WSN Update

Researchers have long been promising that wireless sensor networks will bring a new wave of smart appliances that can communicate with each other, and with central servers when they need attention. The technology for these appliances has existed for some time now. However, the commercial availability of such products has been slow to market for one primary reason - the lack of structured business model.

All that is about to change now thanks to Verizon Wireless's Open Development initiative. The mobile phone network provider has created a new business model for themselves by opening up their network to allow third party devices to be "certified" (they promise quickly and easily, but who knows what that means) for use on the Verizon Wireless Network. With access to the the mobile network, it will now be possible for devices other than your cell phone to communicate. For example, your HVAC can call you when its about to fail or your fridge can call you when you are running low on milk. These examples might not be the most realistic, but the potential for general machine-to-machine (M2M) applications just got a lot more exciting.

Verizon announced certification of the first device under the Open Development initiative this month. Surprisingly, it is a device that has a sensor that dips into a large container (holding the likes food shortening, etc. ) and automatically sends a text message when the contents are low. I guess the examples above were realistic after all.

Beet.tv has a nice interview with Verizon's VP of Open Development, Anthony Lewis.

Labels: low power wireless, m2m, predictive maintenance, verizon wireless

Who said Wireless Sensor Networks had to be small personal area networks? According to Network World, researchers at Penn State University are developing a sensor that would be attached to cars that would allow it to talk to other cars, forming a large area sensor network.

The original concept being developed at Penn State is for the purpose of preventing car theft. It is geared towards parking lots and apartment buildings that might offer the additional anti-theft protection as an added perk. When a user parks their car, the tiny sensor would notify other cars around it that it is there and "joining the network". Periodically, the sensor will broadcast a message to other cars saying "I'm here and I'm safe". When the person unlocks their car to leave, the sensor would notify other cars in the network that it is leaving. If the car were to leave the network without issuing this "goodbye message", in the case of a theft for example, the other cars in the network would notice the car is missing and will notify the network controller to call the police.

The sensors are inexpensive and will eventually be about the size of a grain of rice, making them difficult for a would-be car thief to locate and disable. Even if they were able to tamper with the sensor, the other cars in the network would notice the missing "I'm safe" broadcasts and notify the base station that there is a problem. The wireless network approach will not only make anti-theft protection more affordable, but also more effective.

This concept is a perfect use of Wireless Sensor Technology. Why does it have to be limited to parking lots though? The biggest challenge for wireless sensors is their long term need of power. Cars have plenty of power! Why not put a sensor in every car? Think about what a selling point this could be for Ford or GM? Cars can form a huge wide area network over a large area, routing their information through other cars, until it reaches a controller strategically placed throughout the city. Now it doesn't matter if you car is in a parking lot, parked on the street, or being taken for a joy ride by some punk. A query can be sent through the network to find the exact location of the car, allowing the police to apprehend the thief in minutes.

As with most technological advances, the privacy advocates will have a field day with this one. But I personally believe the benefit would outweigh the potential for concern.

Labels: anti-theft, car alarm, wide area network, wireless sensor

The Baltimore Examiner reported over the weekend that Washington DC is planning to expand its network of wireless gunshot sensors to cover more high crime neighborhoods in the city. By September of this year, police expect 16 of the city's 68 square miles to be covered by the ShotSpotter technology that pinpoints the exact locations of gunshots.

Once the network is deployed, DC will have one of the largest wireless gunshot networks in the country. According to the Washington Post, the technology has guided police to three homicides in Southeast Washington, and in one case officers got there rapidly enough to make an arrest.

Interestingly, the ShotSpotter website states that US Department of Defense has requested that they limit discussion on the how the gunshot detection technology works.

I don't know why they bother as this is no big secret. The basic idea is simple: using the arrival times of the acoustic events at different sensor locations, the shooter position can be accurately calculated using the speed of sound and the location of the sensors. Vanderbilt University was successful in using this technique to pinpoint a gunshot to an area less than 1 meter, which is supposedly more accurate than the ShotSpotter products.

