Phone Pocket Data

This is the training and test data for the Phone Pocket Mode of WearMoCap. It contains ground truth arm poses from an optical motion capture system and the smartwatch sensor data. Subjects wore a smartwatch on their left arm, a phone in their pocket, and were free to move around. Data collection is described in detail in our technical report:

WearMoCap: multimodal pose tracking for ubiquitous robot control using a smartwatch

Data columns

gt_hand_orig_cal_x,y,z: Ground truth hand position relative to the hip
gt_larm_6drr_cal_1,2,3,4,5,6: Ground truth lower arm orientation relative to the hip. Denoted in the 6-dimensional rotation representation which takes the first two columns of th e rotation matrix (details in our technical report).
gt_larm_orig_cal_x,y,z: Ground truth lower arm origin (elbow) position relative to the hip.
gt_uarm_6drr_cal_1,2,3,4,5,6: Ground truth upper arm orientation relative to the hip.
gt_uarm_orig_cal_x,y,z: Ground truth upper arm origin (shoulder) position relative to t he hip.
gt_hips_yrot_cal_sin_cos: Ground truth indicator of body orientation. These are the sine and cosine of the hip rotation around the y-axis (up).
sw_dt: time since the previous smartwatch observation
sw_h, sw_m, sw_s, sw_ns: time stamp of current smartwatch observation in h,m,s,ns
sw_gyro_x,y,z: Smartwatch gyroscope measurements
sw_lvel_x,y,z: Smartwatch linear acceleration multiplied by dt
sw_lacc_x,y,z: Smartwatch linear acceleration measurements
sw_grav_x,y,z: Smartwatch gravity sensor measurements
sw_6drr_cal_1,2,3,4,5,6 : Calibrated smartwatch orientation estimate in 6-d orientation representation
sw_pres_cal : Calibrated smartwatch atmospheric pressure readings
ph_hips_yrot_cal_sin_cos: Sine and cosine of the estimated phone rotation around the global y-axis (up) as a proxy for body orientation.