Skip to main content

Table 2 Overview of subtasks

From: Automatic versus manual tuning of robot-assisted gait training in people with neurological disorders

Subtask (affected DOF: evaluation point(s))Reference and measured trajectories, evaluation pointsStiffness profile, reference trajectoryCalculation of errorLower and upper bound
Weight shift (lateral pelvis position: minimal and maximal position)Perr=max(|Pref,e1Pmeas,e1|, |Pref,e2Pmeas,e2|)2.50cm 4.17cm
Lateral foot placement (hip abduction angle: angle at 100% of gait cycle)θerr=θref,e1θmeas,e11.15deg. 1.91deg.
Leading limb angle (hip flexion angle: maximal angle)θerr=θref,e1θmeas,e12.15deg. 3.58 deg.
Trailing limb angle (hip flexion angle: minimal angle)θerr=θmeas,e1θref,e11.75deg. 2.92deg.
Prepositioning (knee flexion angle: angle at 100% of gait cycle)θerr=θmeas,e1θref,e14.29deg. 7.16deg.
Stability during stance (knee flexion angle: maximal angle between 10 and 40% of gait cycle)θerr=θref,e1θmeas,e14.30deg. 7.16deg.
Foot clearance (knee flexion angle: maximal angle)θerr=θref,e1θmeas,e14.52deg. 7.54deg.
  1. Reference (black dotted lines) and measured (orange lines) positions and joint angles (Pref, Pref, θref, θref), assistance profiles (K) and evaluation points (e.g. Pref,e1) that were used to calculate the error are shown. Each figure shows one gait cycle starting with left heel strike at 0%. Weight shift to the right side, abduction and flexion angles are defined positive. The lower and upper bound are thresholds for adjusting the assistance based on the calculated error with the AT algorithm. If the error was lower than the lower bound, assistance was decreased. An error larger than the upper bound led to an increase in assistance and in other cases the assistance remained constant (see also Fig. 1)