Activity recognition in ecologically valid environments using body-worn sensing

Objective information about human movement and behaviour is crucial to many fields such as medicine, rehabilitation and pervasive computing. Particularly the application of automated activity recognition systems in the private home holds potential to improve the daily life of large parts of the (ageing) population. However, the challenges of sensing and assessment of human movement in the private home and other, ecologically valid environments are significant. In this talk I will introduce a novel approach that addresses these issues and shows promising performance in the assessment of Parkinson's disease.​