Optimisation of the stability of an assistive lower limb exoskeleton

Abstract

Currently, most patients who have suffered from a stroke are treated through-one-on rehabilitation training administered by a physiotherapist. Powered Lower Limb Exoskeletons (PLLEs) have been designed to bring about advantages over traditional treatment. Although PLLEs have been invented overseas it has done little help in South Africa (SA), apart from sparking interest and showing what could be possible if research and development are put into this technology. It is hoped that this research will help solve gaps in the existing body of literature, from a South African perspective. This thesis focused on the optimisation of stability of a PLLE, aiming to enhance the walking gait and overall performance of the device. The research explored two main approaches: the Linear Inverted Pendulum Model (LIPM) and an optimisation technique utilising a cost function and genetic algorithm. The study involved the use of software in a simulation environment using MATLAB, which allowed for the prediction of hardware requirements and optimisation for stability. The requirements and desirable traits of a stable PLLE were identified through a literature review. The design considerations encompassed a realistic size and weight based on existing exoskeletons. Using existing motors and motor controllers, a test platform was designed to closely represent the simulation model. Although the test platform did not achieve successful walking functionality, its development process yielded invaluable insights and knowledge, contributing to a deeper understanding of the challenges and complexities involved in attaining the desired performance. The simulated results were analysed based on predetermined criteria, including the distance walked within 10 seconds, maximum torque of ankle, knee, and hip joints, torso height for upright walking, and human-likeness of the gait. The findings highlighted the achievement of a stable gait through the LIPM approach while the optimisation technique showcased progressive improvements throughout iterations, leading to a refined walking pattern. This thesis contributes to the field of PLLEs, providing valuable insights for the development of more stable and human-like walking gaits. It is hoped that this outcome will help to form the groundwork for a rehabilitation system which would improve mobility for SCI patients and support healthcare professionals such as occupational therapists.