At Futurice we are all about sharing knowledge and offering learning opportunities. This is so much so, that in a recent survey the majority of all our employees rated it as one of the most valuable benefits of working here. We don’t limit this to our own people though, and regularly reach out to the community in the form of meetups, workshops, open source tools, and in this case: student mentoring.
Each of the teams had their own brief and competence area; one focused on AI, and the other on robotics.
The AI team was tasked with assembling an 1:10 scale model race car chassis, loading it up with tech using open source methods, and then having it race around a pre-learned track fully autonomously. This whole thing was kicked off at our race event in Särkänniemi (which was being covered by the show “Teknavi” on Finnish National TV at the time). Although pressure was high, the team stayed cool and even went for a few rides on the park’s prime roller coaster, the Tornado, before heading home 😎
The robotics team, on the other hand, needed to figure out how to find lost things. Oodi, the Helsinki Public Library, has had up to 20,000 visitors per day who can browse over a hundred thousand available books. These books don’t always end up back in the correct shelf though, so the task was to develop a scanning system which can determine where each book is located. Using a robot to automate this task will let the librarians spend less time doing inventory, and let them spend their time on more sophisticated and pleasant tasks.
After a summer of assembling the car chassis and 3D-printing custom PELA Blocks, the Aalto car “Speed Demon” was given the chance to prove itself in an official computer-vision based RC car race at the course gala in Aalto University. For the sake of local Finnish rivalry, we’d also tasked them with the extra challenge of beating Markku, the pinnacle of engineering from the Futurice Tampere office. They not only did this handily but, in true Teekkari style of overdoing things, went on to beat the Futurice Helsinki office cars as well and win the entire competition.
(Left: The first PELA-Block car, Right: The new and improved “Speed Demon”)
Embedded content: https://vimeo.com/375090294/6522f0ce11
(Speed Demon at work, tearing up the track)
Meanwhile at Oodi things were looking dark. Literally. The team had built “RadioHead”, an RFID-reader module for Oodi’s own MiR 200 service robots that could scan entire shelves of books, determining each books exact location and noting at which time it was last “seen”. The platform will enable the MiR robots to autonomously take inventory, roaming the library during the night when guest and staff have already gone home for the day.
(“Duct-tape” iteration of RadioHead, and CAD models of the final prototype)
Embedded content: https://vimeo.com/375090244/194a2430a6
(Library book scanning process in action)
Many thanks to the teams for making these two excellent projects, to all the colleagues who helped with the tech mentoring, and to Oodi for offering both opportunity and tools to contribute to their impressive public service. An insight into Oodi’s future plans for their robots is nicely summarized in this video by resident ICT specialist Romeo Pulli and tech-YouTuber Tom Scott. All of the work is available online and open source (CC BY 4.0, MIT License).
Autonomous Racing Car: Project description | Source code
Library Scanning Robot: Project description | Source code