StrengthSense: A Dataset of IMU Signals Capturing Everyday Strength-Demanding Activities

TL;DR

StrengthSense addresses the lack of IMU-based datasets for strength-demanding activities by aggregating 8.5 hours of multi-sensor IMU data from 29 healthy subjects using 10 IMUs across 11 strength-demanding activities and 2 non-strength controls. The dataset is annotated with video labels and validated through synchronization checks and IMU-derived joint-angle estimates compared against video-based references using a Madgwick-filter orientation pipeline. The paper details data collection, synchronization, pre-processing, and validation, highlighting the potential for improved cross-subject robustness, sensor-placement optimization, continual learning, and health monitoring applications. Overall, StrengthSense provides a rich, open resource to advance activity recognition, biomechanical analysis, and wearable-sensor research in real-world strength-demanding scenarios.

Abstract

Tracking strength-demanding activities with wearable sensors like IMUs is crucial for monitoring muscular strength, endurance, and power. However, there is a lack of comprehensive datasets capturing these activities. To fill this gap, we introduce \textit{StrengthSense}, an open dataset that encompasses IMU signals capturing 11 strength-demanding activities, such as sit-to-stand, climbing stairs, and mopping. For comparative purposes, the dataset also includes 2 non-strength demanding activities. The dataset was collected from 29 healthy subjects utilizing 10 IMUs placed on limbs and the torso, and was annotated using video recordings as references. This paper provides a comprehensive overview of the data collection, pre-processing, and technical validation. We conducted a comparative analysis between the joint angles estimated by IMUs and those directly extracted from video to verify the accuracy and reliability of the sensor data. Researchers and developers can utilize \textit{StrengthSense} to advance the development of human activity recognition algorithms, create fitness and health monitoring applications, and more.

Figures (7)

• Figure 1: Illustration of Activity #3 at different stages: (a) Begin, (b) Middle, and (c) End.
• Figure 2: Illustration of Activity #7 at different stages: (a) Begin, (b) Middle, and (c) End.
• Figure 3: Placements of the IMU9 Sensors on human body. The orientation of the sensors is such that the X-axis points downward, the Y-axis points to the right, and the Z-axis extends out of the plane of the paper towards the observer. However, the sensors on the upper arm and wrist do not follow this orientation rule. Instead, for these sensors, the Z-axis points out of the arm towards the environment, the Y-axis points to the front of the participant on the right arm and vice-versa. Notice that the sensors in use follow a left-hand coordinate system.
• Figure 4: Experimental setups: site, stairs, and slope configurations.
• Figure 5: Protocol of the data collection.