Thesis

Shared control between man and machine in brain computer interfaces

Creator
Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2011
Thesis identifier
  • T12919
Qualification Level
Qualification Name
Department, School or Faculty
Abstract
  • An immersive virtual reality electric-powered wheelchair simulator, controlled by a Brain-Computer Interface (BCI) has been developed at the University of Strathclyde. However, fine manoeuvring is difficult due to the limited number of commands that can be given through the BCI. This project aims to incorporate shared control between the user and an intelligent wheelchair controller to provide a better driving experience, while ensuring the safety of the user. A shared control strategy was developed to provide three types of assisting behaviours - Obstacle Avoidance, Collision Avoidance and Wall Following. Reactive navigation strategies were devised with the help of user input from the keypad and information about the environment provided by sonar sensors, infra-red sensors and velocity levels obtained from ground truth. Respecting the importance of user autonomy, as re-iterated in the literature, the user has the power to override the assisting behaviour provided, except in the case of a risk of collision. The shared control technique was evaluated with the help of five healthy volunteers who performed numerous wheelchair driving tasks in manual and shared control modes in the wheelchair simulator. They rated their driving experience using NASA Task Load Index (TLX) scales. In order to study the effectiveness of the shared control behaviours, a secondary task was used to simulate a state of suppressed ability in the able-bodied people. Average workload scores were calculated for all the tasks and revealed that while workload was decreased in the case of shared control for Wall Following behaviour, it was the opposite for Obstacle Avoidance. The inferences made from the NASA-TLX assessment procedure and the observations made during the experiments have provided a useful insight into the limitations of the system and scope for improvement.
Resource Type
DOI
Date Created
  • 2011
Former identifier
  • 834103

Relations

Items