DIY active cat collar GPS unit?

[casper.bang]

I've often wanted to track my cats wandering, but were unimpressed with the commercial solutions available. The primary problem comes down to size, none of the solutions I have seen, are anywhere near small enough to seamlessly fit onto a collar. A secondary issue, is how many solutions only log data rather than actively pushing it out to a receiver, which is needed for any kind of live monitoring or locating an dead/injured cat.

I just came across this post and was wondering whether people on this forum tried a similar DIY version in the 10-30 gram range? As a post-xmas project, I'm tempted to look for the smallest GPS module I can find, hook up an ATtiny and find a way to transmit location data via free-licence bands (27MHz/40MHz/433MHz/866MHz) but some feedback greatly appreciated.

The GPS module would not need to be ultra precise nor offer a high update rate, but apart from small size it should be fairly easy to use.

[krenzo]

There's a user on here by the name of "joelby" that was working on a project to track animals.  You might try sending him a PM to see how far he got.

[owl_uk]

have you thought about logging the positions periodically for download later, rather than live tracking?  I get that live tracking is cool, but it might take quite a lot more power and cats have a bigger range than lots of radio technologies.

[AndyC_772]

The BBC made a programme along these lines earlier this year:

http://www.bbc.co.uk/blogs/tv/posts/Horizon-The-Secret-Life-Of-The-Cat

[casper.bang]

have you thought about logging the positions periodically for download later, rather than live tracking?  I get that live tracking is cool, but it might take quite a lot more power and cats have a bigger range than lots of radio technologies.

Certainly, being a huge fan of divide-and-conquer, the initial version would be focused on logging to a ring-buffer in the memory of the MCU. Power-wise, I think it should be feasible to transmit small burst of data every 30 sec or so, when using a small 110mAh LiPo cell.

[temmi_hoo]

I'm not sure how bleeding edge things you want, but here's one project: http://www.ruuvitracker.fi

RuuviTracker is an open source open hardware device you can script in a high level language and that can send the logged GPS/Galileo/etc data via cellular comms to a home base somewhere in the internets tubes cloud.

You might need to go creative when mounting anything to a cat, mine will dislodge most anything when I'm not observing.

[calexanian]

I recall seeing a book with the title of something like "How to know if you cat is plotting to kill you"  The less tech we give them the better.. They will find a way to use it against us... Or just nap and play with stuff on the floor.

[Stonent]

Since you already have the GPS coordinates, set it to log do a microSD card and when it detects the cat is within range, start sending data. You could maybe put a timer on it to only do it when the cat has been in the area for 5 or 10 minutes just to be sure that the cat isn't just passing through. Though if you used wifi, you could switch it on within range and probably upload all the data before the cat leaves range.

[apelly]

Just trigger a bark collar when the cat goes out of your garden and you won't have to worry about where it defecates. Cats are smart that way. I once taught a cat not to come near my dinner in one lesson.

[amyk]

How about GPS + GSM (GPRS) solution? Position data is compact enough that even the cheapest data plan you can get would probably be much more than enough for e.g. 10-second update intervals.

[itdontgo]

These Quectel modules we use are pretty small combined with a chip!  The only problem with chip ants is that they need a fairly big ground plane.  They might work stuck to a cat though but that's a guess.  You can get pretty good SNR of up to 50 if you design it right!



http://www.adventcontrols.co.uk/L70_HD_V1.0.pdf

[casper.bang]

I'm not sure how bleeding edge things you want, but here's one project: http://www.ruuvitracker.fi

Interesting project for sure, however the PCB is much too large for my idea.

Since you already have the GPS coordinates, set it to log do a microSD card and when it detects the cat is within range, start sending data. You could maybe put a timer on it to only do it when the cat has been in the area for 5 or 10 minutes just to be sure that the cat isn't just passing through. Though if you used wifi, you could switch it on within range and probably upload all the data before the cat leaves range.

