Itchy and Scratchy
Humans, when faced with the task of lifting something heavy or bulky, tend not to go out and find the biggest, strongest person they can to accomplish the task. Rather they will find other people so that the task is distributed or shared. Traditional trends in robotics have gone the opposite way, bigger and more powerful robots were successively being constructed to deal with tasks such as cutting down and stripping forest trees. However, these robots are expensive to manufacture, difficult to transport and in the case of the forestry industry, they actually damage the ground due to soil compaction.
More recently, cooperative robotic interaction has been explored. There are numerous examples of applications for such collaborative robotics, not the least being that NASA realise the only way a space colony will be built is by robots as humans would be unable to carry all the necessary life support equipment required for the duration of such a project.
To explore this form of behaviour, two identical robots affectionately called “Itchy” and “Scratchy” were constructed. These robots are based on a tricycle arrangement with the steering wheel located at the back (in a similar manner to a forklift) in order to improve manoeuvrability.
Itchy & Scratchy Video.wmv (6.4MB)
Current projects:
Praneel Chand - PhD Student I am developing a cooperative control architecture for Itchy and Scratchy. The architecture will be decentralised and distributed to enable the robots to have equal control and make their own decisions. At the individual robot level, the architecture will incorporate real-time control for motion control and collision avoidance as well as non-real-time control for planning tasks and coordinating the robots. The architecture will be kept as generic as possible to allow the robots to enable a range of tasks. Given the current capabilities of the robots, a cooperative map-building and exploration task will be demonstrated. To perform this task the robots will require a map-building algorithm as well as an exploration strategy. Also, since there are multiple robots, issues such as the sharing of resources (e.g. shared environment), task allocation, map synchronisation and robustness to robot failure also need to be considered.
EResearch