Sensor networks and such

[Jamie Lokier]

Ben Ward wrote:
> Can anyone help me find a suitable way of handling and
> addressing/organising a variable number of Internet of Things sensors
> where the sensors won't have a 1:1 relation to IP addresses? SNMP looks
> right, but is overkill and unpopular, and doesn't self-describe. XML
> is just a syntax without a structure. OGC sensor standards perhaps?

Hi Ben,

Are you looking for an existing framework, or in approaches to that
problem?

I can't point you to an existing framework, not least because I've
been having too much fun exploring my own work in the area of similar
protocols :-) (which is regrettably unpublished at this point).

On the other hand, if you are interested, I've been addressing a few
thousand devices spread across the country via custom, bi-directional
pub-sub networks (over flaky connections!) for a few years now, and
studying the issues of interaction with various IP configurations,
self-organising connections, efficient scaling and synchronising, etc.

(I could talk about where I'm taking that next, but I'd rather wait
until I have something to show rather than talk abstractly).

How many Things do you have in mind? A few? Thousand? Million+? ;)

Best,
-- Jamie

[Jamie Lokier]

Jamie Lokier wrote:
> On the other hand, if you are interested, I've been addressing a few
> thousand devices spread across the country via custom, bi-directional
> pub-sub networks (over flaky connections!) for a few years now, and
> studying the issues of interaction with various IP configurations,
> self-organising connections, efficient scaling and synchronising, etc.
>

> How many Things do you have in mind? A few? Thousand? Million+? ;)

By the way, in the event that anyone stumbles over my name on the ietf
hybi list (i.e. WebSocket protocol design), my contributions in
respect of how to do multiplexing protocols that work, are derived
from my experience with the above network, where multiplexing is useful.

It reduces traffic (and connections) when you have large numbers of
devices in clusters sharing internet connections. This is especially
true of pub-sub patterns and can be quite an extreme saving with large
numbers of devices per cluster.

-- Jamie

[Ben Ward]

Hi Jamie, thanks for turning this into a thread, 

On 26 July 2012 22:46, Jamie Lokier <ja...@shareable.org> wrote:

Ben Ward wrote:
>    Can anyone help me find a suitable way of handling and
>    addressing/organising a variable number of Internet of Things sensors
>    where the sensors won't have a 1:1 relation to IP addresses? SNMP looks
>    right, but is overkill and unpopular, and doesn't self-describe.  XML
>    is just a syntax without a structure.  OGC sensor standards perhaps?

Hi Ben,

Are you looking for an existing framework, or in approaches to that
problem?

Well, not being a fan of reinventing the wheel, nor a fan of searching through standards documents I'm up for either.

 

I can't point you to an existing framework, not least because I've
been having too much fun exploring my own work in the area of similar
protocols :-)  (which is regrettably unpublished at this point).

Ah, then I'm in the right place! The closest I've found so far is OGC Sensor Web Enablement. http://www.opengeospatial.org/projects/groups/sensorwebdwg

I don't yet know enough about it to know if it's applicable, but is the only answer I've had that takes things beyond either proprietary M2M style databases or transport protocols.

Seems to me that huge numbers of sensors is a scalability/manageability problem, and not one that's currently a problem for many people, otherwise they'd all be clamouring for it, but after doing lots of work with SNMP I realise that managing millions of sensors scalably is Hard Work.

 

On the other hand, if you are interested, I've been addressing a few
thousand devices spread across the country via custom, bi-directional
pub-sub networks (over flaky connections!) for a few years now, and
studying the issues of interaction with various IP configurations,
self-organising connections, efficient scaling and synchronising, etc.

Yes, that does interest me :-)  I think it betrays the IoT movement's ability to talk about billions of connected devices, whilst having no real solutions how to get beyond farming a few thousand. 

Would your things be GPRS/UMTS?  How do you feel about the promise of Whitespace?

 

(I could talk about where I'm taking that next, but I'd rather wait
until I have something to show rather than talk abstractly).

How many Things do you have in mind?  A few?  Thousand?  Million+? ;)

all of them!

Heh, hundreds for now. Ultimately, though, we'll need a globally addressable scheme. Like a URI. Yeah, a URI :-)

Ben

--
b...@crouchingbadger.com | http://www.crouchingbadger.com

[Nick Ryder]

I've been thinking that setting up some simple, hopefully cheap, sensors around Oxford would be interesting. They could log the temperature, noise and light levels fairly easily and cheaply. They could be a fairly small battery powered package that we could put around town. I just can't think of a cheap way to have them communicate back to a central server so we could have some decent web based ~real time display. I have a 3G dongle that a microcontroller can use to get online but they require quite a bit of power (whilst connecting/in use) and aren't very cheap (~£25). I think if we could design something for <=£20/piece I'd be happy to buy 5-10 to put around town. Looking at some sensors on RS then a temp + colour light + sound sensor package could be <£5, another £5 for a uC and cheap chinese PCBs for £1 could form the base of some sensors if someone could help me find a good communication method. Worst case scenario it might still be interesting to just have them dump data to an SD card and manually collect it periodically (presumably when batteries need changing too), but having some real time communication would be preferable.

