Research

Robotics and Automation

Robotics and Automation UWA

Providing expertise and solutions to spatial awareness and remote autonomous operations of robotic systems. Unmanned and remote autonomous systems and self-directed, maneuverable and interactive robots will allow us to go boldly where no one has gone before.

Representing the nexus between artificial intelligence, computer vision, control robotics and signal and image processing robotics and automation will be instrumental in engineering projects which require unmanned operation of remote systems.

From the remote outback to outer space and from the domestic robot vacuum cleaner to hands-free, self-driven cars, robotics and automation will be pervasive. Our aim is to develop new techniques that enable systems to sense and move within their own environment in either a fully self-directed or human guided fashion:


Problems that we solve

Robotics and Automation will address the challenges and problems of applications that revolve around engineering systems where continuous human presence is either undesirable or impossible, and where tele-operation and automation are desirable or necessary.

Automation of remote operations cannot rely on remote control alone, as is currently the case in a number of remote mining operations. Remote control can only be the first step in any autonomous operation, where gradually more and more operations will be automated and the role of the human operator is being changed from “driver/operator” to “supervisor/monitor” of largely autonomous operations.

Our team brings together a multidisciplinary group of expert researchers in Mechatronics, Computer Vision, Autonomous systems, and Signal and Image Processing. Our work will focus on the following key areas:

Robotics

Group members have been involved with robotics and automation for 40 years. The Robotics and Automation lab has designed and built hardware, electronics, and software for almost 100 driving, walking, swimming/diving and flying robots over the last decade.

Computer Vision

Group members have an extensive experience and an international reputation in this area through the development of theoretical tools e.g., for machine learning and computer vision with applications in various areas including 3D biometric, RGB-D object segmentation and recognition, robot grasping, shark mitigation, and sub-sea ecology related projects based on computer vision

Signal and Image Processing

Group members has proven expertise in audio-visual spatial awareness involving multiple blind source separation of acoustic signals using remote sensors and scene analysis and recognition using audio- visual data. Remote detection, separation, and recognition of multiple acoustic and visual objects is an integral component for the navigation and spatial awareness of any autonomous system. This has been extended to human speech and speaker recognition systems crucial for man-machine authentication and interaction with robotic and automatic systems.

Autonomous Drive Systems

Autonomous drive systems are one aspect for automation in remote areas. This includes people transport (as in autonomous cars) as well as transport of goods, such as in autonomous trains and autonomous LHDs in mining operations. The REV group has significant background in autonomous drive systems and has built two autonomous cars, an autonomously driven BMW X5, as well as autonomously driven electric Formula-SAE car.

Electric Drive Systems

Many types of remote operation equipment rely on electric drive systems, e.g. when it would be difficult to refill the petrol tank. Solar and wind energy is available in most remote operation sites and can be stored in battery banks, in order to drive equipment used in remote operations. The REV group has significant background in electric drive systems. This includes DC, AC, and brushless DC drive systems, including motors, controllers, battery banks, battery management, and charging systems.

Telepresence and Remote Control

UWA helped to create the field of Internet telerobotics using teleoperation and supervisory control methods where local intelligence assists remote operators, also with the help of augmented reality visualisation. In creating remotely operated systems, designers need guidance to help work out the best combination of local workers, remote teleoperators and selective automation by understanding the strengths and limitations of each possibility