Yeah once an MCU is onboard, it opens up for all kind of interesting optimizations. As always though, those are best left for a post-step.

Just trigger a bark collar when the cat goes out of your garden and you won't have to worry about where it defecates. Cats are smart that way. I once taught a cat not to come near my dinner in one lesson.

Not sure what you are trying to say here. Electrocuting dogs because you are not capable of humanely train your dog is pathetic and btw. illegal in my neck of the woods.

How about GPS + GSM (GPRS) solution? Position data is compact enough that even the cheapest data plan you can get would probably be much more than enough for e.g. 10-second update intervals.

Clearly a very compact and long range solution, but apart from needing a carrier and a lot of power it's also quite an expensive kit to start messing with... but I might look into this some more.

These Quectel modules we use are pretty small combined with a chip!  The only problem with chip ants is that they need a fairly big ground plane.  They might work stuck to a cat though but that's a guess.  You can get pretty good SNR of up to 50 if you design it right!

Interesting, though they look bit large. I can't help to wonder why there isn't a commercial product made, with X number of small SMD PCB's strung together inside some heat-shrink tubing.

[PaulAm]

How about the Quectel GPS L80?  That even includes the antenna.  Pretty darn small.

http://www.quectel.com/product/prodetail.aspx?id=62

There's a guy in Hungary selling the L50 on Ebay, which is a little larger for around $24 US.

I found this place:
   http://store.comet.bg/en/Catalogue/Product/18130/
Listing the L80 for $12.67 US in single quantities.  Doesn't look like he has stock though.

[lgbeno]

There's a user on 43oh.con forum doing something similar with a cc430 radio.  I think it is called hay tag.  It might not have gps though, not sure.

[itdontgo]

Forgot about the L80 yeah that's a decent idea.  Those chip antennas also have another problem you need to consider in that you need to keep them away from the enclosure by a couple of mm too not just metal surfaces.

The L70/80 are the smallest modules I've come across.

[casper.bang]

Thanks for info about the The L80, looks very interesting indeed - going to look for a supplier.

[Sigmoid]

I've already been wondering what is the smallest affordable GSM/EDGE or UMTS/HSPA+ module available... I've been thinking about remote controlling an aerial drone through cellular networks as opposed to conventional RC - it would solve a lot of problems.
I guess the same technology could be used to trace cats.

Given how the actual GSM / UMTS component of a modern cellphone is only a tiny fraction of its size and cost, I'd say it sounds feasible...

[amyk]

If you want a ready-made module look at these: http://wm.sim.com/wirellessen.aspx?id=1090
GPS + GSM, AFAIK it's based on the same Mediatek all-in-one chipset in a lot of cheap phones.

[Maxlor]

A couple of years ago, I saw a demonstration of the Nanotron NanoLOC system. It's an RF transmission system that uses chirp spread spectrum to transmit data:


The receiver uses a frequency-dependent delay line, after which the signal looks like this:


The advantages of this system are robustness against interference, especially narrowband interference (the frequency sweep uses all of the 2.4GHz ISM band), a pretty good signal efficiency, and the pulses you get after the delay line can be used for very precise timing. Because of those timing characteristics, the signal can be used for triangulation. In that demonstration I attended, they showed 1m accuracy when tracking a couple of dozen tags in an office building, despite heavy 2.4GHz interference. Outdoor range was claimed to be several km, with similar distance measurement accuracy.

So what does this have to do with your cat tracking project? I saw that the chips are available on Element 14. Maybe you could build a tag for the cat, and 3 receivers positioned at the outskirts of your property, and triangulate?

Seeing how the sender only needs to be active for an instant every, say, 10 seconds, this solution would probably be a lot more power efficient than a GPS/GSM solution, which could result in a lighter collar.

Oh well. I realize that it'd probably be a lot more effort to build something based on this technology rather than what was mentioned before, but hey, maybe you or someone else likes the idea.

