Hövding is a company that develops and sells airbag helmets for cyclists. The product is worn like a collar around the neck, when the cyclist is experiencing an accident, a system detects an unnatural movement pattern and deploys the airbag. To detect movements, two different sensors are used: a gyroscope and an accelerometer. The sensor data is analyzed by an algorithm based on machine learning. The algorithm has been trained on accidents performed by stunt performers as well as non-accident data generated by volunteers. Hövding is continuously developing new methods to improve the algorithm and methods to analyze data gathered from real accidents and tests. This requires different types of tools which often requires heavily customized solutions.
Hövding are interested in artificially generating and visualizing data. We need a system that can efficiently simulate sensor data given a movement trajectory in space. This is a simple problem of calculating a derivative, but the real world application is a more complex task, since sensors are prone to noise and we are interested in realistic visualization as well.
We would like a tool to draw paths in space (with orientation of the sensors) and generate corresponding sensor data from a simulated accelerometer and gyroscope. The project has two main parts: a) generating movement trajectories and b) from the trajectories generating sensor data.
Who are we looking for:
One or two students with a robust coding background and good understanding of MEMS sensors and data analysis as well as some experience from GUI programming. You are expected to communicate your project to non-technical staff and deliver a system with an intuitive interface.