results and visual data. Polaris also analyzed and generated reports on information that would have required humans to navigate through various systems.

At the exhibition today, a robotic arm with a 'hand gripper' attached moved wire used to fasten electric vehicle parts, mimicking a human's hand. This was previously done by the worker themselves. Hyundai-Kia's manufacturing system integration technology 2 team's manager, Joon-hun Yu, explained that they are developing hand grippers that mimic human hands to handle various parts.

The technology is still in the early development stage but is expected to become a core component in future automotive automation. There are also technologies where robots replace human workers in tasks like workers bending down to install door latches (door 'garnish'). Hyundai-Kia plans to implement this technology at their Kia Gwangju 1 Plant in December 2027.

Quality checks for finished vehicles also utilize AI. At the tunnel-type exterior inspection device, the AI analyzes the position and size of defects after capturing high-definition images of each vehicle, displaying them on a screen overlaid on the vehicle's body map. Hyundai-Kia plans to implement this technology at their Kia Hseongseong EVO Plant in July next year.

The level of technology readiness varies from experimental operation to mass production preparation and early development stages. The automation expansion impacts the nature of labor; Hyundai-Kia's center director for IpoForest, Young-ta Kim, emphasized that the goal is to enable employees to focus on safer, more valuable work, while AI helps the factory operate more flexibly and sustainably.