In-Use Estimation of the Yaw-Rate Motion Output of a Wheelchair Using Joystick Inputs

Open Access
Blundi, Kathleen Elizabeth
Area of Honors:
Mechanical Engineering
Bachelor of Science
Document Type:
Thesis Supervisors:
  • Sean N Brennan, Thesis Supervisor
  • Jacqueline Antonia O'Connor, Honors Advisor
  • wheelchair
  • robotic wheelchair
  • assistive technology
  • inertial effects
  • health-monitoring
People with disabilities that impair their ability to walk may become reliant on an electric wheelchair for mobility and independence. This thesis presents an in-use model to predict the yaw-rate of a wheelchair using experimentally-collected joystick data, wheelchair position, and wheelchair orientation to understand how user motion intent compares to actual measured motion output. Experimentation used wheel odometry and on-chair sensors to measure steering inputs and motion outputs. Using the joystick steering inputs, a discrete time-domain model was developed using a Tustin Transform. Then the model was used to perform a least-squares identification of the system dynamics. Finally, the accuracy of the model was verified using experimentally collected data. The model for the predicted yaw-rate is dependent on the joystick inputs as well as the in-use measured yaw-rate. Furthermore, the yaw-rate is controlled by the vertical rotational moment of inertia of the wheelchair and the vertical rotational moment of inertia is based on the overall mass and positioning of the mass of the wheelchair. Therefore, the results show that changes in body mass, body position, and body orientation may be tracked by detecting changes in the vertical rotational inertia of the wheelchair. The application of these results can then be used to determine if a patient is gaining or losing weight over time, requires repositioning assistance, or may require modification of the joystick inputs for increasing a patient’s comfort