"IMU Based Human Motion Dataset for Full Body Kinematics Reconstruction"

"Each participant was fitted with a modified Conventional Gait Model comprising 64 markers, each 9.5mm in diameter. IMU sensors were strategically placed on their feet, wrists, pelvis, and head. Data was captured using a twelve-camera optical motion capture system (MTX160, VICON, Oxford, UK) and IMU sensors (ZurichMove, Zurich, Switzerland), sampled at 200 Hz. The IMUs featured accelerometers (range: \u00b116g) and gyroscopes (range: \u00b12000\u25e6\/s). The synchronisation between the systems was achieved through an external TTL signal linked to VICON. TThe dataset contains recordings from 60 participants divided into two age groups. The young cohort consists of 36 participants (20 male, 16 female), while the elderly cohort includes 24 participants (13 male, 11 female). Data are organised into training, validation, and test subsets. The training and validation sets include 32 young participants and 20 elderly participants, providing 704 minutes and 440 minutes of walking data, respectively. The test set contains 4 participants from each age group, with 24 minutes of data per group. In total, the dataset contains approximately 1,192 minutes of motion recordings, corresponding to over 14.9 million synchronised IMU and motion capture sample points recorded at 200 Hz."