Often, high accuracy activity recognition can be performed using relatively simple methods, such as through the use of sensors like accelerometers and gyroscopes. This means that activity segmentation, meaning the extraction of continuous sequences of a specific activity, is readily available for many applications. However, so far there has been relatively little work into the detailed analysis of these segments, in spite of the benefits information about the way people perform could offer for many different domains and applications, in particular medicine.
The main aim of this PhD project, therefore, is to develop and investigate suitable methods that will allow for the quantifying of differences in motor performance, as observed through the use of body-worn and pervasive sensors. Working closely with people with Parkinson’s Disease, we intend to apply and evaluated these methods in order to develop suitable outcome measures for motor deficiencies such as Bradykinesia and Dyskinesia.