Smart forklifts test theories in the real world

The smart forklifts are being tested in Toyota’s factory in Mjölby, which employs about 3,000 people. In total, about 80,000 electric indoor forklifts are manufactured here every year. In more concrete terms, smartness means that the forklifts are fitted with integrated telematics which describe in detail what they are doing 24 hours a day. We spoke to Jonas Klang, Program Manager Innovation, whose job includes making operations more innovative.

–Our forklifts have a computer that controls things. Working with the university, we’ve trained an algorithm that automatically classifies what the forklift does based on data streams in the system. The algorithm uses combinations of messages and data to understand what the forklift is doing. The question is whether the same ‘approach’ used for image recognition can be used for this high-resolution data stream, he says.

We are interviewing Jonas just two weeks after the pilot project launched, and it’s going to continue for another month. The hope is it will lead to more efficient use of the forklifts, improved service and maintenance and that more knowledge about how customers use their forklifts will help to avoid breakdowns.

–We connected our forklifts 15 years ago, but we’re convinced that there’s still a lot to learn about what we can do with that data. Halmstad University has expertise in machine learning that means we can develop an even better control and monitoring system. At the same time, we can give the university evidence that their theories work in the real world, Jonas Klang says.

Valuable cooperation between companies and academia

Boris Ahnberg, Solution Architect at Toyota Material Handling, is one of the initiators of the partnership with Halmstad University. He says that the expertise on algorithm-driven knowledge systems has been a driver of collaboration since 2015.

–I think it’s incredibly valuable to work with Halmstad University and with KTH and Linköping University. Students and universities can learn from our reality while we need access to their cutting-edge expertise. I would also like to see greater cooperation between us and the university’s other partners, he says.

Thorsteinn Rögnvaldsson is Professor of Computer Science and runs the Center for Applied Intelligent Systems Research (CAISR) at Halmstad University. He sees making Swedish industry more competitive using AI and machine learning as an important task.

– Then of course we need to bring out lots of interesting research too. Collaboration with the business sector is exciting because it makes great demands in terms of problem solving. It isn’t just about taking an algorithm and running it. Researchers must look at reality too. I’m convinced that collaboration with our industrial partners both improves the skills of our researchers and boosts the competitiveness of Swedish companies, he says.

Text: Carina Järvenhag