Elevation Tracking Using Wearables and Machine Learning
Loading...
Date
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The increased integration of technology in emergency management situations has enhanced communication among first responders during critical incidents. Current initiatives in the commercial health and fitness tracking spheres aim to improve indoor location tracking, but these techniques have not been applied to emergency scenarios due to a lack of precision in elevation. While most commercial wearable devices can classify lateral movement with precision, the z-axis poses a challenge for these devices and requires increased precision for high-stress situations. The objective of this research is to develop a compact wearable device, approximately the size of a phone and worn on the bicep, that uses machine learning to accurately classify elevation. This implementation uses a suite of sensors including absolute orientation, acceleration, pressure, temperature, and heart rate sensors. The integration of these sensors aims to facilitate precise classification of activities linked to elevation gain for use in emergency scenarios and general fitness scenarios. An example of the emergency scenario use case is a firefighter in a burning building. If a firefighter passes out within a burning building, this device can notify a fire captain and determine the exact location of the emergency responder with improved elevation accuracy. This device also improves the general fitness example by allowing for more precise statistics on floor climbed data and distinguishing automatic versus manual movement. Notably, existing research in this domain lacks comprehensive coverage of edge cases, particularly in distinguishing between manual and automatic activities during elevation changes. The primary insight for this prototype is floors climbed. The collaborative efforts of the Hardware and Power Supply Lead, Microcontroller and Database Lead, Machine Learning Lead, and Android Application Developer are integral to the successful realization of this solution. Overall, this research contributes to advancing the field of wearable technology by enhancing the accuracy of elevation change classification and expanding the capabilities of fitness and health monitoring devices.
Description
Keywords
machine learning, microcontroller, elevation tracking