Today, cars are primarily glued and plugged together. This makes it increasingly important in Industry 4.0 that the clicking into place of plug connections is reliable and automatically documented. Why? Because the error warning in a car or on a washing machine is »pre-programmed« if the plug does not sit correctly when it leaves the factory, that is, does not close the connection with the typical »click« of the locking lug. This often leads to high logistics costs and time-consuming rectification. The remedy is a »click detector« in the shape of an acoustic sensor system for in-line monitoring during assembly, which can be integrated as both a mobile device, for example at the end of work, as well as a stationary application, depending on customer requirements. In this way, we at Fraunhofer IDMT in Oldenburg are making a concrete contribution to AI support for shopfloor management on the basis of automated data processing.
Intelligent ears for automated and contact-free production control
The sensors are sensitive to airborne noise and can recognise deviations from the sound. When the plug connection engages, this triggers a »click«, which the microphone or sensor picks up. If there is no click or the sound deviates, the acoustic monitoring system signals an error, which is accurately documented. At the same time, the worker is informed. In fully automated manufacturing, the system can also be used for process documentation and quality assurance. What is special about this solution is that the maintenance system can distinguish between many different types of »clicks« and mechanical impacts, while also filtering out noise interference in loud production environments. After all, not all clicks are the same. A ballpoint pen sounds different to a plug connection – and in addition there are many different kinds of clicking sounds. These innovative automatic detection technologies are uncomplicated and flexible in their handling. Other detection tasks, e.g. recognition of distress calls or accidents, can be easily integrated. For human-machine interaction, speech recognition and voice control are also available.