Using router as a computer

[ggadrian]

Today I assisted to a First Lego League event where the teams (of kinds between 12 and 16 years old) have to present some scientific and make a robot with lego parts.

One of the projects caught my attention:

The kids proposed to install raspberry pi's in the forest forming a grid with an humidity and temperature sensor which will be sending the data to a central server so if a fire starts the firemen can know exactly where it is, making it easier to drop water with a plane in a very short amount of time.

It's actually a very interesting project, because in Catalonia 60% of the land are forest and it's a hot and dry place, so fire is a real problem and the response time when a fire starts is very important.

The problem I see is that they don't really know how to connect the pi's. They said that it would be really cheap (40-50€ per unit) but they where also saying something about sending the data via satellite, which I think would be incredibly expensive.

The units would be like 20 meters apart on the top of the trees, so I thought that maybe the optimal solution would be to make wireless ad-hock connections between the pi's; given that they need very few bandwidth (just sending 2 numbers like every minute) it might work, but the price per unit would be north of 100€ if you have to buy capable wireless adapters and a USB hub to power them.

So I remembered that dd-wrt uses a linux kernel, and the question is: would it be possible to connect the humidity and temperature sensor to a cheap router with dd-wrt and then code a simple program to retrieve the data from the sensors? That would be like 50€ per unit + the power supply (they want to use small solar panels)

Do you think that it's possible? I mean running the program in the router and making a huge wireless network with a lot of dd-wrt routers.

 

[Fardringle]

I'm sure it's possible, but if you don't already have someone with the expertise to program router firmware, then it will probably cost a lot more to hire one to do it than it would to just buy equipment that will work.

 

[VirtualLarry]

What about using those Android-on-a-stick devices? They usually have some form of built-in Wifi support.

 

[ggadrian]

What about using those Android-on-a-stick devices? They usually have some form of built-in Wifi support.

Yes, that's actually a good idea, but there are two points of concern:

-Wireless range (are they going to be able to communicate with another stick 20m apart in a forest? It seems like a difficult matter for a cheap wireless solution)

-Are they capable to create a network?

 

[ViRGE]

-Are they capable to create a network?

It's all a matter of software. They can create a network if someone writes the software to do it.

Ultimately your problem won't be the hardware. You can get simple devices that will do the processing, though you'll probably need an external antenna.

Oh, and I'd suggest seeing if anyone has done something similar. Mesh sensors are not really new (though the application you're proposing may be), so the hardware/communication problem may have already been solved.

http://www.libelium.com/products/meshlium/

http://ieeexplore.ieee.org/xpl/logi...re.ieee.org/xpls/abs_all.jsp?arnumber=1488955

 

[JackMDS]

http://activefiremaps.fs.fed.us/activefiremaps.php

http://www.libelium.com/wireless_sensor_networks_to_detec_forest_fires/

http://www.fire-watch.de/

And many more.

As a Hobbyist there are very inexpensive USB Sensors, if they do not work with a USB port of a Router it is probably Not a big deal to accommodate a 3rd party firmware to do so.

http://www.amazon.com/Powered-Thermometer-Temperature-Sensor-Data/dp/B0054U4YKI

I actually use one connected to my HTPC that can be connected form remote from any Internet connection.

😎



😎

 

[velis]

For the volume (number of units) this needs, it is much more prudent to simply design a custom PCB with a SOC that already has some comms chip built-in and has the sensors on-board. Top-side == solar panel, bottom side == components.
This would end up much cheaper than any of the proposed variants. Building a custom PCB isn't THAT expensive, especially if you have volume.
Initial developement would be done on a Dev board for the said SOC. Low cost + all the necessary tools are already provided by the SOC manufacturer.