MATLAB file containing data from all participants (N = 6) from training session and evaluation session. For each experimental session, data are presented for each exoskeleton assistance pattern as well as normal walking and unassisted walking. These data include EMG signals, ground reaction forces, inverse kinematics, inverse dynamics, exoskeleton torque, exoskeleton joint angles, and the results of EMG-informed simulations (estimated knee contact force, simulation-adjusted EMG, muscle forces, and simulation reserve torques). All data has been trimmed to a single stance phase of the instrumented (right) leg and averaged across strides for that condition. Forces and moments have been normalized to participant body mass. Additional materials, such as scaled anatomical models, are available upon request.

Feature set and coefficients for early-stance and late-stance regression models to estimate peak knee contact forces from EMG, GRF, and knee angle data. These models were trained on data from Session A from all participants in this study (N = 6). For each participant, all EMG, GRF, and knee contact force training data were normalized to the peak value from walking without the exoskeleton. Early-stance features were taken as the peak value between 15 and 40% of stance; late-stance features were taken as the peak value between 60 and 90% of stance. Input features were standardized to have zero mean and unit variance before model fitting. We provide mean and standard deviation of our training data in this file to scale your input data.