The new artificial intelligence system can detect when a malfunction occurs on the robot assembly line

According to foreign media New Atlas, in factories, when the two interlocking components of a product are together, a satisfactory "click" sound is usually emitted. By listening to this sound, a newly developed system can detect when a malfunction occurs on the robot assembly line. This technology was created by the Fraunhofer Digital Media Technology Research Institute (IDMT) in Germany, which includes audio sensor units that can be directly installed on robotic devices, as well as on nearby walls or other surfaces.

Each unit is no larger than a pack of cigarettes and contains a microphone, a battery, and a microprocessor that can analyze audio signals within the module.

After being initially trained to recognize the sound of correctly connecting specific components, the system can then determine when the robot has undergone an action that should have led to a successful connection, but the sound has not occurred. Then, it will display error messages, alert human supervisors, and record events.

Using artificial intelligence based algorithms, the device can also ignore irrelevant background noise and distinguish multiple target sounds. This means that a single sensor can be used to check the successful connection of multiple components.

"In fact, we provide machines with an auditory system to ensure quality," said Danilo Hollosi, Director of Acoustic Event Detection at IDMT. "This enables manufacturers to promptly identify signs of damage in the early stages, reduce unplanned downtime, coordinate workshop workflows, and improve the efficiency of the entire factory."

Original title: Quality control system based on artificial intelligence can be used to monitor the "click" sound on assembly lines

Prev: Inspired by insects, scientists have developed “flexible skeleton” robots