[owl_uk]

I like your idea Maxlor because it removes a lot of the complexity from the cat, giving the potential for a very small circuit on the cat's collar.    How far apart would the 3 receivers need to be placed?  And how much external circuitry is needed with these chips?  I don't have a cat or a garden so I'm not going to build one, but I think it is a very interesting project.

[Maxlor]

I don't know about the external circuitry, I haven't actually used those chips myself. As I said, only saw the technology demonstrated.

As for the distance, it depends on how precise you want to be. In this situation where the distance between the receivers is much smaller than the distance between the transmitter and the receiver, you'll be able to accurately determine the distance of the transmitter, but have some error on the angle. I did a quick worst-case calculation for just two receivers 10 meters apart, and the error angle ends up being 11.5° if the distance measurement error is 2 meters. That translates to 205m at 1km distance. So, not great. But it halfes as the distance between receivers doubles, and if you have more than just two (which would only cover a 180° arc anyway) the error could be further reduced through least squares fitting or similar.

So... without spreading your receivers, angular accuracy is limited. But I can still imagine some use cases where its good enough, cat tracking being among them

[casper.bang]

A couple of years ago, I saw a demonstration of the Nanotron NanoLOC system. It's an RF transmission system that uses chirp spread spectrum to transmit data:


The receiver uses a frequency-dependent delay line, after which the signal looks like this:


The advantages of this system are robustness against interference, especially narrowband interference (the frequency sweep uses all of the 2.4GHz ISM band), a pretty good signal efficiency, and the pulses you get after the delay line can be used for very precise timing. Because of those timing characteristics, the signal can be used for triangulation. In that demonstration I attended, they showed 1m accuracy when tracking a couple of dozen tags in an office building, despite heavy 2.4GHz interference. Outdoor range was claimed to be several km, with similar distance measurement accuracy.

So what does this have to do with your cat tracking project? I saw that the chips are available on Element 14. Maybe you could build a tag for the cat, and 3 receivers positioned at the outskirts of your property, and triangulate?

Seeing how the sender only needs to be active for an instant every, say, 10 seconds, this solution would probably be a lot more power efficient than a GPS/GSM solution, which could result in a lighter collar.

Oh well. I realize that it'd probably be a lot more effort to build something based on this technology rather than what was mentioned before, but hey, maybe you or someone else likes the idea.

I looked into something like this in the past, and the NanoLOC looks great from a technology point of view. However, the price of the IC's looks a bit steep and also I am not as impressed by the relative high complexity and large size of the transponders. I would actually be Ok with a manual triangulation a la what the loc8tor [http://www.loc8tor.com/uk/primary-products/loc8tor-pet.html] offers, or perhaps simply construct several directional receivers to give me a rough N, NE, E, SE, S, SW, W, NW indication. Perhaps the easiest for me would be to simply order such a loc8tor and tweak it a bit.

[casper.bang]

It seems that the loc8tor is not easily hack-able, relatively expensive and not that great in reality. Wouldn't it be relatively easy to construct a similar unit using small modules like these coupled with a small ATtiny10 (to wake up every 4'th second and transmitting a byte sequence) and a small 30-50mAh LiPo?

I get the impression that 433MHz is the sweet spot for penetrating walls and escaping water droplets and 300 meter range sounds about ideal for my purpose. To track I would probably listen using SDR on my laptop before trying to construct a discrete dedicated tracking unit (this has worked well with intercepting data from other ISM band devices in the past).

[SArepairman]

whats even better is to put a go pro camera on cat.

i dont think its very ethical, I would not do it to my cat.. but it is what it is

[DakLak]

I bought a similar commercial device for dogs in HongKong.

I attached it to my daughter's electric bicycle and can view the track on a web site. It uses GSM cell technology. Saves me the hassle of asking where she is

Your problem will be the weight/transmission distance ratio, you don't get anything for nothing. And the cat a stiff neck.