AI-Powered Robot Control Optimizes a Challenging Handling Process

HWL Löttechnik GmbH (HWL) is a Berlin-based, family-run thermal treatment company with customers in power plant engineering, automotive, electronics, aerospace, and tool industries, as well as plant and fixture construction. Like other small and medium-sized enterprises, HWL faces the complex realities of growing labor shortages, the ongoing energy crisis, and the intricate nature of production processes, particularly in “High-Mix, Low-Volume” manufacturing. To overcome these challenges, HWL directs part of its investments toward the adoption of state-of-the-art technologies and the optimization of operational processes, positioning itself as an attractive employer.

At HWL, a commitment to the highest quality standards and continuous innovation goes hand in hand with prioritizing employees’ expertise and satisfaction. Recognizing the physically demanding and repetitive work involved in thermal treatment and the increased workload for staff due to labor shortages, HWL’s managing director, Kai Lembke, set out to identify the task that could benefit the most from automation.

Collaborating with his dedicated team, Lembke chose the loading of metal parts onto racks from washing baskets. Before heat treatment, the serial parts handled by HWL undergo a meticulous cleaning process in a cabin washing facility. Each customer typically delivers around 900 to 1,000 parts in large loading carriers, which employees manually transfer into washing baskets. The parts in the baskets are then carefully placed onto multi-tiered racks using a peg system before being transported to the oven for heat treatment.

The immense diversity in size and weight of the metal parts, combined with their fragile nature prior to heat treatment, makes the handling process extremely complex. The following types of variances in the process made automation with conventional solutions impossible:

• variance in the size and weight of the components

• variance in the positioning of the components within the washing basket

• variance in the position of the racks

• variance in the appearance of the component due to reflection and changing lighting conditions.

In October 2022, HWL automated the part loading task with a robot equipped with MIRAI, an artificial intelligence-powered vision system designed for tasks that are too difficult or costly to automate with traditional solutions. Using AI, MIRAI generates robot movements in real time, responding to changing conditions in the production environment (e.g., lighting, shape, position, color, or background). Robot skills (or specific tasks) are trained, not programmed, through human demonstration, without users needing programming or AI knowledge.

To start, the robot is repeatedly shown both a task and the environment with the help of a camera that is typically mounted on the robot’s wrist. The recorded information is then transformed into a skill capable of handling variances autonomously.

After the washing process is completed, the MIRAI system takes control of the robot. Guided by MIRAI, it individually lifts the delicate parts from the washing baskets and places them on the pegs in the racks. The AI-based vision system automatically adapts to the varying sizes and weights of these highly individualized parts. Unlike traditional vision systems, even changing lighting conditions and resulting reflections on the surface of the workpieces do not affect the functionality and precision of the system. “By automating processes with MIRAI, we can save costs and position ourselves as an innovative and modern company while enhancing the attractiveness of our workplace,” Lambke explained.

Since the implementation of MIRAI, HWL employees have taken on the responsibility of managing and retraining the system, especially important when faced with changing requirements or environmental modifications.

“The fact that these adjustments can be made by our employees without the need for experienced programmers from the manufacturer is a significant advantage,” Lambke said. “We have even been able to optimize cycle times by 20 percent. That’s exactly the flexibility we need. By automating processes, we can save costs, position ourselves as a progressive and modern company, and enhance the attractiveness of our workplace.”

As a result of these positive experiences, HWL envisions adding more AI-based applications soon.

This article was contributed by Micropsi Industries (Berlin, Germany). For more information, visit here  .

---