Human Weight Compensation with a Backdrivable Upper-Limb Exoskeleton: Identification and Control

Abstract

Active exoskeletons are promising devices for improving rehabilitation procedures in patients and preventing musculoskeletal disorders in workers. In particular, exoskeletons implementing human limb’s weight support are interesting to restore some mobility in patients with muscle weakness and help in occupational load carrying tasks. The present study aims at improving weight support of the upper limb by providing a weight model considering joint misalignments and a control law including feedforward terms learned from a prior population-based analysis. Three experiments, for design and validation purposes, are conducted on a total of 65 participants who performed posture maintenance and elbow flexion/extension movements. The introduction of joint misalignments in the weight support model significantly reduced the model errors, in terms of weight estimation, and enhanced the estimation reliability. The introduced control architecture reduced model tracking errors regardless of the condition. Weight support significantly decreased the activity of antigravity muscles, as expected, but increased the activity of elbow extensors because gravity is usually exploited by humans to accelerate a limb downwards. These findings suggest that an adaptive weight support controller could be envisioned to further minimize human effort in certain applications.

Influence of the Physical Interface on the Quality of Human–Exoskeleton Interaction

“…c) Electromyographic measurements: A band-pass filter (Butterworth, fourth order, [20,450] Hz cut-off frequencies) is applied on the EMG signals [44]. They are then centered, rectified and normalized by the maximal value of the measured muscular activity during the whole experiment for each subject and each muscle [13], [40]. The envelope of the signals is determined by applying a low-pass filter (Butterworth, fifth order, 3 Hz cut-off frequency) as recommended by previous studies [44].…”

“…This JM is produced by the differences between the human elbow and robot elbow positions. These position differences create both angular misalignments as illustrated (they have proven to be particularly impacting when trying to apply an accurate force on the user[40]) or bending effort (τ y in our case). They also induce either a varying distance between the robot elbow and the physical interface (if a slider is present) or shear effort (F x in our case) along the human forearm while moving[11],[37].…”

Fast reoptimization of human motor patterns in non-Earth gravity fields locally induced by a robotic exoskeleton

“…Residual errors will inevitably occur due to inherent limits of the control law. These errors were quantified in a previous paper (48). Although relatively small, slight variations from a true gravitational filed are present and it is difficult to estimate the extent to which it could affect the participants' behavior.…”

“…The first one is that exoskeletons may disturb the human motor behavior even in transparent mode (our 1g condition here), although interaction efforts are minimized. We have quantified the nature of the perturbation for elbow flexions/extensions in previous works and the remaining disturbances were mainly due to the additional inertia caused by the exoskeleton (48,(71)(72)(73). It could be possible to compensate this inertia by using predictive control methods (74).…”

“…EMG data was first filtered (band-pass Butterworth, [20,450] Hz cutoff, fourth order, zero-phase distortion, "butter" function from the "scipy" package), centered and rectified (83). Signals were normalized by the maximum peak value from all trials observed during the experiment for each participant and each muscle (20,48,71). The envelope of the signal was then obtained by filtering the pre-processed signal (low-pass Butterworth, 3 Hz cutoff, fifth order, "butter" function from the "scipy" package) (83).…”