Sensing robots

If you’re visiting Dale Carnegie and his Mechatronics team at Victoria University, don’t be scared when you are approached by MARVIN (short for Mobile Autonomous Robotic Vehicle for Indoor Navigation), a very sophisticated security guard.

MARVIN is designed to wander the university corridors and detect intruders, but to do this, he must have a wide range of sensors and a means to interpret the information.

MARVIN is equipped with a number of different sensors that he can use to detect changes in his environment. He ‘sees’ using infrared sensors, he ‘hears’ through electric sensors sensing sound and ‘feels’ through touch sensors. These sensors send electrical signals to other components that may trigger a response, for example, MARVIN can change his body posture to assume a more aggressive stance if provoked.

MARVIN is also a research tool that Dale Carnegie and his team use to develop new ideas, for example, a new sensor is being developed that is similar to a human eye that will allow MARVIN to find the range of every item in front of him.

Search and rescue robots

When the World Trade Centre in New York collapsed in 2001, a robotics team from Florida was sent in to try and find survivors. Unfortunately, they were unable to find a single living person. Part of the problem was that these robots were very large and couldn’t penetrate much of the rubble. Their large size was due to big batteries needed to run a lot of electrical circuits and components used for sensing and responding mechanically.

Dale Carnegie’s team have developed a different way of designing rescue robots. Instead of one large robot, they are constructing a system of robots that communicate with one another electronically. They have developed a 3-tiered system consisting of a ‘grandmother’ robot, several ‘mother’ robots and up to 200 ‘daughter’ robots.

The grandmother is a large, complex robot designed to act as a launch platform for the mothers and as a base station where it processes and controls the information sent to and from the other robots. The grandmother does not enter the disaster zone, so her size is not a problem. The smaller mother robots are able to travel over rough ground and contain many sensors for mapping and navigation. When they sense a gap in the rubble, they dispatch a daughter robot into the gap.

The daughters are launched from the mother, with a range of only 100 metres. They are small enough to crawl into small gaps in the rubble and are equipped with electronic sensors to detect a trapped human. These sensors may detect movement, heat, sound and even gases such as carbon dioxide. The daughter robots are very small, about the size of a credit card, and are designed to be completely disposable. It is not expected that any of them will be recovered once launched from the mother.

The information from the daughters’ sensors is sent back to the mother robot, and she relays the information from the daughters to the grandmother for processing.

So instead of putting all the information gathering and analysis of data into one robot, the team at Victoria have separated the tasks and designed robots of different sizes to carry out different tasks.

An analogy of how the system functions is that the grandmother acts as the brain, the mothers act as the arms or legs, and the daughters act as the skin (detecting heat), eyes and ears. Specialised electrical circuits form the basis for these sensing activities.

The robots that were sent into the World Trade Centre disaster were expensive and required skilled operators, but Dale’s grandmother robot is the skilled operator, controlling up to 200 daughters – this means a disaster area can be covered with many more sensors than the ‘do everything’ robots used in previous rescue situations. The daughter robots aren’t recovered from the disaster scene, but because they are disposable, their design can be kept much simpler and cheaper.