Groups of cooperating robotics working in hazardous environments will require a networked approach to systems design. One Australian research laboratory is leading the way.
Any human mission to Mars will require a high degree of robotic support to accomplish a wide variety and difficult tasks prior to human arrival. Tasks may require a large number of robots, cooperating with each other to handle multiple scientific and engineering objectives.
This becomes more complicated by the fact that the missions can change once humans arrive, requiring the “system of systems” of robots to now be redesigned for a completely new scenario.
What system can solve such a complicated problem? The system must understand itself in a highly intelligent way, by self definition based on the group’s capable behaviour. With this knowledge it can reconfigure itself by adding or removing individual robots, or by changing their organization to fit mission requirements.
Researchers at the Australian Centre for Field Robotics, a part of the Centre of Excellence for Autonomous Systems (CAS), are tackling the problem from a network centric perspective. Their approach¿network all of the people and platforms in the system.
If the platforms in the system are all networked they can communicate their performance metrics to everyone in the group. This information is then used to quantify a systems level understanding of the group’s behaviour, called a system map.
System maps are created for each combination of platforms and organization. The best system map is then the system with the best behaviours to complete the mission. And what if the mission changes? The solution may be as simple as choosing a new system map.
