Battery-free Wireless Communications – How It Works
August 22, 2013
The University of Washington recently announced a new concept that allows Wireless devices to go battery-free with new communication technique. The devices use a technique the UW researchers call ambient backscatter.
The concept is simple – you have a device with an antenna that is either terminated so it absorbs RF or unterminated (or shorted) using a simple switch. For this experiment, the researchers used existing over-the-air DTV signals as the RF source. The receiver has an antenna the same size as the transmitter (both tuned to UHF TV channels for the testing) and is able to detect the variation in the RF field caused by the transmitter switching its antenna from a reflective to absorptive mode. The devices operate at low data rates, from 100 bps to 10 kbps, and do not require any active RF signal processing or DSP circuits. This allows them to use the ambient RF not only as a way to communicate (similar to a lost hiker using the sun and a mirror to signal a helicopter) but to rectify the RF and store it to power the devices. The receiver is designed so that it turns on only when data from the transmitting device is detected.
Joshua Smith, a UW associate professor of computer science and engineering and electrical engineering and co-author of the paper Ambient Backscatter: Wireless Communication Out of Thin Air explained, “Our devices form a network out of thin air. You can reflect these signals slightly to create a Morse code of communication between battery-free devices.” The paper has more details on the device and an experimenter should be able to duplicate at least a crude form of this technology with the information provided.
The transmit side of the device is easy to understand, but I was wondering how a simple “zero power” receiver could detect the changes. It turns out it isn't that difficult! According to the paper the median power ratio is about 1.4, similar to the levels targeted by traditional backscatter communication in RFID devices. Due to multipath (you've seen the difference in signal levels moving around an indoor TV antenna) the power ratio can be as high as 4.3. Figure 6 in the paper shows a circuit diagram for the demodulator. It consists of a diode, some RC filtering and a TS881 ultra-low power comparator. The filter component values are listed for 1 kbps and 10 kpbs bit rates.
By now you've probably figured out how the receiver works. The incoming signal is detected in the diode, just like in a crystal radio AM receiver! The filter removes the 8VSB modulation leaving only lower frequency components. The RC network on the input of the comparator allows it to adjust to slow changes in the RF level (people moving around a room) while decoding the simple on/off data stream. The physical layer uses FM0 encoding to put data onto a stream of pulses by switching between reflective and non-reflective states. TI has a good explanation of FM0 encoding in its UHF Gen 2 System Overview (Slide 20). The UW team created a packet format for the device with room for a preamble, header, data and CRC. An alternating sequence of 1's and 0's at the start of each packet allow for carrier sense. The prototype device was designed so it will only turn on when the rectified RF used for power charges the device's storage capacitor to greater than 1.8 V. The paper provides detail on power demand by component. The “RF” side of the device is responsible for around 1% of the total power consumption!
Since these devices “modulate” the RF field from a TV transmitter, can they cause interference? They aren't generating any RF on their own, so FCC rules don't apply. The UW researchers actually tested device impact on TV receivers. At 100 bps no noticeable glitches were created unless the device was less than 2.3 inches from the TV antenna. At higher data rates of 1 kbps and 10 kbps, the median distance for interference was 4.1 inches and 3.7, respectively. At distances greater than 7.2 inches no glitches were seen at any data rate, so TV interference shouldn't be a problem in most cases. If the device was included with a smartphone or tablet with a TV tuner it may cause some problems, not only to the TV reception but perhaps to other radios in the device.
As expected, performance declines as the TV signals' field strength declines. The researchers tested the system up to 6.5 miles from the TV tower. The prototype devices could be separated by as much as 2.5 feet outdoors and 1.5 feet indoors.
Let me know if you try building these devices. A simple proof of concept would be to test it using just a square wave into a low power RF switch, perhaps just a 555 timer and some FETs, and a demodulator similar to the one shown in the paper. An interesting experiment would be to see if a higher-order active filter on the comparator/demodulator centered on the data frequency allowed reception over longer distances.
I doubt these battery-free devices will ever have the range and data capacity to be used for human-to-human wireless communication, but they are perfect for close-range remote sensing where providing power isn't practical or as a low cost substitute for RFID (the researchers give one example of using the device to check store shelves). They may also be useful for short range remote control, assuming enough TV stations are on the air to provide the RF and the power!