Let me know if you have any ideas,

Nick

[Eileen]

Can’t comment on the techie side, but presumably if you plan to distribute these around Oxford you’ll need permission to attach them (to something) from the Council. They may be interested in your plan anyway, if presented the right way. I used to work very near Carfax and some years back (maybe 12-15) there was a survey of air quality – the city centre came up with very poor results. If you’re looking for a dogsbody to approach the Council, I’d be happy to have a go at the relevant stage.

Cheers

Eileen

[Lauren Hutchinson]

Agreed, I was just thinking in response to this that it's an interesting project in itself, but even more interesting if it somehow produces air pollution ratings and we make a colour-gradient Oxford area pollution map.  I'll have one of the sensors in my back garden, and maybe others would too, for a start.  Thus it would use my wifi and I can switch out batteries (within reason)--bonus?  Perhaps some of the museums and colleges would have them on their roofs too.  Wouldn't it be cool to train up the gardeners of each of the colleges to switch out the batteries and clean the solar panel once a season?  ;)  We'd have pollution and/or weather ratings for each of the 45-odd colleges!

...maybe the engineering dept (with its link to the old power station management) would even become aware of our good tech works through this mechanism?

L

--
You received this message because you are subscribed to the Google Groups "Oxford Hackspace" group.
To post to this group, send an email to oxford-h...@googlegroups.com.
To unsubscribe from this group, send email to oxford-hackspa...@googlegroups.com.
Visit this group at http://groups.google.com/group/oxford-hackspace?hl=en-GB.
 
 

[Jamie Lokier]

Lauren Hutchinson wrote:
> Agreed, I was just thinking in response to this that it's an
> interesting project in itself, but even more interesting if it somehow
> produces air pollution ratings and we make a colour-gradient Oxford
> area pollution map. I'll have one of the sensors in my back garden,
> and maybe others would too, for a start. Thus it would use my wifi and
> I can switch out batteries (within reason)--bonus? Perhaps some of the
> museums and colleges would have them on their roofs too. Wouldn't it
> be cool to train up the gardeners of each of the colleges to switch out
> the batteries and clean the solar panel once a season? ;) We'd have
> pollution and/or weather ratings for each of the 45-odd colleges!
> ...maybe the engineering dept (with its link to the old power station
> management) would even become aware of our good tech works through this
> mechanism?

When you put it that way, it is sounding more and more worth doing as
an official oxford hackspace event.

Using the quadrotor to fly the sensors and land them in difficult to
reach spots :)

-- Jamie

[Lauren Hutchinson]

OH, THIS!

Mind you, dropping the sensors a few dozen times could get expensive, heh.

;)

How awesome that would be.  Maybe if we're feeling kind of sluggish and unfocussed about formal community outreach stuff (and we seem to be at the moment), we should relax for the moment on that front and do a "Cool PrOXjects" initiative instead.  That is, we choose two or three big projects that are "cool projects for the good of the oxford community" (=Cool, OX) and that will allow us to hack stuff.  Then we roll it out in the community, and the rollouts slowly start osmosising our good name out to council, like-minded groups and companies, and so on.  We get our profile raised through hacking stuff for the community.  Works, no?

We'd need proposals for what the projects should be, and we have the luxury of time so we don't need timelines and can do rollouts when the project is organically ready.  Can't lose.

Ie:   (ok just riffing now, join in!)

Cool PrOXjects 2012
-----------------------

1--Oxford Pollution Skymap

An interactive, digital, colour-gradient pollution map reading daily and nightly from each of the 45 Oxford colleges, private homes and more, the Oxford Pollution Skymap is regularly updated.  Scroll back through the seasons to see how pollution data fluctuates in YOUR area.  Use it to raise awareness of air quality issues in your neighbourhood, to choose the most healthy routes if you commute by bicycle or run regularly for exercise, to help you track personal asthma and allergy severity, or even to help you choose your next home or school for your child.  Here's to your health!  Brought to you by your friendly neighbourhood Oxford Hackspace.

