Robots and beyond

"I notice that all these articles are in the same format – problem/solution," comments Professor Alcherio Martinoli as we sit down to talk about his work. "I have to warn you, we don’t work like that. We work on enabling technology. We can’t afford to build ad-hoc solutions for a single application. We develop methodologies and recycle their resulting "solutions" as much as we can, applying the basic technology we’ve developed to a variety of problems."

Martinoli works with multi-robot systems and sensor/actuator networks– "mechatronic substrates", in his terminology. That means devices that involve mechanics, electronics and software. He is mainly interested in systems that are distributed in space, and that operate quickly, simultaneously, and flexibly. One of his projects involves fleets of 12cm robots equipped with odor detectors that wheel around sniffing out potential gas leaks or noxious chemicals. Once a robot gets a whiff of the desired odor, it quickly communicates its find to the rest of the robots, who then reorient themselves and refine their own searches accordingly.
In another project he contributes to the development of sensor networks that can record a range of environmental variables, in addition to adapting to the field under observation. "In the swiss experiment," he explains, "we essentially have a system of robots without wheels." These are the autonomous, networked meteorological stations developed in collaboration with enAC’s environmental Fluid Mechanics and hydrology laboratory and two laboratories in the iC school (Profs. Vetterli and Aberer). Deployed en masse in mountain valleys, around campus, within cities – anywhere except under water, in fact – these tripod-based stations collect data on environmental variables, and then communicate between each other, bundling and "hopping" their data from one station to the next, down the chain to a "mother" station which assembles the lot and makes it instantaneously available over the internet. This is a far cry from typical atmospheric data collection and a real leap forward for environmental engineers. But Martinoli is already thinking about pushing the technology to the next level. "Now we could add mobile robotics technology to the picture. It could bring a lot of good information, but at a cost. You have to provide additional energy for self-locomotion. Could they fly? hop? Float in the air?"

At first glance, Martinoli, who has been trained as electrical engineer and computer scientist, might seem out of his element in enAC, working on complicated bundles of software, electronics and mechanical parts while surrounded by engineers and architects more obsessed with issues like construction projects, global warming and toxic waste removal. But the kinds of things he and his team can contribute at the application level are more than just interesting – they may open up possibilities that could revolutionize the way that civil and environmental engineers tackle their problems. And for his part, Martinoli enjoys the challenge of working on systems that operate in the very challenging conditions of the real world, instead of the typical laboratory settings electrical engineers and computer scientists are used to. Martinoli is already thinking about pushing the technology to the next level.

"Civil or environmental engineers typically use off-the-shelf modules and patch together equipment," Martinoli explains. "We go one design step below that, and buy off-the-shelf components, wire them together using printed circuit boards, and write our own firmware and software." It’s not always easy to convince environmental and civil engineers that their way of doing things might be improved upon. But the possibilities are intriguing. Imagine a distributed fleet of floating robots equipped with sensors and actuators that, in the presence of a trigger – a pollutant spill, say – could join together in a chain, encircle the spill, and release remedial bacteria, for example. Or imagine a bridge that could adapt its shape to the force of the wind via a distributed network of sensor-equipped movable elements. These mechatronic challenges have to work in the real world and be cost-effective. And they also challenge ingrained ideas of what is possible and safe in engineering. Nonetheless, Martinoli firmly believes that interdisciplinary work like this can lead to breakthrough technologies in all the disciplines involved.