Published in Science Robotics, this dataset consists of sensor data from a robotic hip exoskeleton and time-synced with ground-truth human lower-limb biomechanics. All data were sampled at 200 Hz.

• 34 subjects, 9 ambulation modes (e.g. level ground, ramp ascent/descent, stair ascent/decent, standing), 3 controllers (e.g. gait phase-based control based on biological torque, unified joint moment controller)

• Angles and Moments computed from Inverse Kinematics
  • Processed sagittal plane hip, knee, and ankle angles
  • Processed hip and knee extension moments
  • Processed ankle plantarflexion moments
• Exoskeleton Sensor Data
  • Actuators position and velocities from encoder
  • Acceleration and gyroscope from IMUs
  • Hip moment estimated by TCN
  • Torque commanded to actuators
  • Torque from the actuators (calculated by multiplying measured current by motor torque constant and gear ratio)
• Gait Phase
  • Gait phase for each leg segmented by toe-off
  • Gait phase for each leg segmented by heel strike
• Ground Reaction Forces
  • Ground Reaction forces for each foot in 3 directions
  • Center of Pressure for each foot in 3 directions