Two years ago, University of Utah assistant professor of electrical engineering Neal Patwari demonstrated how radio signals could be used to "see" people through solid walls. Now, he is leading a team that is using that same technology to wirelessly monitor peoples' breathing as they sleep. The system could be particularly useful for observing patients who are recovering from surgery, people with sleep apnea, and babies who are at risk of sudden infant death syndrome (SIDS). While respiration-monitoring systems do already exist, Patwari's doesn't require anything to be physically attached to the subject's body, plus he claims that it should be cheaper.

Named "BreathTaking," the system consists of a network of off-the-shelf wireless transceivers (similar to those used in home computer networks), set up along the sides, top and bottom of the subject's bed. Each of the transceivers - or nodes - transmits to and receives signals from all of the others, and they each take turns sending signals. As the subject's chest and abdomen rise and fall with each breath, they impede the wireless signals that are being beamed across the bed. The data from the nodes is processed by a computer algorithm, that detects when too long of a gap has occurred between the signal interruptions. This would indicate that the subject had stopped breathing.

Patwari used 20 transceivers when testing the system, and decided that a minimum of 13 would be required for an insignificant rate of error, while at least 19 would result in no errors at all. The height at which the nodes were placed appeared to have little affect on accuracy.

Serving as his own test subject, Patwari also monitored his respiration using a conventional carbon dioxide monitor, which was connected to his nostrils by tubes. Compared to it, the BreathTaking system was found to be accurate to within 0.4 to 0.2 breaths per minute based on 30 seconds of data. This discrepancy is considered negligible, as most systems round to the nearest breath-per-minute anyway.

The University of Utah team is now looking into using a different or multiple radio frequencies for the system, and the possibility of using it to detect how many different people are breathing in one room - a feature which could prove useful in hostage-taking situations. Patwari has stated that it will probably still be at least five years before it's on the market.

Another experimental non-invasive respiration-monitoring technology known as BabyBeat uses a video camera to detect the subtle changes in skin tone that occur when a baby stops breathing.

View gallery - 2 images