Descriptor: Solar and Kinetic Energy Harvesting Dataset for Human Activity Recognition (SKEH-HAR)

This paper presents the multimodal Solar and Kinetic Energy Harvesting dataset for Human Activity Recognition (SKEH-HAR), collected to support research in self-powered activity recognition and wearable systems. The dataset was collected using custom-built wearable devices equipped with solar cells and kinetic energy harvesters, alongside accelerometer sensors. It includes data from 40 adult and healthy participants performing 5 common daily activities, including walking, running, sitting, standing, and going upstairs/downstairs, in both indoor and outdoor environments. The SKEH-HAR dataset captures energy harvesting signals across diverse light and motion conditions, enabling not only the analysis of energy availability but also supporting HAR based on energy harvested signals. This unique combination enables novel investigations into sustainable HAR systems and adaptive sensing strategies based on energy availability. The dataset is expected to serve as a valuable resource for research community working on activity recognition, low-power and batteryless sensing, and self-powered wearable technologies.