Wouldn’t it be nice if sensor networks using radio frequency (RF) signals for monitoring air quality, the structural integrity of bridges and roads, and traffic or weather conditions—creating the so-called “Internet of things”—could simply draw their power out of thin air?
Now, using a technique called “ambient backscatter,” University of Washington researchers are doing just that: They have transformed existing wireless signals into both a power source and a communication medium for a sensor network. Each device within the researchers’ network is a very basic sensor made up of a credit-card-size circuit board with a transmitter, a receiver and a component for harvesting energy from TV signals, all connected to the same antenna. Embedded in roadways or bridges, such networks could monitor structural integrity and signal authorities about faults via e-mails or text messages. Someday they might even provide back up messaging capability if your phone’s battery is dead.
Earlier efforts had leveraged RF signals specifically for sensor power. This is done by capturing energy from TV or cellular signals and converting that energy into electricity to power a circuit. “What’s new here—we’re doing that [harvesting signals for power], but we’re also communicating by reflecting those TV signals,” says Joshua Smith, a computer scientist and electrical engineer of the University of Washington. The researchers presented their work (pdf) last week at the Association for Computing Machinery’s Special Interest Group on Data Communication 2013 conference in Hong Kong.
Devices in the researchers’ sensor network send and receive messages to one another by either absorbing or reflecting ultra high radio frequency TV signals produced by a nearby tower. Each device also has an LED light that flashes to indicate to observers when it has received a signal from another sensor in the network.
Repurposing existing RF signals does not interfere with their intended purpose, whether it is delivering TV programs or mobile communications. “All we’re doing here is creating additional reflections of existing wireless signals around us, whether they are TV, wi-fi or cellular,” says Shyam Gollakota, a University of Washington computer scientist who worked with Smith on the project. The sensors in the network require little power because they are not creating any radio signals of their own.
The researchers tested their proof-of-concept sensor network in several locations—inside an apartment building, on a street corner and on the top level of a parking garage—all within about 10 kilometers of a TV tower in the Seattle area. Outdoors, the sensors communicated up to 0.75 meters from one another at a rate of 1 kilobit per second. That’s enough throughput to transmit a small amount of data—the basic info that might be contained in a sensor reading or a text message. Indoors, the signals are weakened a bit, meaning the devices could exchange information at this rate within a slightly shorter range of half a meter. The researchers might increase transmission range in subsequent experiments by developing a more efficient coding scheme or by having the sensors send redundant signals that improve the chances of a message getting through.
The researchers expect they will eventually be able to increase data transfer rates to 600 kilobits per second or even 1 megabit per second. One way of achieving this could be leveraging the structure in the TV signals instead of treating them as completely random and uncontrollable.
The primary restriction of ambient radio signal harvesting is that devices can draw limited amounts of power—measured in microwatts—even from large numbers of signals. Although the technique is feasible for powering small sensors, it is impractical for, say, recharging mobile phones or tablets. Nevertheless Smith imagines a scenario in which a mobile phone could use ambient backscattering to send and receive e-mails or text messages even if its battery is dead. “Or,” he adds, “you could build this type of capability into a smartphone just to reduce overall power consumption and improve battery life.”
