In manufacturing, even the most efficient assembly lines regularly grind to a halt. Professor of Artificial Intelligence Lambert Schomaker is working with regional and international industry to develop smart systems that can detect and solve problems along an assembly line. As part of a European programme involving 47 partners from academia and industry, the project has been awarded € 30 million by Europe.
‘Although the sub systems in industrial production lines are usually heavily optimized, the productivity of the total line is often still disappointing. One of the main reasons is that a general context model of the line as a whole is often lacking’, explains Lambert Schomaker. The Mantis project hopes to change this state of affairs. The project has been awarded more than € 30 million in funding from the European Commission Horizon 2020 programme, € 540,000 of which will go to the University of Groningen.
The aim of Schomaker’s part of the project, which he will carry out together with regional industry, is to design a system that combines all the knowledge obtained from the entire assembly line and uses it to respond to problems. ‘The system will be able to identify problems and intervene where necessary, in the same way as the body recognizes and tackles problems like infections. It may even be able to take pro-active measures.’
Sensors compile information from the entire production chain into a self-learning system, which generates solutions based on the information it receives. The next step is to build a flexible shell of hardware that will allow the smart system to implement its own solutions. The production line will react like a living organism: it will recognize faults and take the appropriate corrective action.
Schomaker’s research group at the institute of Artificial Intelligence and Cognitive Engineering (ALICE) will provide the machine-learning software. ‘The input for the system comes from big data generated by the sensors in the production line, as well as from the operators.’ They will be given an iPad to use on the work floor, enabling them to make detailed reports of any faults and take photos where necessary. ‘The iPad belongs to the production line and will be passed from shift to shift. Software will read and perform text mining on the comments made by the operators, so that we can link the data from the sensors to human knowledge about the causes of problems.’
Three years from now, this should eventually result in robustly designed and controlled systems that keep going even if, e.g., one of the bearings wears out. ‘This is an important lesson we have learned in artificial intelligence: It does not count whether a system is merely beautiful, it should be robust and resilient in the first place. Smart factories are now beginning to take the same step, says Schomaker. ‘A delegation from the regional production industry was visiting us, so we showed them a robot that could grab its own can of coke. After the verbal command we secretly took the can of coke away and put it in another place. The robot was still able to find and retrieve it. “Now that’s what we want!” was the reaction of the delegates.’
All 47 parties in Mantis will share their experiences with each other. Schomaker: ‘The solutions we develop in Mantis will be useful to other companies too.’ The ultimate aim is to lower the production costs in European industry so that Europe will continue to be a serious market contender.
More information about the Mantis project
Contact: Lambert Schomaker
Prof. Slotboom co-applicant in awarded ZonMw project application
On Thursday 10 June, a partial solar eclipse will be visible in the Netherlands. The eclipse will start at 11.21 a.m., with a peak at 12.27 p.m. At 1.36 p.m., the eclipse will be over. Eclipse glasses will be handed out at five locations.
De vier Nederlandse solar teams (Delft, Eindhoven, Groningen, Twente) slaan de handen ineen tegen klimaatverandering De studententeams bouwen hun eigen zonneauto’s en ontwikkelen innovaties voor een duurzamere toekomst, maar zien tegelijkertijd dat...
The UG website uses functional and anonymous analytics cookies. Please answer the question of whether or not you want to accept other cookies (such as tracking cookies).
If no choice is made, only basic cookies will be stored. More information