Innovation

SHWETAK PATEL likes to put his ear to the wall. In September this unusual penchant earned him a $500,000 "genius prize" from the MacArthur Foundation, a not-for-profit outfit that spontaneously rewards individual intellectual exploration. Dr Patel, an assistant professor at the University of Washington, uses plug-in sensors to listen out for minute fluctuations in the flow of power and water in buildings. The noise can be used to infer when, say, a television set is turned on, whether a toilet is leaking, or which part of the house has people milling about.

The idea of relaying information from the home to make it function more efficiently is not new. But many efforts have foundered, falling victim to the vagaries of the utilities market. Google.org, the internet-search giant's philanthropic arm, retired its PowerMeter project in September. A similar project by Microsoft, called Hohm, is scheduled to wind down in 2012. Both required smart appliances and specially equipped meters to collect and analyse data.

Dr Patel's sensors, by contrast, dispense with all that. They paint an accurate picture of household resource consumption, but leave it up to the denizens whether they act on that information. Corporate and other partners were quick to spy tremendous commercial potential in this and other work emerging from his lab. The grant, paid out quarterly over five years, will therefore add to an already ample kitty. Last year Dr Patel sold the company he co-founded in 2008 while in graduate school to Belkin, a maker of peripheral devices like routers, and which is banking on consumers' interest in green technology.

The 29-year-old boffin has received other accolades, as well as government grants and corporate research funds. However, the MacArthur prize is different. It brings not just dollops of money, but also a responsibility, he says. Flush with cash, Dr Patel now has to figure out the best way to spend it. He is thinking of setting up a not-for-profit outfit which would offer his technology to low-income households.

The foundation's largesse frees some fellows from the drudgery of supporting themselves, letting them focus on blazing new trails in their fields. As in previous years, the 2011 crop includes a gamut of talented individuals. There is a condensed-matter physicist, a producer of a public radio show, a cellist—even a silversmith. Many fellows would previously eke out a living as poets or other skint professionals. Computer scientists and engineers are rare among them; they can, after all, always tap other sources of wherewithal.

The grant is meant to be awarded at a key point in a creative career, around the age of 40. That is often when an individual has achieved, or is on the verge of achieving great things, but has yet to receive broader recognition. Dr Patel is one of the younger recipients. Last year Babbage spoke with Matthew Carter, a master type designer, who won his award at 73. Both Dr Patel and Mr Carter view the grant as encouragement to pursue novel ideas.

Dr Patel has no shortage of those. When your correspondent paid him a visit at the computer science and engineering department, he sketched out several of them. Each converts epiphenomena of existing technologies, currently treated as noise or waste, into useful information. All could one day become billion-dollar businesses.

Another common feature is a reliance on machine learning, where users' feedback is used to hone the automated system's analytical skills. For instance, a water sensor developed by the firm Dr Patel sold to Belkin can tell not just that a toilet was flushed, but which one was. He says toilets, faucets and other water flows behave in unique ways, and the further they are from the sensor, the lower the amplitude of the recorded sound.

Dr Patel shows off one tiny circuit board with a long antenna wrapped around it several times. The device could, for instance, come equipped with a sensor to measure the number of mould particles in the air, moisture content or other parameters. Such devices exist today, of course. But Dr Patel's big idea is to dispense with complicated wireless routers typically needed to transmit the data from the sensor to the control console. Instead, the widget's antenna produces a signal that is picked up by the wiring in the walls of a house, slightly altering the electricity flow through it. An adapter plugged into the wiring anywhere in the home can then interpret that signal.

The sensor board uses almost no battery current. It could run for 50 years on a single coin-sized battery. It will corrode long before it runs out of juice, jokes Dr Patel. Low cost and long battery life means that walls could be riddled with such devices. He suggests some could be stuck through holes in a wall and left to produce signals until they expire.

Dr Patel has also looked into using a home's wiring to detect occupants' location and movements. The idea, called the Humantenna, is supported in part by Microsoft Research. Microsoft's interest might be explained by the technology's potential for consumer electronics, such as games consoles. Unlike its popular Kinect optical motion detector, the Humantenna would not need any additional sensor. Instead, it would calculate movement based on fluctuations in electromagnetic noise. (The researcher has so far shunned government and defence contractors, despite the obvious police and military potential.)

The team Dr Patel leads at the University of Washington has also looked beyond a home's innards to those of human beings. In one project, a smartphone app uses the built-in microphones to listen to a person's breathing, and determine whether an asthma attack is imminent. Usually, a test must be administered with specialised equipment in a doctor's office. With the app, a person could screen himself and administer prescribed medicine before asthma strikes. The group will soon test the device with the nearby university medical centre.

A relative stripling, Dr Patel is set for a rip-roaring career extracting meaning—and dollars—from noise. Listen carefully, and you can expect to hear more from him.