FhP-AICOS has been working in the field of monitoring physical activity and activity related metrics, and already has a solution for Activity Level and Fall Risk monitoring integrated in the Smart Companion environment. However, few or no efforts have been made in analysing the data that is generated while users carry the monitoring application. By understanding the users’ behaviour patterns, we can better suggest appropriate intervention strategies and personalized recommendations.
The inertial sensor data that is received by the application is processed and stored in a backend server. This data comprises physical activities, postures recognition and related metrics such as walking, running, sitting, standing and resting time, number of sit-to-stand transitions, gait inter-stride variability, walking speed, number of steps, energy expenditure, distance travelled and number of falls along the day. Among activity data, information from GPS location and points of interested visited by the users are also available. For the purpose of this project, it is expected to retrieve all the useful information from the user’s database and to develop appropriate algorithms for activity patterns recognition based on hourly, daily and weekly records. For each epoch, different findings and analysis should be performed.
In order to accomplish the objectives of this project, it is necessary to develop methods for data representation, data visualization and relevant outcome identification by analysing major trends and patterns in daily routines for each user. Moreover, data from similar users can be crossed to find group patterns and to infer disruptions in daily activity and early signs of physical and cognitive decline. Whenever it is relevant, meaningful recommendations should arise from the data analysis, as for example, when a user has a sedentary behaviour and poor social life, suggestions of outdoor physical activities and social events are useful when crossed with the user’s location.
Author: Mário Ferreira
Type: MSc thesis
Partner: Faculdade de Engenharia da Universidade do Porto