The Future Of Wireless Sensor Networks

Frisky070802 writes "In the 12/03 Wired, Intel's Tiny Hope for the Future describes a fundamental transformation as Intel's Research director David Tennenhouse realized the importance of sensor networks. He saw a Berkeley project on 'motes,' little sensors that communicate on ad-hoc wireless networks. 'The company now foresees networks consisting of thousands of motes, located wherever there's a need for data collection, streaming real-time data to one another and to central servers. Intel imagines the day when every assembly line, soybean field, and nursing home on the planet will be peppered with motes, prodding factory foremen to replace faulty machines, farmers to water fields, and nurses to check on something unusual in room E214.' Intel was impressed enough with the technology to fund a whole 'lablet' to develop it. Intel sees a huge potential market in developing both the sensors and the computation to process the huge amounts of sensor information. If this rings any bells, note that the Intel lablets are also behind the Planetlab Internet emulator, previously discussed in Slashdot."

Comments

Potential application

[mattjb0010]

This could be really useful for monitoring kids at the sleep lab where I do some work. It's hard enough just getting a myriad (EEG, EOG, ECG, O2, CO2, etc.) of sensors stuck on a kid, the fact that you then end up with huge mass of wires causes all sorts of problems, making it hard for the kid to get to sleep, plus there's the tendancy to pull on the leads, totally destroying the signal (often several times a night).

Nursing homes being done

[nb caffeine]

The company I interned for this summer had some of this going on: we had tags we would place on residents for certain purposes (ones that would detect urine in an adult diaper, would alert nurse if a resident pissed themself, and wouldn't sit in thier own piss till a nurse came to check, cut down on urinary tract infections, as well as ones that would detect if a person with alzimers wandered too far from their room, that sort of thing). While not in an ad hoc network style, these would at least alert the nearest CNA that something was ary. And if the alert went unchecked, it would go up the chain of command, possibly to the point where the head nurse or director of the home would be notified, and someones ass would be in trouble. Was an interesting application, though i wasnt lucky enough to have worked on them. I got to work on the CRM software. woo and stuff.

Re:Let's not forget the military applications

[Comatose51]

Not true. According to my professor who's working on sensor networks, the military has already tried them out in Iraq. Iraq is well-suited for such things because it is mostly deserts.

Sensor networks would have a lot of difficulties in an urban setting because of buildings. Buildings present challenges to localization of the sensor network. Two sensors can be right next to each other but still can't communicate because a wall is in the way. Furthermore, GPS is hard to receive in an urban setting. Thus, the network must first localize relative to each other then hope that a few of the nodes can recieve GPS to serve as "beacons" to localize the rest of the network.

In other words, sensor networks are more likely to succeed in an non-urban environment first than an urban environment.

Localization is a major problem for these networks because of the lack of processing power and lower transmission radius. However, localization will inevitably involve graph theory and graphs are not the easiest thing to solve. Thus, you can see the trade-offs as the nodes get smaller and smaller and have less resources available.