Diagnosing for printer maintenance with AI
When exactly does a building or item of equipment require maintenance? Carrying out too much preventive maintenance can be costly. But if you leave it too long, the expense of an emergency repair can be much greater. And we all know that machines have a habit of breaking down at exactly the wrong time. So predictive maintenance is high on the list of priorities of the manufacturing industry. Here, too, artificial intelligence (AI) can make all the difference.
A customer is itching to receive the printed matter he has ordered. The printers are running flat out – as are all the other machines. It is not just tough quality requirements that the graphic industry has to meet, but in many cases tight deadlines as well.
That makes it an ideal sector for introducing predictive maintenance based on artificial intelligence. And that is exactly what TNO and Canon Production Printing (formerly Océ) are currently aiming to do. Together, they are carrying out research into an AI system that stands out in terms of reliable diagnoses and prognoses for professional printers.
AI that understands printers
A lot of different data is needed to be able to estimate the condition of a machine or machine components. But much of that data is incomplete or unreliable. To be able to extract the right conclusions from this mix of data, you need AI that is entirely at home in the world of printers.
This calls for 'hybrid AI' – a combination of data-learning AI and domain knowledge. The AI system is able to modify the likelihood of causal links in the domain model on the basis of a machine’s user data. It can also work out what problems could occur and in what kind of time frame.
Still essential: human expertise
And what about people? The part they play should not be underestimated. Machine learning, then, is just one part of the story. It is precisely the combination of artificial intelligence and human expertise that make this solution so powerful. The big challenge here is to develop an AI system that is capable of accurately combining the input from people and machines alike.
Get inspired
System integration for robots in greenhouses
Good collaboration between humans and AI robots in greenhouses is essential. We are working on a semantic explanation and navigation system for improved communication.


AutoAdapt: Self-adaptive machine learning to propel us into the future
Self-adaptation in computing is a concept that might prove to be the missing link in making AI more transparent and accelerating innovation.


Developing moral models for AI systems
It is vital that values such as safety, trust and well-being are integrated in the decision-making process. How far are AI systems that we can trust them?


AI Systems Engineering & Lifecycle Management
The AI system for the future. At TNO, we work on AI systems that remain reliable and can handle new functions in the future.


Responsible decision-making between people and machines
Bias in facial recognition and accidents with self-driving cars. AI must be developed further. The fastest way to do this is in close cooperation with people.

