Speed estimation Algorithmics

The project

The Netherlands is one of the best performing countries in speed skating. As one would expect, the current level of training is very high among Dutch speed skaters. There is a lot of data readily available, mainly by using IMU-sensors (Inertial Measurement Unit) while training. One of the areas that could be further expanded upon is speed tracking using these sensors. Currently it is not possible for an accurate speed estimation to be made at any given time. This is because most of the training is done indoors where GPS-services are unreliable.
A solution for this problem is our Speed Estimation Algorithm. This algorithm uses real-time skate data provided by the IMU, this data is then passed through a kalman-filter and will result in an accurate speed estimation at any time that the skater is on the ice. This data will then be presented to the user in the form of a graph, and can be further developed into a Java application to be used real time on mobile devices on the skating rink. This algorithm would allow the user to have a greater and much more specific insight in their performance, which would allow for more accurate training and thus enhancing the performance.

The customer

Our customer is Jeroen van der Eb. Jeroen has been working with skaters and their IMU data to give an insight into their performances and has asked us to take a look into developing an algorithm for accurate speed estimation. Communication went via Microsoft Teams and WhatsApp. Using WhatsApp meant that our communication was quite informal and accessible. Meaning that we could send quick updates on our progress over WhatsApp, and meet for a more in depth discussion via Microsoft Teams. At the start Jeroen warned us that this was quite a difficult subject and that we may not be able to complete the whole java application. Despite this, our communication was clear and we mainly focused on what we could do, than rather focus on what we couldn't do.

"Working with a 'real' customer brings forth a sense of responsibility."
The team

Our team consists of seven people. three of them are studying bioinformatics, three of them are studying computer science and one of them is studying physics. So our team was quite large and pretty diverse. The background of our project did not align fully with the background of most of our team members. The project was focused mainly on the physics side of the calculations (how can you calculate the speed given certain variables). This meant that the team member with a physics background had to do a lot of explaining, and the other team members had to do a lot of study/reading to learn about these 'new' subjects. Due to the coronavirus, all of our meetings were online, and were held via Microsoft Teams. This worked pretty well, although meeting in person is preferred (but sadly, not possible).

The technologies