mActivityClassify – Smartphone Real Time Classification of Activities of Daily Living


A stroke is the condition that occurs when a rupture or blockage of the blood flow to the brain happens. In Europe, Japan and United States, this disease is the third leading cause of death. Each year 750.000 people have a stroke and from them, 200.000 don’t survive. Most of the time, stroke survivors have to live with serious impairments in one or a few parts of the body. The rehabilitation after stroke must be a continuous process of several sessions of physiotherapy and of continuing staff and patients/families/carers education. It easily becomes tedious, repetitive and not motivational but it is extremely necessary for them, to recover their functional independence in activities of daily living.

The choice of a stroke rehabilitation treatment and of methodologies of assessment of patients’ impairments is made according to the therapists’ practical experience and the occurrence of abnormal or unwanted movements (compensations with non-paretic muscles) is only detected by direct observation during physiotherapy sessions. The empiric nature of physiotherapy can benefit from an extra source of information gathered outside a health facility. The exposure of the number of tasks that are indeed performed in patient’s everyday life and the associated difficulty to each one of them can help therapists to tailor rehabilitation sessions that are more appropriate and adapted to each individual patient everyday difficulties.



The goal of this project was the development of a real time classification algorithm using an Android smartphone and an external smartwatch/BLE (Bluetooth Low Energy) tag to classify activities of daily living. The project extended our knowledge in the fields of activity classification. The correct detection of activities of daily living has big potential to be applied to the area of fall prevention. The correct assessment of different activities can also be directly applied in the gamification of rehabilitation exercises.


For any additional information regarding this project, please contact us using the inquiries form.