University of Twente Student Theses

Login

Development and evaluation of a safe Fall Controller for a Lower Limb Exoskeleton using reinforcement learning

Vogelzang, Kevin (2020) Development and evaluation of a safe Fall Controller for a Lower Limb Exoskeleton using reinforcement learning.

[img] PDF
6MB
Abstract:Lower limb exoskeletons allow people with paraplegia to walk again. Currently, this walking is rather limited because the pilots have to use crutches to maintain balance and steps are made by following pre-defined trajectories. Additionally, most exoskeletons are used in a clinical or research setting, generally with some safety mechanism in place in case of a disturbance. It is likely that the field of lower limb exoskeletons will continue to develop, allowing for more exoskeletons to be used without any external safety mechanisms. Not having these makes the exoskeleton rely more on a good balance controller to maintain balanced when facing disturbances and a good safe fall controller if this balance cannot be maintained and damage has to be minimised. Due to the assumption that another safety mechanism is in place to catch the human-exoskeleton system, the state of the art fall controllers do very little to actually break the fall, risking injury to the human pilot and damage to the exoskeleton. This work, therefore, aims to create a better controller for minimising damage when falling with a lower limb exoskeleton. This controller needs to lower both the amount of times the head is hit as well as the impact force on and impact velocity of the head if it does hit the ground. This is tested for the forward and backward fall using a three-dimensional, humanoid model.
Item Type:Essay (Master)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:50 technical science in general, 52 mechanical engineering
Programme:Systems and Control MSc (60359)
Link to this item:http://purl.utwente.nl/essays/82748
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page