2--Oxford Kiss Map

As part of our ongoing quest to bring the romance and mystique back to the thousand-year-old twilit Oxford streets, Oxford Hackspace presents the Oxford Kiss Map.  Modelled on the Toronto Kiss Map of Toronto, Canada, (http://torontokissmap.com/), the Kiss Map records the stories and locations of the first and special kisses that happened in these streets, both modern and pulled from literature, and encourages you to add your own special story.  Select stories may be projected onto walls or, using motion sensors, narrated by British actors at the outdoor spaces they occurred at, during summer evenings, to remind passersby of all of the joyous forgotten stories these ancient streets could tell!

3--Next Gengineering Day 2012

On October 7, 2012, join Oxford Hackspace and 100 other Oxford schoolchildren for an afternoon and evening of code breaking, programming, electronics training, hobby rocketeering and astronomy in Christ Church Meadow, near where parts of the Harry Potter series were filmed.  Watch or join in as Hackspace technical guides work to kindle a passion for science and DIY in your children and our next generation.  Be there just before dusk as the Hackspace Quadrotor minihelicopter launches with a payload of environmental pollution sensors for the College roof.  As night falls, we will set off homemade rockets, carve jack-o-lantern pumpkins with special tools, make sparks in our mouths using wintergreen mints (demonstrating triboluminescence), and observe the Draco the Dragon constellation and the Draconid meteor shower through telescopes in the field. 

-----


/ok ok daydreaming over


Anyway, those are just examples of what we could get up to if we were so inclined ;)  Please excuse the gimmicky names that I made up right now ;)  Any other propositions??

We just need 3 events or "projects" that we propose to do to eventually get our name out there.  If the enviro sensors could be one, anyone have ideas on others?  What would you love to see us do??



(note: December 13 at 9-10pm might be a better meteor shower, 50 per hour, but anyway.  See http://earthsky.org/astronomy-essentials/earthskys-meteor-shower-guide for interesting meteor shower info!)



Lauren

[Nick Ryder]

Eileen,

I must admit that I hadn't even thought of asking the council for permission. For a large scale/long term roll out of some sensors getting permission would definitely be the right thing to do. Thanks for offering to approach the council for us, I hope we need to take you up on the offer in the not too distant future. I think the right time to do so will be after we have an actual sensor package that we want to deploy so we can show them and explain exactly what it is.

Lauren, 

I also hadn't thought about measuring air pollution. Off of the top of my head I cannot think of a simple way to measure this but we should look into different possibilities. What I was thinking of was basically measuring the noise and light levels (which are arguably both pollution of different kinds) at different places. The advantages of measuring these are that they require pretty cheap and simple electronics. 

For example I fond this light sensor: http://uk.rs-online.com/web/p/colour-light-sensors/6424339/?  The £2 chip can measure total as well as red/green/blue light levels. Having the colour measurement is interesting because street lighting is usually orange and so it could be distinguished from natural lighting by looking at the different colour levels.

Another measurement that may be quite simple/cheap and could be very useful would be a measurement of the traffic. I was wondering whether it could be done with an ultrasound range finder from the side of the road. vehicles could be detected by seeing a transition from the maximum measurement distance down to a distance a little further away than the edge of the road. This could be used to count the rate at which vehicles are passing, or could be used to estimate the speed of the traffic too if you measure the time each vehicle takes to pass the sensor.

I really like the idea of a map based display of whatever data we measure, especially if it is in approximately real time. This adds some complications to the device we'll design since they need a way to communicate back their data regularly, but it's something I'm looking into. I think this kind of project could really help the group if it has a well designed, easy to use public web interface. Especially if we measure something "good" or useful then it can really project that we're an organisation doing things for the benefit of the local area (as well as for fun).

There are a lot of possibilities with this kind of project and I'm really glad that a few different people have started contributing ideas and offers for help. I've started a wiki page:   http://wiki.oxhack.org/wiki/Project:SensorNetwork where we can collect ideas in a more organised way. Please add any ideas you (anyone) have for things that we could measure, ways of measuring them, how we should share the data, etc.

Cheers,

Nick

[Adrian Godwin]

There was a project done in Amsterdam recently that might have some
useful ideas for you, driven by someone from London hackspace who
moved there. Seems as though it was repeated in NYC.

https://groups.google.com/forum/#!msg/airqualityegg/EVkTG4WxJ5c/Y3U0RgWXAU0J

-adrian

[Eileen]

Thanks very much, especially to Nick and Lauren,  for their ideas on this. Before I migrate over to the new wiki - actually there are lots of websites out there looking at how to measure air pollution, not least here, which might give you one or two ideas http://www.tsi.com/air-quality-monitoring/. We could at some stage even consider doing outreach to schools (though it would be pretty labour-intensive) using simple pollution-testing equipment, along the lines of the project outlined here: http://www.sciencefairadventure.com/ProjectDetail.aspx?ProjectID=112. That might be for a much later stage, but could certainly raise our profile.