Real-Time Full-Body Gesture IMU Dataset

This dataset contains IMU recordings of participants performing full-body gestures for mobile real-time gesture recognition.10 participantsSeven isolated gesture classes: idle, right punch, left punch, jump, right kick, left kick, duckParticipant demographics: 6 male, 3 female, 1 not available; age range 20-29 years for 9 participants with available metadata; height range 155-200 cm for 9 participants with available metadata413 isolated-gesture recordings35 continuous-sequence recordings with intended gesture sequences and sample-wise labels448 recording files in CSV formatFour Silicon Labs Thunderboard Sense 2 boards per recordingSensor positions: left arm, right arm, left leg, right legSensor configuration: accelerometer range +/-2 g, gyroscope range +/-250 deg/s, sampling rate 50 HzRecorded in July 2019 by Connor Sachs under supervision of Philipp Niklas MüllerSee the included README.md for more information.