Date of Award

Spring 2016

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Biomedical Engineering - (Ph.D.)

Department

Biomedical Engineering

First Advisor

Richard A. Foulds

Second Advisor

Sergei Adamovich

Third Advisor

Eugene Tunik

Fourth Advisor

Gerard G. Fluet

Fifth Advisor

Gail Forrest

Abstract

The robotic lower extremity exoskeletons address the ambulatory problems confronting individuals with paraplegia. Paraplegia due to spinal cord injury (SCI) can cause motor deficit to the lower extremities leading to inability to walk. Though wheelchairs provide mobility to the user, they do not provide support to all activities of everyday living to individuals with paraplegia.

Current research is addressing the issue of ambulation through the use of wearable exoskeletons that are pre-programmed. There are currently four exoskeletons in the U.S. market: Ekso, Rewalk, REX and Indego. All of the currently available exoskeletons have 2 active Degrees of Freedom (DOF) except for REX which has 5 active DOF. All of them have pre-programmed gait giving the user the ability to initiate a gait but not the ability to control the stride amplitude (height), stride frequency or stride length, and hence restricting users’ ability to navigate across different surfaces and obstacles that are commonly encountered in the community. Most current exoskeletons do not have motors for abduction or adduction to provide users with the option for movement in coronal plane, hence restricting user’s ability to effectively use the exoskeletons. These limitations of currently available pre-programmed exoskeleton models are sought to be overcome by an intuitive, real time user-controlled control mechanism employing admittance control by using hand-trajectory as a surrogate for foot trajectory. Preliminary study included subjects controlling the trajectory of the foot in a virtual environment using their contralateral hand. The study proved that hands could produce trajectories similar to human foot trajectories when provided with haptic and visual feedback. A 10 DOF 1/2 scale biped robot was built to test the control paradigm. The robot has 5 DOF on each leg with 2 DOF at the hip to provide flexion/extension and abduction/adduction, 1 DOF at the knee to provide flexion and 2 DOF at the ankle to provide flexion/extension and inversion/eversion. The control mechanism translates the trajectory of each hand into the trajectory of the ipsilateral foot in real time, thus providing the user with the ability to control each leg in both sagittal and coronal planes using the admittance control paradigm. The efficiency of the control mechanism was evaluated in a study using healthy subjects controlling the robot on a treadmill. A trekking pole was attached to each foot of the biped. The subjects controlled the trajectory of the foot of the biped by applying small forces in the direction of the required movement to the trekking pole through a force sensor. The algorithm converted the forces to Cartesian position of the foot in real time using admittance control; the Cartesian position was converted to joint angles of the hip and knee using inverse kinematics. The kinematics, synchrony and smoothness of the trajectory produced by the biped robot was evaluated at different speeds, with and without obstacles, and compared with typical walking by human subjects on the treadmill. Further, the cognitive load required to control the biped on the treadmill was evaluated and the effect of speed and obstacles with cognitive load on the kinematics, synchrony and smoothness was analyzed.

Share

COinS