Scientists from CXI TUL are developing laser systems that can handle the task without the need for manual intervention by a technician. The laser adjusts itself precisely to the need and detects by sound whether the process is running correctly. The research in the LasApp project is aimed at space and regular production.

Adjusting a laser is a delicate job. It only takes a fraction of a millimeter for the mirror to move and the entire system stops working. Today, this step is performed manually by a technician, repeatedly and as the situation requires. But teams at CXI TUL are working on making the laser self-sufficient.

A detuned laser? It's not just that

The heart of the device is the laser cavity, the place where the beam is created, reflected many times and amplified. The mirrors must be in place with micrometer precision. But over time they become out of tune . Temperature, vibration, simply operation .

The CXI TUL team is testing a system that adjusts the position of the mirrors using piezoelectric motors – small but extremely precise actuators. The laser tracks where its own track lands using a sensitive sensor. If it deviates from the correct position, the system recognizes it and corrects it itself.

At the current stage, the accuracy is in the order of tenths of a micrometer. To give you an idea: the thickness of a human hair is around 70 micrometers.

AI returns the laser to the optimal position

Recognizing the error is one thing. Fixing it quickly and efficiently is another. That's why the team is testing machine learning methods that learn how to navigate the mirrors to the correct position as quickly and with as few moves as possible.

The goal is clear: to get the laser beam exactly into the center of the sensor, with an accuracy of less than three micrometers. In May 2026, a new IR viewer device was added to the laboratory, thanks to which the infrared laser trace can be observed directly inside the system. This helps to set the optical path more precisely and prepare the laser for live testing.

Why is this important? In places where a technician can't be sent for service, like in space, a laser must be able to repair itself.

When mistakes are heard

While the first part of the research looks towards space, the second is aimed at industry, where it could bring concrete practical applications. The team is focusing on a technology called laser peening . This is laser processing of the surface of a material, which is used, for example, in aircraft or automotive components.

Each laser pulse produces its own sound signature. If the process is going well, the sound is different than if something goes wrong.

The researchers collected and analyzed these short audio recordings, and based on this, they created their own application that can display and compare signals in both the time and frequency domains.

Accuracy over 90 percent

Based on more than 21,000 recordings, they created a model that can distinguish between a correct pulse and an incorrect one. They used the XGBoost method, a commonly used machine learning tool for classification.

The result: 93 percent accuracy . This means the machine can check the quality of production in real time, just by how the process sounds. Without having to inspect each piece under a microscope.

A unique combination of lasers, artificial intelligence and smart diagnostics

All this is part of the LasApp project, which connects top laser centers of the Academy of Sciences of the Czech Republic with other workplaces. The coordinator is the Institute of Photonics and Electronics of the Academy of Sciences of the Czech Republic, other partners are the Institute of Physics of the Academy of Sciences of the Czech Republic - HiLASE Centre, the Faculty of Science of Charles University - BIOCEV Centre, the Institute of Instrumentation of the Academy of Sciences of the Czech Republic, the Institute for Nanomaterials, Advanced Technologies and Innovations of TUL and the Institute of Plasma Physics of the Academy of Sciences of the Czech Republic - TOPTEC Centre.

Six institutions, one goal: to bring laser technologies closer to smart manufacturing and future space missions. The research in the LasApp project combines optics, automation, artificial intelligence and diagnostics of manufacturing processes into one whole.

And when something goes wrong, there will be no need to send a technician into the bowels of the device like a hero into the dark corridors of a spaceship, where an intruder may be lurking. A laser that adjusts itself and recognizes in time that something is not going according to plan can, thanks to research at CXI TUL, prevent similar dramas before they get serious.

Source: Science and Research