Finally!!! After sooo many setbacks, our preliminary research on human activity recognition (HAR) in extreme environments, specifically for mountaineers is out. Have a look at https://www.mdpi.com/1424-8220/23/3/1416.
For this paper we wanted to investigate factors associated with the placement, number, and combination of accelerometer sensors for accurately identifying activities in these high-risk scenarios.
We found that the use of mountaineering boots and a 12 kg rucksack had a significant effect on classifier accuracy (just look at the image and confirm how red (==bad) the classifiers get when these factors are not considered!) and that using two accelerometer sensors placed on the hip and right tibia provided the most accurate classification of simulated mountaineering activities.
Thanks Steve for the hard work and Sijung for all the help along the way.
The follow-up paper is on its way out! Stay tuned!
Ward, S.; Hu, S.; Zecca, M. Effect of Equipment on the Accuracy of Accelerometer-Based Human Activity Recognition in Extreme Environments. Sensors 2023, 23, 1416. https://doi.org/10.3390/s23031416