Theses Doctoral

Modular Cable-driven Leg Exoskeleton Designs for Movement Adaptation with Visual Feedback

Hidayah, Rand

Exoskeletons for rehabilitation commonly focus on gait training, despite the variety of human movements and functional assistance needed. Cable-driven exoskeletons have an advantage in addressing a variety of movements by being non-restrictive in their design. Additionally, these devices do not require complex mechanical joints to apply forces on the user or hinder the user's mobility. This accommodation of movement makes these cable-driven architectures more suitable for everyday movement. However, these flexible cable-driven exoskeletons often actuate a reduced number of actuated degrees-of-freedom to simplify their mechanical complexity. There is a need to design flexible and low-profile cable-driven exoskeletons to accommodate the movement of the user and be more flexible in their ability to actuate them.

This thesis presents cable-driven exoskeleton designs that are used during walking and or squatting. These exoskeletons can be reconfigured to apply forces that are appropriate for these functional tasks. The three designs presented in this thesis are non-restrictive cable-driven designs that add minimal weight to the user. The first design shown is the cable-driven active leg exoskeleton previously developed by the Robotics and Rehabilitation Laboratory (C-ALEX, 10kg). The second and third designs are novel cable-driven architectures: (i) the modular C-ALEX (mC-ALEX, 3kg) and (ii) the soft C-ALEX (SC-ALEX, <1kg). A preliminary evaluation of the latter two devices was performed, and the results of these studies are presented to better understand the limitations and abilities of each design. The functionalities added to the latter two designs include the ability to reconfigure the robot's cable routing and attachment geometry, allowing the devices to apply torques through cables in the non-sagittal plane. These features will enable the robot to assist in tasks other than gait while still using the original C-ALEX design methods. Another feature added to the exoskeleton controller is to allow visual feedback through an Augmented Reality headset (the HoloLens) to incorporate visual feedback during tasks better. This feature is currently missing from the rehabilitation field using exoskeletons.

The effects of using the C-ALEX with post-stroke participants were carried out to ascertain the efficacy of using a cable-driven system for gait adaptations in persons with gait impairments and compare their effectiveness against rigid-linked exoskeletons. The C-ALEX was assessed to induce a change in the walking patterns of ten post-stroke participants using a single-session training protocol. The ability of C-ALEX to accurately provide forces and torques in the desired directions was also evaluated to compare its design performance to traditional rigid-link designs. Participants were able to reach 91% ± 12% of their target step length and 89% ± 13 % of their target step height. The achieved step parameters differed significantly from participant baselines (p <0.05). To quantify the performance, the forces in each cable's out-of-the-plane movements were evaluated relative to the in-plane desired cable tension magnitudes. This corresponded to an error of under 2Nm in the desired controlled joint torques. This error magnitude is low compared to the system command torques and typical adult biological torques during walking (2-4%). These results point to the utility of using non-restrictive cable-driven architectures in gait retraining, in which future focus can be on rehabilitating gait pathologies seen in stroke survivors.

Visual and force feedback are common elements in rehabilitation robotics, but visual feedback is difficult to provide in over-ground mobile exoskeleton systems. A preliminary study was carried out to assess the effects of providing force-only, force and visual, or visual-only feedback to three independent groups, each containing 8 participants. The groups showed an increase in normalized step height, (force and visual: 1.10 ± .13, force-only: 1.03 ± .23 visual-only: 1.61 ± .52) and decreased normalized trajectory tracking error (force and visual: 42.8% ± 23.4%, force: 47.6% ± 18.4% , visual-only: 114.2% ± 60.0%). Visual normalized step height differed significantly from force and visual and force-only normalized step height (p<0.005). Lap-wise normalized tracking error differed significantly ($p < 0.005$) within participants.

The mC-ALEX and the HoloLens were used to test the effectiveness of robot force feedback compared to visual feedback with a squat task. The squat task aimed to have the user reach targets of 25%, 75%, and 125% of baseline squats depths through each feedback modality. The kinematic and foot loading effects were considered to establish the differences in user behavior when receiving both types of feedback. The results show that visual feedback has lower errors from targets with similar lower variability in user performance. The force feedback changed joint flexion profiles without changing foot loading biomechanics. When looking at the sessions in sequence, both feedback modalities reduced depth error magnitudes further along with the sessions time-wise. This is the first study where augmented in-field-of-view visual feedback and robotic feedback are used with the aim of changing the kinematics of a squatting task.

Overall, this thesis contributes to expanding the capabilities of cable-driven exoskeletons in lower limb rehabilitative tasks. Three designs are evaluated to understand their on-user performance, with the latter two devices being novel designs. The devices are used in protocols that include visual feedback to ascertain their effects on movement adaptation through the two feedback modalities.


  • thumnail for Hidayah_columbia_0054D_16858.pdf Hidayah_columbia_0054D_16858.pdf application/pdf 42.6 KB Download File

More About This Work

Academic Units
Mechanical Engineering
Thesis Advisors
Agrawal, Sunil K.
Ateshian, Gerard Agop
Ph.D., Columbia University
Published Here
October 6, 2021