Javascript is not activated in your browser. This website needs javascript activated to work properly.
You are here

Klas Nilsson | Cognibotics

Klas Nilsson, LTH researcher focusing on robotics, previously took the view that there is too much administrative work involved in commercialisation. "I am not the person to commercialise things. I work for scientific purposes, focusing on developing new solutions and have not paid much attention to who takes care of them. I usually give away my ideas and then move on to the next exciting project."

But when the idea of robots which are capable of independently detecting and identifying their own properties without any external input took shape and was tested for a year, Klas began to consider a patent application. The solution was so basic in all its simplicity that the companies which took part in the project did not completely understand the idea – and were thereby not interested in getting it free of charge for their own patent. 

Technology for increased precision in robotic control

Klas consulted with various experts and they all agreed that the new technology was as simple as it was brilliant. And yet nobody had thought of it before!

Klas had previously tried to patent a few ideas together with companies, got a little money but mostly found it difficult and cumbersome. He had also thought of starting his own company, but dismissed the idea on the grounds that it would mean "too much administration and tedious tasks". This time Klas decided to contact LU Innovation, which he had heard could help with patents.

– I got a quick response and then it was just a matter of getting going. The various parts were explained to me, business models and suchlike.

The technology entails the robot connecting its tool or its joint to a fixed point and performing a stress test on itself.

Take one step at a time

Even though Klas had worked in the industry previously, it was as a development engineer, which can be compared to research work.

– Your whole approach needs to shift its focus somewhat. But LU Innovation did not overload me to begin with, proceeding instead in suitably small steps: this is what we will do with the patent application, what should a business plan look like, what customer categories are there? One question at a time, so that you get into it gradually, while getting a lot of support.

Support and impetus

The plan was not initially to start a company, but it became a natural step as the process and the discussions with business developers progressed. Klas thinks that the company would never have been formed without the impetus provided by LU Innovation. 

– When it dawned on me that there was so much support for the formal aspects, that they provide chairmanship and help with accounting costs – everything I find so boring, then I saw the issue in a whole new light!

LU Innovation was also clear that its task is economic growth, which you do not hear so often in the knowledge-oriented university world. I think it is a good thing that someone has that job, and I appreciate clarity!

A timely idea

The patent application was approved without any amendments, and Klas, with his interest in robotic movement and large networks within the sector, had already determined that there was a market for the solution. At the same time, he started another project focusing on making robots usable and financially available for small companies, which gave Klas further documentation to map customer benefits. 

– My idea is really timely. The market is ready!

Cognibotics was started in 2013 by Klas and five of his research colleagues. The company provides software and technical support to license its method to robot manufacturers, among others. Klas is the team member who meets customers and leads discussions on performance and prices.

– The fact that I am the salesperson in the company really means that I talk with a lot of people who understand and demand technology within my field. We do not talk so much about money, more about the ideas and technology in terms of applications and customer benefits.

Research in the company

Klas underlines that researchers and doctoral students should apply for patents for their ideas more often.

– It is interesting how you describe your idea in a patent application so that the patent becomes sustainable. The Swedish Patent and Registration Office is a good reviewer, and the structure is useful in the thesis writing as well. A patent application also means that you can publish the findings in research publications without sacrificing growth potential.

And if the project is patentable, you can move on to the next step.

– Starting a company is interesting in itself. There are ways of doing it step by step, so that you do not need to expose yourself to a major risk at once. You see whether it seems right or whether to postpone it, if it is a tough market for example. But you should still take the step and see how it feels!

Running a company does not have to entail setting research aside, either. 

– Starting a company does not necessarily mean getting stuck in administrative matters, even in a company it is possible to do research. What we are doing here today is of the same research level as what we did at the department. This is not the end of interesting work duties, it may even be only the beginning!


Facts Cognibotics


Wear and tear, loosening and potential production faults limit precision in the hundreds of thousands of robots currently in use in the world's factories. In order to compensate for the faults, previously it was only possible to purchase more expensive robots or specialised machines, or to conduct regular measurements with expensive and complicated specialist equipment.


Klas Nilsson, researcher in industrial robotics and leader of the research team on Robotics and Semantic Systems within Computer Science, together with five researcher colleagues, developed a new technology which leads to increased precision in industrial robots. The technology entails the robot connecting its tools or its joint to a fixed point and “flexing its muscles”. The robot can then, with its regular sensor, measure strengths and deviations itself. The information is used to identify the properties of the joints and to calibrate the control system so that mechanical faults are not noticeable in the robot's movements.


Cognibotics was founded in 2013. The business concept is to sell the license for the method to robot manufacturers, system integrators, large industrial companies with their own fleets of robots and to companies which work with robot calibration. LU Holding AB is a part owner.