Regardless of it accuracy, the ShotSpotter is an excellent example of a company that is USING the technology to solve a SPECIFIC problem. As we have mentioned here in the past, it will take a lot more APPLICATIONS of the technology before the market really takes of. It is great to see that ball starting to roll.

Labels: gunshot detection, low power wireless, wireless

In my last post, I explained how the lack of adoption of Wireless Sensor Network technology was due to the limited amount of money being invested in actual applications of the technology.

It was refreshing to see this article today from the European Commission Research Information Centre. The EU is investing EUR 2.44 million in the WINSOC project which is developing innovative sensor networks that mimic biological systems. With these new sensors, the project partners hope to detect imminent catastrophes in time for action to be taken.

WINSOC, or Wireless Sensor Networks with Self-Organization Capabilities for Critical and Emergency Applications, is a technical science project that is inspired by living organisms around us. The sensors being developed by the project aim to address three areas in environmental monitoring:

1. detection or prediction of landslides


2. detection of gas leakage to prevent hazard situations or simply avoid unnecessary wastes of energetic resources


3. monitoring of temperature fields as a way to detect fires or, even better, to predict the risk of a potential fire in a given area.

I commend the EU for taking the lead in funding these types of projects. This type of technology can have a huge positive impact on our everyday lives, allowing us to be safer, use less energy, and understand more about our environment.

Full Article...

This blog has been touting the promises of Wireless Sensor Networks since February of 2005. It was predicted back then that WSN would be a disruptive technology that changed the way we viewed the world. Today, market research still predicts that the WSN revolution is "just around the corner". The reality is that the technology is no closer to commercial adoption that it was over three years ago when this blog began. EETimes.com had an interesting article today that sheds some light as to why - just like every other new technology, it needs a "killer app".

The article discusses a UK study that concluded that commercial adoption has been slow. It is the traditional sensing applications that are currently commercially exploiting the advantages of wirelessly networked sensors. There have been very few new uses of the technology. The report goes on to suggest that now that the major system integrators are getting involved, the new expertise in the field will drive further adoption. I am not sure I really agree with that.

Having been in the WSN field for nearly 4 years, I have learned that there is a much bigger problem - MONEY. No one wants to pay for the development of the "Killer App". Integrators will learn the technology, but will only integrate it when they are paid to do so. For the most part they don't develop commercial applications. The traditional sensing markets are very stuck in their ways. They have wired sensors and SCADA systems that work and are reliable. Often these systems monitor critical systems where saftey is a huge concern. So why would they add risk by adopting something new with unproven reliablily?

WSN is a great technology, but that's just it. Its another technology. The technology alone doesn't do anything, it needs to be applied. And for most of the things that it can be applied to, there is often a cheaper, lower-tech, more reliable method.

I still believe the technology has the potential to be disruptive, but its going to take a lot of money and a big leap of faith to develop that Killer App. Until then, everyone is stuck waiting on the sidelines for the game to begin. If Google or Apple are not pushing it, it might be a long time before the technology really takes off.

Embedded.com has a good article today that is Part III in their series of choosing the right low power wireless sensor network. This article shifts away to the general criteria discussion of parts 1 and 2, and dives into the technical details by comparing TI's proprietary SimpliciTI wireless network with Zigbee.

In a nut shell, the article shows how the need for standardized reliability and security implementation as well as an interoperability requirement would make Zigbee the protocol of choice. Relax or remove these requirements and the optimum network would be a general or proprietary 802.15.4 network.

Full Article...

Labels: low power wireless, pan, protocol, wireless, Zigbee

For those who are into the low level electronics of Wireless Sensor Networking, EE Times has a good article today on the importance of ensuring low power in wireless mesh sensor networks.

The article discusses how current consumption is not the only factor to be addressed in order to achieve low power in wireless sensor applications. Also important are managing the peak power conditions, graceful power failures, and acheiving low-power mesh routing.


Full Article...