"Single Triaxial MEMS Walking Dataset"

"This dataset comprises structural vibration time-series data collected to facilitate the development of privacy-preserving pedestrian monitoring systems for smart building applications. The data was acquired using a single triaxial MEMS accelerometer placed on the floor of a controlled indoor room, capturing the mechanical response of the structure to human footsteps. It includes 4 kHz raw acceleration recordings ($a_x, a_y, a_z$) annotated with the ground quadrant relative to the sensor (1-4) at the peak of each step, and the walking direction and sense (A-K) of each group of steps. This collection supports research into non-intrusive occupancy sensing, enabling the validation of algorithms for event detection, signal processing, and direction estimation without the privacy concerns associated with optical or acoustic methods."