Industrial collaboration worth copying

Marcelo Costa, researcher at the Federal University of Minas Gervais (UFMG), and Aline Kraemer, doctoral student at Instituto Tecnológico de Aeronáutica (ITA), both in south-eastern Brazil, arrived in Linköping during the abnormally snowy winter. They have also experienced a May that Swedes are not accustomed to – the warmest in living memory. But they both just shake their heads when the northerners complain about the heat. “This is what the winters are like in Brazil.”

Six months visit

Marcelo Costa and Aline Kraemer have been guest researchers at the Division for Automatic Control, Linköping University, for six months. They are in Sweden under a Swedish Brazilian exchange programme developed by CISB, with support also from the competence centre LINK-SIC. Successful research in close collaboration with industry has long been a feature at the Division of Automatic Control, and LINK-SIC, which is financed by Vinnova, operates in collaboration with for example ABB, Saab and Scania.

Aline Kraemer, half-way through her doctoral education, is working with the analysis of flight simulation in order to determine what can go wrong. Her goal is to develop failure diagnosis algorithms that will help to improve flight safety.

“I have met and held discussions with researchers at LiU, and with folk at Saab. I’ve become more certain about what I’m going to focus on in my doctoral research. In Brazil we don’t have industry-based doctoral students, and it’s interesting to discuss with the people who work here at the division”, she says.

The largest difference between being a doctoral student in Brazil and in Sweden is the way in which doctoral students are viewed. “In Brazil we are placed with the students, but here I get to work close to professors and other researchers. This has meant a great deal.

Industrial collaboration

Marcelo Costa is also working on faults, but he’s developing machine-learning algorithms that enable a robot or an aeroplane to know when something is starting to go wrong, and raise the alarm before an accident happens. He has been given access to a large database from ABB Robotics, and his research concerns how to benefit from all of the sensors in the robot, enabling it to draw its own conclusions about what is starting to go wrong and when it is time for maintenance. “The same algorithms and mathematical methods are used in machine learning for robots and for aeroplanes,” Marcelo Costa concludes.

He sees the greatest difference as the collaboration with industry. “Some of the professors here split their time between industry and the university. I haven´t seen anything like that in Brazil. The exchange is interesting in both directions: we find out which questions need to be solved in industry, while at the same time theoretical advances can be used in industrial applications. My university is owned and financed by the Brazilian government, but my department, which carries out research within production, has some projects in collaboration with industry. This is not by any means common,” he says.

He is also impressed by the way in which researchers at the Division of Automatic Control work. “I am impressed by how the division is managed, how the work is planned, in both the short term and the long term. We have meetings every Monday where such matters are discussed. The collaboration with industry is also interesting, as is how the students get involved through degree projects.” 

Marcelo is planning on taking some of these ideas home with him.
“It’s difficult to change the way things are organised at home, but we can work together in small groups. In collaboration with colleagues, we can apply for financing and establish closer ties to industry. People are interested in this”, he says.

Source: SVENSELIUS, Monica Westman. Industrial collaboration worth copying.Available on Access on February 12nd, 2019.