IoT of active and healthy ageing: cases from indoor location analytics in the wild

Abstract

ACKNOWLEDGMENTThis work was supported in part by the UNCAP Horizon 2020 project (grant number 643555), as well as, the business exploitation scheme of the ICT-PSP funded project LLM, namely, LLM Care which is a self-funded initiative at the Aristotle University of Thessaloniki (www.llmcare.gr). 3 AbstractRecently much research has been conducted on early detection of cognitive and physical status deterioration in elderly adults. Primarily the focus is on gait analysis methodologies exploiting average speed, however this presents an issue when used for context aware applications. Additionally data capture tends to be in short bursts over a long period, allowing for localized temporal factors, such as short term injury, to potentially skew measurements. As such this work collects gait and trajectory IoT data from elderly adults in senior homes ("in the wild") over a sustained period of time (1 year). Density based clustering algorithms are then applied to the data to provide long-term insights into how the high density regions change over time. The data is collected, analyzed and made available by the indoor analytics client utilizing available processing resources and delivers the analytics outcome even when it is hosted in hardware with constrained resources. Promising results are obtained from the long-term study, suggesting that this form of evaluation has strong potential in the analysis of cognitive and physical status deterioration.

Clustering Indoor Location Data for Social Distancing and Human Mobility to Combat COVID-19

“…Generally, DBScan [7] has been used for location data that contains regions with scarce data as it can handle noise well. The algorithm was adopted to find high dense regions using aggregated data, to reveal places which are more likely to be visited [29][30][31]. In contrast, clustering in the spatio-temporal dimension is more complex and data are typically expressed in different forms, such as geo-referenced variables, moving objects and trajectories [32].…”