Comparing Clothing-Mounted Sensors with Wearable Sensors for Movement Analysis and Activity Classification

Abstract

Inertial sensors are a useful instrument for long term monitoring in healthcare. In many cases, inertial sensor devices can be worn as an accessory or integrated into smart textiles. In some situations, it may be beneficial to have data from multiple inertial sensors, rather than relying on a single worn sensor, since this may increase the accuracy of the analysis and better tolerate sensor errors. Integrating multiple sensors into clothing improves the feasibility and practicality of wearing multiple devices every day, in approximately the same location, with less likelihood of incorrect sensor orientation. To facilitate this, the current work investigates the consequences of attaching lightweight sensors to loose clothes. The intention of this paper is to discuss how data from these clothing sensors compare with similarly placed body worn sensors, with additional consideration of the resulting effects on activity recognition. This study compares the similarity between the two signals (body worn and clothing), collected from three different clothing types (slacks, pencil skirt and loose frock), across multiple daily activities (walking, running, sitting, and riding a bus) by calculating correlation coefficients for each sensor pair. Even though the two data streams are clearly different from each other, the results indicate that there is good potential of achieving high classification accuracy when using inertial sensors in clothing.

Towards Human Stress and Activity Recognition: A Review and a First Approach Based on Low-Cost Wearables

“…Tables 1 and 2 summarize a review of recent works, sorted by year of publication, related to human activity recognition through wearables, while Figure 1 depicts some relevant conclusions. Overall, most of the studies detected a variable number of human activities, from one to nine, including the identification of sports [14][15][16], activities of daily living (ADL) [17][18][19], also featuring house activities [20] such as washing, shelving items, sweeping, ironing, vacuuming, driving, etc. In most studies, the typical attempted activities were: walking, jogging, standing, sitting, lying, squatting, going upstairs/downstairs, running, among others.…”

Ecological inference using data from accelerometers needs careful protocols

“…It is also germane to consider that tag attachment stability may change over time in longer-term deployments. These issues have long been recognised in the wearable sensors industry for humans (Jayasinghe et al, 2019). Consequently, we cannot, in good faith, compare VeDBA or wingbeat amplitudes of tropicbirds between seasons although the wingbeat frequency will be unaffected.…”

“Chatty Devices” and Edge-based Activity Classification

“…For example, Tabia et al [47] identified human activities such as running, jogging and running by using Naıve Bayes and Support Vector Machine (SVM) classifiers. Jayasinghe et al [22] conducted experiments using a combination of clothing mounted sensors and wearable sensors for movement analysis and activity classification with a focus on running, walking, sitting and riding in a bus. By calculating correlation coefficients for each sensor pair, they suggest that even though the two data streams have some notable differences, results indicate high classification accuracy.…”