GPS-based Life-space Analytics of Older Adults
Sayeh Bayat (1,4), Naglie Gary (2), Mark Rapoport (3), Elaine Stasiulis (2), Alex Mihailidis (1,4)
1. Institute of Biomaterials and Biomedical Engineering, University of Toronto
2. Baycrest Health Sciences
3. Sunnybrook Research Institute
4. Toronto Rehabilitation Institute
Mobility is a fundamental factor in healthy aging. It is an important concept from both individual quality-of-life and public health perspectives. Out-of-home mobility is frequently measured in terms of life-space, defined as the spatial area through which a person moves. Life-space is traditionally self-reported using questionnaires or travel diaries and is thus subject to inaccuracies.
We have developed a comprehensive and autonomous GPS-based mobility model for older adults that: (1) extracts individuals’ more interesting destinations (i.e. points of interest), (2) infers activity types conducted at the POIs, and (3) creates GPS-based spatiotemporal mobility attributes and computes life-space score.
In the pilot trial, ten cognitively intact older adults carried a GPS device when traveling outside their homes for 4 weeks. Participants recorded the details of their trips in a travel diary. Participants’ stop points were extracted from the collected GPS trajectories. Then, their stop locations were clustered using DBSCAN algorithm to find their points of interest (POIs). Next, Google API places was used to infer the types of activities the participants conducted at their POIs. Finally, spatial and temporal features of mobility, including area, perimeter, and frequency of trips, were computed from the GPS trajectories. Moderate agreements were observed between the results from the algorithms and the travel diaries for stop locations, points of interest, and activity types. Strong correlation was observed between participants’ scores on life-space questionnaire and their GPS-based life-space scores.
The new GPS-based life-space construct shows much promise in objectively and accurately measuring life-space, which is valuable for long-term monitoring of older adults’ out-of-home mobility behaviours. It can also be used as guidance for care-partners of older adults with cognitive impairments such as dementia.