Project to help save endangered animals, I could really use some help!

Hi my name is Arnold, a few months back I saw a TED talk about Tapirs (a mammal about the size of a large pig). I volunteered my time to help design some tech to help the researchers who study these animals and help prevent their deaths. I managed to get in touch with them, and they were so thrilled they invited me along for a week of getting sucked dry by ticks and mosquitoes to observe how they work and what their problems were.

I came back from this trip, full of ideas and projects that could have a huge impact, maybe save billions of animals worldwide. At the moment I'm trying to develop a very cheap IOT type GPS collar that could slash tracking collar costs . They cost a fortune in 3rd world countries, right where most endangered species are!

Secondly im trying to develop a ,cheap and easy to deploy, solution to reduce roadkill. I was shocked to find out more that 400 million , yes 400,000,000 animals are run over every year in Brazil alone! In the one week I spent in the field with the researchers I saw countless dead tapirs, giant ant-eaters, armadillos, tucans(frootyloops bird).It's INSANE! Thrid-world governments don't give a crap , the NGOs doing the research are underfunded and lack the knowledge of designers and engineers, so I feel I must do something.

Both of these ideas are still on paper, so hopefully if enough people help we could have a prototype ready in a few weeks instead of a few months.

I'm doing this for free, Im going to help them not for status or profit, and the NGOs agreed that if we succeed to create a solution , that we would do everything opensource and non-profit.

If your willing to help please get in touch. You don't have to be a arduino ninja or be in Brazil to help, ANY help is welcome. I intend to do this sticking to arduino and xbee so it can be further developed openly.

My email is rudge.design@gmail.com

My name is Arnold ( Arnaldo in portuguese). I live in Atibaia near Sao Paulo in Brazil.

Here is the link to the TED talk that got me involved.

http://www.ted.com/talks/patricia_medici_the_coolest_animal_you_know_nothing_about_and_how_we_can_save_it

Iniciativa Nacional para Conservação da Anta Brasileira (INCAB)

Ok. I have deleted the entire post I had just written, and believe me it was a beaut. But I can see it went a little over-the-top with the skepticism. Here's the main thing:

Both of these ideas are still on paper, so hopefully if enough people help we could have a prototype ready in a few weeks instead of a few months.

More info, please.

What would you like the people here to help design and build? You will not get a lot of feedback here until you take these ideas of yours off paper and either put them in a post, or put them on a website somewhere. "Oh, I got an idea, who want's to commit time to building it?" really isn't going to cut it.

Your post spends a fair bit of effort outlining the problem, citing various pearl-clutching statistics. Guys here aren't really motivated that way. You have to outline what action you intend to take, you have to suggest a concrete project.

What, exactly, would you like us to make your Arduino do?

He wants someone to help him develop a GPS collar that transmits the location of the animal wearing it via... something. I don't know what.

Anyway, the devices exist. I've considered developing one for horses because I live with horses and I have an application where it would be useful. But my needs aren't pressing enough to move forward on what is such a time-consuming project.

In the OP's situation he will run into multiple problems. The biggest of which I see is the availability of a communications channel: there's no WiFi in the jungle and leaves are excellent at absorbing the frequencies that WiFi or cellular would work on anyway. Then there's the fact that animals don't like having stuff strapped to them and you need to worry about how to secure it.

Now, as far as the roadkill problem is concerned, if he frames it as protecting vehicles from animals, he may get far more support. There's a lot of roadkill here in the Midwest, but most people are more concerned that the vehicle that they need to get to work was totalled than that another deer is dead. All of my vehicles have either hit, or been hit by, a deer at some point. I'd love a solution!

Thanks for the reply guys, sorry if I didn't go too deep into my ideas, I didn't want to make it "my idea" but I was hoping to find people to think of solutions team effort.

first the problems , then my idea

Iridium type GPS collars, are 3000 usd + take taxes, import costs, it's outrageous. Add on top of that, transport, salaries, accommodations etc. so it comes out to 5000 usd per animal caught and tagged. Given our exchange rate it's like buying a new Car, that's how much it hurts the pocket. NOW imagine if that animal then manages to take it off, or it gets stolen! Then it's like setting fire to your new F150 after one ride around the block.

VHF collars, are half the cost, but require a lot of work to triangulate, time, gas, and gives you a hell of a lot less real time data, valuable but not as sexy as getting data every 24h.

Gsm, good option where there is service, but it's still expensive(in our 3rd world reality). Many areas don't get gsm coverage, especially the pristine and hunter prone areas.

Store-on-board, cheap, but you have to physically recover them. many times impossible.

Okay so I suggest using an zigbee type rf module with a large store-on-board system. So I imagine a system where every animal that wears our collars passes along and stores the GPS data of every other animal that it has come within radio distance of.

So one animal might carry the data of another 20.

Then a few select animals and a system of towers placed around places known to be frequented, such as palm trees patches, relay the information via Iridium or a more powerful RF transmitter. Once the info is relayed then an erase memory pointer is passed around freeing up memory.

So many cheap collars with only a gps reciver, zigbee compliant module, and atmel uM and maybe a 1gb of memory. I estimate they could be 1/4 to 1/15th the cost. So they could tag maybe 10 times more animals. It would have more redundancy, be cheaper to run,build,maintain, gather data and it would be open source and manufactured at cost at any maker space etc, anywhere in the world.

Given a large enough net, we might get parcial but frequent coordinates and vitals ,with enough overlap to make the data as good as the data from a 2 way gps collar.

There is a lot more to this idea but this is the crux of it.

arrumolane: So I imagine a system where every animal that wears our collars passes along and stores the GPS data of every other animal that it has come within radio distance of.

