Assistance or collaborative robots aim to preserve the health and well-being of people, in particular by promoting interactions during phases of fragility. There are many ways to interact with robots, but this project only focus on human motion. The goal of the BioMotion project is to find the modalities of representation of movement allowing to extract physical and cognitive information.
A first problem is to collect physical information and, in a more original way, cognitiv information of the state of the person through the analysis of the person’s motor strategies. We wish to explore this through the decomposition of human movement. However, this observed movement accumulates a set of information: what is due to the task (e.g. the walk), what is intrinsic to the person (e.g. the gait), what is a matter of ‘mood’ of the person (e.g. emotion, calm, fatigue).
The expected results can enable us to:
- to make the identification of primary movements of the person more robust by removing some of the personal variations of the reference signal;
- to extract characteristics of the person’s gait that could be exploited for behavioural analysis.
People involved:
- Jessica Colombel (PhD Student)
- François Charpillet (thesis advisor)
- David Daney (thesis advisor)