Background

Background on data and errors, activity prediction, as well as client geoprivacy.

Accelerometer data and errors

Accelerometers are used to measure the time-varying differences in the movement of body mass known as human physical activity. Measurements are made with regard to the intensity, duration, and frequency of physical activity. Accelerometers have evolved over the years. Amongst many factors, the size, weight, battery life, and accuracy of accelerometers have improved with technological advances. Despite technological advances, some uncertainties remain. For example, extremely high count activity values are indicative of an accelerometer malfunction caused by voltage signal saturations. Another example would be the issue of inaccuracies caused by the position of an accelerometer on the body of a study participant as set forth in our introduction section. These skew results and threaten data accuracy.

GPS data and errors

GPS provides geolocation in the form of latitude and longitude coordinates alongside time information. It is important to truly understand how GPS works in order to analyze the errors from the GPS. A major contributor to GPS error is the inconsistency of atmospheric conditions which contributes to multipath effects where signals are deflected or delayed. Another major contributor to GPS error is the uncertainty of GPS coordinates when a limited amount of satellite signals are accessible. Furthermore, GPS receivers can create positional errors depending on the quality of its components.

Importance of activity prediction

With regard to the error introduced by the free movement of accelerometer devices, a solution rises in the form of activity prediction by combining accelerometer and GPS data. A determination must be made as to whether the accelerometer data should be attributed to active or sedentary behaviour. This can be done by examining the relationship between the accelerometer data and the GPS coordinate data. For instance, if we notice increased accelerometer data knowing the person is physically moving around based GPS coordinate data, we can easily make a positive determination. At the same time, if we notice increased accelerometer data knowing the person is in the same spot based on GPS coordinate data, we can easily make a negative determination.

Client data and geoprivacy

As these accelerometer and GPS technologies advance, so does the volume of detailed data being collected. Where the protection of privacy of individuals data comes from is of great importance, especially if there is a need to release precise geographic information (such as in healthcare research fields). Since data anonymization is critical for privacy protection, there is a tension between attempting to comply with privacy policies while attempting to prevent the loss of geographic information. As the data for our project contains sensitive information such as the latitude and longitude values of the study participants, geoprivacy is important. This is especially exuberated by the fact that our study participants are children. SERO will employ data anonymization techniques to safeguard their sensitive information, especially for the map of our study area showing the locational movements of study participants.