Holy balls! Mesh network for tapirs!

Ok, I am glad I didn't post my other comment, because that is a very, very interesting idea. Perhaps the most notable thing about it is that it leverages new hardware. People have been radio tracking animals for ages, but cheap 2GB storage cards are relatively new. This is do-able today in a away that wasn't possible before.

Now, a thing that occurs to me is that you aren't storing that much data. Not unless you are planning to put a camera on these collars. This means that a lower radio frequency will have sufficient bandwidth to do the transfer, and lower frequencies travel better. Bluetooth won't cut it - it has a range of a few meters.

A problem is antenna: lower the frequency, the longer the antenna. I'd suggest taking this to a ham radio enthusiast board. You are going to need a small team to pull this off: electronics, fabrication, programming.

Oh - if you are not using bluetooth or whatever, then legals becomes an issue. You need to find a radio band that can be used for telemetry.

As to the programming:

  • I'd be inclined to store your data in blocks, and to take a 4 or 8 bytes message digest of the block. … actually, not necessary. A collar #/block serial # would be sufficient to uniquely identify them. Collar number == mac address? Same idea, really.

  • When the collars communicate, the salient info is what blocks they have, and whether the blocks are known to have already been uploaded. Blocks known to have already been uploaded are marked can be erased from the storage card.

  • Hmm - if we are using serial numbers, then you could store a range rather than the individual numbers. That would work better.

  • You know what this is very much like? Bittorrent.

  • I wonder if it might be possible to rank the collars over time in terms of how well-connected they are to the upload point. As the network becomes aware that some particular collar is on the fringes of the network and its transfers are spotty, its data becomes marked as 'precious'.

  • I wonder if it might be possible to implement this as a very … very slow IP network. That is, the collars rank themselves in terms of how "close" they are to the upload point, and make an effort to transfer data from more distant collars to closer ones.

  • Along with the GPS data itself (why not add a light sensor?), a salient bit of info is how long collars spend in radio contact, how good the connection is. That alone might reveal tapir social networks. That is: if the GPS units are pricey, you could get interesting data out of the system just with radio signal strength.

All this has probably been done before, though.

http://gizmodo.com/turn-your-valuables-into-a-mesh-network-so-you-never-lo-1683325737 - probably too short range and expensive

https://www.movebank.org/ - very interesting. Why not make this database or another one like it your "home" for the data? It can be integrated with other data on other animals.

http://imos.org.au/animaltracking.html - more interested in marine info. Accoustic, rather than radio? Works underwater, I suppose.

https://en.wikipedia.org/wiki/Wireless_Medical_Telemetry_Service - focuses on legals in the USA. Typical of the kinds of issues you face.

I wonder if the guys over at CSIRO already do something like this. If they already have this gear, and you are affiliated with a well-known NGO, maybe they can do you a deal.

Woah! It seems I went off half-cocked when I wrote my reply:

arrumolane: So many cheap collars with only a gps reciver, zigbee compliant module, and atmel uM and maybe a 1gb of memory. I estimate they could be 1/4 to 1/15th the cost. So they could tag maybe 10 times more animals. It would have more redundancy, be cheaper to run,build,maintain, gather data and it would be open source and manufactured at cost at any maker space etc, anywhere in the world.

They already make 'em? Of course they do! A whole world I didn't even know about. So the only problem becomes actually doing it.

Yeah digi has a great solution, there is also a very cheap , but shorter range module (+/- one mile) from microchip that seems interesting. I need help tackling the complexity of the protocol need to manage incoming messages and designing network flexibility on top of those which are built in to zigbee. Also i'm not that knowledge able when it comes to battery life management and charging. So since I've been working with this sort of stuff , I've only handled simpler systems that didn't need to be so robust.

Guys, I don't mean to rain on your parade, but anything you do will require budget.

Perhaps you should start by deciding how much you can afford to pay for each device, installation and the supporting infrastructure before running off with a "solution?"

I'm not too worried, as it might be a little more or less than 300 usd, depending on batteries, but it's sure gonna be cheaper than 3000. my BOM came out to 170 usd min to 780 usd max, the routers used to relay the signal is a standard comercial system at anywhere between 500- 2000 usd. so if we have more than 10 animals in any given area it's a hell of a saving. Anyway that's why Im asking for some help here to harness experience to build reliability and bring down costs further.

Hey paul , you nailed it! i imagine using, less powerful, cheaper modules and having a "relative position system" so as to make sub 50 dollar collars. As an animal comes within a mile of each other they can transfer data, so we a drastic saving in antenna and battery costs and add another 2 gb of storage, and bob's your uncle. Sure the data from these would be less "rich" , but you would still get a good positioning , and you do get social interaction status, and they still serve as a bridge or node , in the network, multiplying the chances of frequent upgrades. I'm starting to put together a team, I might even be getting funding from conservation projects, I've sent out what must be 100 emails by now , and it's looking good.

But so far you seem to get it the most! Could really use your input, can I add you as a contributor? I've also thought of adding transmitters to inter municipal, or inter state buses that cut right through these areas , mostly isolated roads, with a only few trucks per minute.

The buses would pick get up dates from animals along the roadsides. Buses travel regularly through these roads maybe 5 times a day , and they are usually the same vehicles. I'm even getting help from a drone maker, to get aerial updates, which got me thinking of putting relay collars on big birds like condors and vultures. There is a Peruvian project that did just that . Ill send you the link when I remember the name.

And as a long shot, building cheap irrigation systems for local farmer with built in network modules "ninja modules", quietly working way, like the cuckoo, get the the farmers , who usually hunt these animal to inadvertently help save them by powering and maintaining the nodes.