Gait analysis using wearable sensors in Parkinson's disease

Revision as of 16:28, 30 September 2019 by Taha (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Title Gait analysis using wearable sensors in Parkinson's disease
Summary The project aims to develop a machine learning tool for the assessment of Parkinsonian gait in a natural environment
Keywords Signal processing, Machine Learning, Parkinson's disease
TimeFrame Autumn 2019 - Spring 2020
Supervisor Taha Khan
Level Master
Status Open

Gait impairment is an important symptom in Parkinson's disease. The project aims to investigate if wearables such as inertial sensors could be used to perform long-term continuous monitoring of Parkinsonian gait in a natural environment. Data consist of time-series of accelerometer readings and pressure insole sensors recorded from Parkinson patients during their assessment of gait by a doctor at a hospital in Johannesburg South Africa using a standard clinical protocol, as well as, during their walks around the clinic without the protocol in a natural setting. The project will explore and develop novel methods and gait features that could be extracted from the time-series and used for training machine learning models for automatic classification and monitoring of the severity of gait impairment in a natural environment.