Overground vs. treadmill: where is my horse running? Context detection based on IMU data
Problem Statement:
Horses locomotion is evaluated either in-hand overground or on a treadmill, using for example wireless inertial measurement unit (IMU) systems. Treadmill evaluations enable the collection of highly standardised data at a constant speed, which is valuable especially in the research context. To better automate our gait analysis system, we would like to develop overground vs. treadmill locomotion classification models. Using different feature selection methods, the differences between overground and treadmill signals can then be evaluated.
Task:
In this assignment, your task will be to develop one feature-based machine learning model (e.g. Support Vector Machine, SVM) and one deep learning model (e.g. Long-Short Term Memory, LSMT) that can classify overground and treadmill locomotion from walk and trot (labelled) equine data.
You will then compare the features and signals used by the models to better understand which ones are the most useful for overground vs. treadmill locomotion classification, and thus how the locomotion is affected by either condition.
Work:
40% theory, 40% Implementation, 20% Writing
Contact:
Jeanne Parmentier (j.i.m.parmentier@utwente.nl)