Applied Artificial Intelligence - Robotics

Robots are a practical application of artificial intelligence of rapidly growing importance for industrial, domestic, entertainment and military tasks. The capabilities of today’s robots go far beyond the factory robots of the last century, whose operations depended on the strictly controlled conditions of an assembly line. Today’s field and service robots must be able operate in the unstructured world of a shop floor, a home, a school or a battlefield. Furthermore, the software controlling such machines must be far more flexible, robust and autonomous than the robot programming languages of the past. They must be able to be commanded to go about their tasks far more readily than is today possible. In order to solve such demanding problems, it is not only necessary to build and test new kinds machines, but also to create new and powerful classes of algorithms, the general applicability of which can be demonstrated by the successful operation of those machines on challenging problems. We invite research students to join us in this important work; skills in electronics, mechanics or software development are welcome, but an interest in robots is essential!

Current Projects

Our existing projects are carried out in the School’s Robotics Laboratory, which is now undergoing expansion. This laboratory has accumulated a range of tools and electronic instruments, battery chargers, wheeled worktables etc. , for construction and maintenance of our machines. There are also a number of existing robots from older projects (in various conditions!) which might be modified, rebuilt or cannibalised for parts.

Mascot Field Robot

This field robot is designed to demonstrate that six-legged locomotion can provide good speed, traction and stability in uneven or broken terrain that cannot be matched by wheeled machines. Initially, it will be a remotely teleoperated, fast, stable, mobile platform which can carry a pair of cameras anywhere outdoors, including off-roadways, up stairs, through the bush and in the rough, and send useful imagery back to a video recording station. Such a machine could be used for maintenance and security photography in remote locations, capturing exciting but difficult sports or live animal motion shots and for entertainment in the form of embodied video games, to name only a few applications.

To avoid the well-known problem of unreliability in complex, jointed leg systems, the mechanism has been reduced to its simplest form. Six simple passive spring legs, each mounted on an independent revolving axis (6 DoF in total) are driven by 18V 100W DC motors fitted with 150:1 planetary gearboxs. The motors are driven by six servo control boards connected to a 32-bit RS485 multidrop network controlled by an onboard subcompact computer running Windows XP.

The machine is 590mm long, 570mm wide at the middle legs and 700mm from the ground to the top of the current camera mast. It weighs approximately 14kg. Imitating the gait of an insect, the machine moves by “tripod walking”: at any instant three legs are on the ground, and these alternate between the sides of the robot. The machine is steered by altering the phase relationship between the tripods on either side. The main power supply for the motors consists of two 18 volt, 13Ah Lithium-Ion battery packs with built in voltage regulators and thermal shutdown circuitry. A 12v 2Ah Lithium Ion battery supplies logic power.

Walking test video: http://www.youtube.com/watch?v=OskuPsk9DRI

The camera platform is designed around a pair of EO5-380 CCD cameras mounted on a tilt-pan head. Each camera is capable of transmitting 380-line PAL colour video over at 2.4GHz wireless link. At this stage the cameras are not used by the robot as a vision system, but only as part of a low-cost teleoperation control system. This also depends on a commercial 6-channel 36MHz FM wireless remote control system, designed for model aircraft. Two channels of this control the effective left and right steering, and two channels control the tilt and pan motors of the camera head.

When completed, this project will allow a remote operator to control high-speed locomotion of the robot while viewing real-time video from the cameras on a small LCD monitor. Still and video cameras will also be remotely controlled. Depending on the performance of the machine, the project can progress to a high-level commanding mode or to full automation.

Principle Investigator – Dr. G. Mann

Behaviour-Based Visual Servoing with the SmartArm

These experiments are being conducted with Prof. Arcady Dyskin of UWA’s School of Civil and Environmental Engineering. Prof. Dyskin has been experimenting with new kinds of bricks, which tessellate and interlock in three dimensions, the better to stabilise structures built with them. Compared to rectangular prism stone blocks or ordinary house bricks, structures comprised of his osteomorphic (“bone-shaped”) bricks would be far more resistant to collapse in the event of collision by a vehicle, or in the event of earthquakes. Prof. Dyskin believes that if houses and public buildings in earthquake-prone areas of the world, such as mainland China, Japan and northern Pakistan could be constructed of such bricks, thousands of lives could be saved every year.

It is important that these new bricks can be produced very cheaply, for only then would they be economically competitive with respect to existing bricks. To convince brick-making companies such as Boral Australia that such bricks could easily be formed, handled and produced in large quantity, it is necessary to demonstrate that they can be deftly and easily manipulated by robot arms. (Boral uses robots to produce almost all its bricks now.)

We also want to explore the possibility of automating construction using the bricks. This is because deaths in earthquakes come not only from people being crushed by the collapse of buildings, but from being deprived of the shelter that those buildings afford in extreme weather conditions. The need for many emergency shelters to be quickly assembled or constructed can be extreme in disaster scenarios, and there is always a shortage of time and labour. Conditions can be very harsh. The idea, then, is to provide an option for construction of emergency shelters or reconstruction of crucial building infrastructure, even when almost no human labour is available. Therefore, a range of site-levelling, block placement and cable-threading task requirements is also desirable for a construction robot, and the feasibility of this must be demonstrated experimentally.

The SmartArm is a small-scale model of a robot capable of accomplishing these tasks. Although superficially it resembles a conventional factory robot, it is in fact a service robot which operates on a very different principle. Instead of being programmed with long and rigid sequences of steps which will only operate under ideal condition, the SmartArm depends on a very flexible, behaviour-based visual servoing program which allows organisation of basic behaviours for the robot – finding, grasping, moving and stacking blocks – to be assembled hierarchically into more complex behaviour in a very robust way. A short movie of the visual tracking behaviour can be seen on YouTube at http://www.youtube.com/watch?v=b1U-XzFV7rE

This draws on the work of Swedish researcher Zbigniew Wasik, who succeeded in reducing the problem of composing simple behaviour-based units into complex behaviours of this kind – that is, his techniques can both arbitrate between multiple possible behaviours and fuse their parameters into optimal behaviours at the same time. Using this method, the fuzzy rules controlling these behaviours can be made simple, robust and easy to expand.

Principle Investigators – Prof. Arcady Dyskin, Dr. G. Mann.

Possible New Projects

Research students are encouraged to develop ideas of their own in consultation with academic supervisors within the Applied Artificial Intelligence group. Possible new projects include:

  1. Testing LCIRCLE family of algorithms for obstacle avoidance
  2. Development of a low-cost teleoperation system for mobile robots
  3. Upgrade of TarBaby biped to a self-contained humanoid
  4. Development of enabling robot technologies for human exploration of Mars