Physical activity level estimator (ALE) (ended)

ALE: Physical Activity Level Estimator

On a growing scale, we witness the use of mobile phones for diverse daily life activities, e.g., entertainment or education, and therefore we also research methods for such a phone to provide pervasive services for wellness and healthcare.

Towards this end, we assess the phone’s feasibility to unobtrusively, continuously and in real-time track its user’s physical activity intensity and quantity per day based on the World Health Organization (WHO) recommendations.

The identified gap in the existing approaches is a lack of accurate monitoring of the physical activities performed along with persons daily life activities. The existing approaches are also providing insufficient evidence-based feedback to the user to encourage the increase of such a physical activity. Finally, we did not found systems that give feedback according to the WHO recommendations.

Towards this end, we developed an Android OS application called Activity Level Estimator (ALE) that enable to accurately track user’s daily walking activities, computes the intensity and the quantity of the physical activity and that provides feedback to the user, characterizing him as a sedentary or not, based on the above-mentioned recommendations from the WHO.

ALE is a prototype application made for an Android OS mobile phone, running unobtrusively, continuously and in real-time as a phone’s background application without using a dedicated external device. We focused on the walking activity only because it is the most probable activity that a user can perform along a day and also because it is the most popular kind of moderate activity to reach the physical activity goal based on the WHO recommendations. Lee and Buchner concluded that because of its popularity, the walking activity has substantial importance to public health.

ALE estimates the user’s physical activity intensities based on the user’s movement patterns using the phone’s built-in 3D accelerometer. The acceleration is monitored and transformed into intensities during a defined time interval, assuming that the phone is worn in the user’s pocket, near the user’s hip. The intensity is computed in METs in order to match the activities’ level as defined by the WHO (low, moderate, vigorous). Then, ALE counts and classifies times spend by the user per these levels of activities.

The ALE application was developed to be used on a regular day and can display to the user his/her daily results and weekly results. To measure the walking activity intensities, we let choose the user to activate it upon own will. A daily result could be enough to represent a user’s physical activity status during a regular typical day and deriving a typical week’s overview. The user can understand and evaluate his physical activity habits and compare them to the WHO recommendations while using the ALE application for one or more days or throughout the whole week. After some time of using ALE, the user will be able to estimate and adapt the effort needed per day or per week to reach the WHO recommendations.

Calibration and evaluation

A study was conducted in collaboration with the Centre d’Analyse et de conseil à la Santé et aux Activités Physiques et Sportives at the University of Geneva.

The aim of this study was to calibrate our application and assessed the accuracy of ALE against a Gas Analysis System, a “gold standard”. The study was conducted with 12 participants.

Our results showed an overall average accuracy of 90.06% per minute for walking.


The following short movie presents the utilization of ALE and also a set of preliminary results acquired during the last user study.