Mechanical Engineering The University of Adelaide Australia

Section Navigation

Robotics Group



School of Mechanical
Engineering

THE UNIVERSITY OF
ADELAIDE
SA 5005
AUSTRALIA

Telephone:
+61 8 8303 5460
Facsimile:
+61 8 8303 4367

Soccer Server - Artificial Intelligence (AI) for soccer playing mobile robots Browsing by Title Strain sensing using a 3D scanning laser vibrometer
Project Picture
Photo of Ben Cazzolato

Ben Cazzolato

Photo of Nicholas P. Baker

Nicholas P. Baker

Photo of Cameron Brown

Cameron Brown

Photo of David Dowling

David Dowling

Photo of Justin L. Modra

Justin L. Modra

Photo of Daniel J. Tootell

Daniel J. Tootell

SON of EDGAR

Ben S. Cazzolato, Nicholas P. Baker, Cameron Brown, David Dowling, Justin L. Modra and Daniel J. Tootell


Keywords: Segway, scooter, control, inverted pendulum, self balancing scooter, segway clone

(Commenced: 01-Feb-2006,Concluded: 01-Nov-2006)

State-Space Control of Electro-Drive Grav-Aware Ride

Self-balancing scooters, like the Segway are often thought to be technological miracles, but they are not actually very hard to build. Building on the successes of the 2005 EDGAR project, this honours project aimed to further develop a self balancing scooter.

The refinements to SON of EDGAR include (but are not limited to):

  • Motors - More powerful motors were sourced and installed.
  • Batteries - Larger batteries were used. In addition, the control system monitored the battery charge and modified the gain of the controller accordingly.
  • IMU - Replaced existing Inertial Measurement Unit with much cheaper solid-state rate gyros and accelerometers, incorporating complentary filters.
  • Encoders - Incremental encoders were added.
  • Steering - Implemented closed loop steering via feedback from the encoders. This approach mitigated the problems that arose when the surface of the scooter was uneven or when one tire is inflated less than the other.
  • Wireless Coms - Wireless comms were added to allow tracking of controller performance.
  • System ID - Experimental system ID was performed to validate analytical models of the system dynamics.
  • Controller - The plant dynamics were derived (with and without rider). With this model available, a State-Space controller was programmed and implemented.


SON of EDGAR Deliverables

Media

Image Galleries: