Student start-up improves electric mobility with AI
COMPANY PORTRAIT: Most start-ups want to be noticed. The Innovation District Copenhagen-based start-up Lumina, hopes that those who benefit most from their product will never know they have been helped. Lumina is using artificial intelligence to improve the reliability of charge points for electrical vehicles. Barely a year into their journey, they are already piloting their solution with four international operators. By Jes Andersen.
A search for real-world data-rich challenges
Start-up advisors always tell would-be entrepreneurs to solve real-world problems. Lumina’s four co-founders took this advice to heart. With three computer scientists and an interaction designer on the team, they asked industries ranging from hospitals to music venues for data-rich challenges.
We were looking for a big enough problem with a big enough data-set. What really surprised us, was that all the companies we got in touch with, were incredibly open about their challenges”: William Heuser, Partner Lumina, Msc Data Science, UCPH.
One in ten visits could yield no charge
Your average Electrical Vehicle charge point is only 85 to 90 percent effective. For an electric car driver, this means that one in ten visits to the charger may result in frustrations ranging from technical difficulties to no power for the trip. This provided the type of challenge that Lumina’s founders were looking for. 10 to 12 million logs a day from every charge point provided the data.
No one would ever accept ten percent down-time at a traditional petrol pump. To solve this problem, we got to work building a machine-learning system, that can read and analyse these logs. We have also built a dashboard making it easy for operators to understand the problems and we have ended up with a system, that significantly increases reliability”: Mikkel Weikop, Partner, Lumina, BA Visual Communication, Interactive Design, DMJX
From real-time warning to prediction tool
Machine learning excels at discovering patterns in huge sets of data. Lumina’s artificial intelligence can discover problems in real-time. As it crunches ever more charge point logs, it starts identifying what happens before a technical glitch. It starts predicting faults.
One of our pilot-customers says, that our system saves them a lot of time, because they used to read these logs manually. We also hope to help them save on technicians, because a lot of these problems can be solved with a remote reset, if you catch them before they happen”: Sofus Laub Erdal, Partner Lumina, Msc Mathematical Modelling & Computation, DTU
Benefitting from start-up community
The company is currently located in the University of Copenhagen innovation centre UCPH Lighthouse. Here, they have benefitted from free office space, entrepreneurship training and mentoring. They recently won access to the Danish non-profit innofounder programme. This will help them further develop and improve the technology. Their long-term dream is to remain an R&D company within artificial intelligence.
The EV-charger market is really niche. We need to sell our solution to those few actors who have thousands of charger parks. Fortunately we are entering a market, that is expanding at incredible rates. And even better. The data from the charge points are all in the same format, so our solution can be used by any charge point operator in the world”: Andreas Patscheider, Partner Lumina, Msc Business Analytics, DTU.
The power you need, to go where you want to go
If the four pilot projects are any indication, this start-up should go far. The latest addition is the global Danish company Spirii. Another of Lumina’s initial collaborators successfully used the promised reliability to raise capital. Meanwhile the drivers who get more dependable service may never know, that a Danish student start-up is helping them get the power they need, to drive where they want to go.