discussion / Open Source Solutions  / 23 March 2019

What would an open source conservation technology toolkit look like?

Hi. We've had a nice discussion about this topic in a different thread and I wanted to break it out into a separate discussion thread. For context, here is the previous discussion:




[hackerfarm] 

Would it be interesting to brainstorm what an ideal open source device toolkit would be for wildlife conservation tech? Not just one device, but a suite of tools that are open source, customizable, and can be used for various tasks specific to wildlife conservation? Something along the lines of a universal datalogger, camera trap, acoustic monitor, wildlife tracker (Small, Med, Large) that has a base design and can then be modified for specific instances? I'd love to discuss something like this and how it could fit into the larger systems of data science, machine learning, etc. 

[Rob Appleby]

hackerfarm, are you reading my mind (given your interests, I think this is a distinct possibility)?!?!? That's exactly what we need! I am loathed to reinvent the wheel here though, so perhaps we need to start by consolidating what's already available etc., and work out what's missing? 

One thing we've been working on and off on (this is setting up a pun) is a low power timer switch with multiple independent channels. It's still very clunky, but we have managed to get it to turn a GPS on/off on one schedule, a VHF on another and finally a drop-off once. It's all up on Github: https://github.com/Wild-Spy/OpenSourceTimer but still needs work. We thought a universal timer with pretty sophisticated shceduling could be useful (but there's also the cheap but great units from Adafruit). 

What are your thoughts on it all? Some sort of super-powered modular AI Arduino (or AIduino if you would)? I reckon powerful and modular is one route, but another is super low power. 

Cheers,

[hackerfarm]

Actually, I agree that there shouldn't be reinvention of wheels. I think that's the idea for having things be open source. If that's the case, then we can build on each other's shoulders which makes things infinitely faster. 

The timer switch sounds amazing. Timers are useful for all kinds of things. We use them for our animal feeders and our irrigation switches. You said it still needs some work. If you have ideas on what your ideal low power timer is, I can offer some design advice on implementation. I looked through the github and it looks like it's quite well developed. 

I'd like to start a conversation on what an ideal toolkit is. It could be things that exist already and it could be a wishlist of devices and features. The important thing is to start putting it together and gluing it together with application software. It would be amazing to have a toolkit like that. I think that there's enough design skill here and domain knowledge that we can put something together that can enable some pretty amazing things.

[Rob Appleby]

I totally agree. We've almost gotten a GPS LoRa off the ground using Feather boards, but as always, there are bugs. Happy to share that too as it's probably fairly close. It's almost certainly a code issue(s), as it's mostly a problem getting the correct date read from the GPS data stream (and several rather terrible problems from ham-fisted attempts to fix it). I can explain more...

Re: the timer, we tried to build as much of it as we could before we ran out of development funds. We got the basics up and running, but never really got to test it properly. It's a difficult board to hand assemble (there are some parts that are just way too difficult) and we didn't have enough to do even a small run. Utlimately, one idea might be a "no guarantee" or "this is strictly a proof-of-concept" Group Gets or something to get enough on order to make it cheap for a few people to try. I'd love to add some sensor inputs to make it a lot more flexible. 

I should stress, I am not an engineer, and unfortunately, our engineer has moved on to greener pastures. I might be able to coax him back for a bit of a brainstorm re: further design etc., but it might be a case of only being able to work with the stuff that's already there. Maybe if you peruse the repo and find some glaring mistakes or missteps etc., I'd be super grateful, as I am for the offer on design help. I'd love to see this damn thing take off a bit. Ultimately, I'd love to publish an open-source paper on it too, if that's of any interest. 

So what sorts of things would you put in a toolkit? Something that sort of works like Blockly might be good for non-engineers to piece together bits of code for sensor add-ons etc. And as sensors or modules get added, the hardware and assembly instructions can be updated somewhere/how (I am picturing something like Fritzing). Matt, our old engineer made a nice Java interface for scheduling the timer, where you could essentially write in english what you wanted (there were a few rules), and then a timeline would appear in that particular channel, so you could check if it was the schedule you wanted or not. That sort of intiutive, building block approach could be useful for noob wannabe engineers like me. Looking forward to getting your thoughts, and (hint hint Talia) this is looking like a pretty cool live discussion topic maybe...

I think this is a bit of wishlist thinking right now, but a  possible open source device toolkit for conservation tech seems to be:

  • General purpose timer that is customizable and can control different devices
  • General purpose datalogger than is customizable and can interface to different sensors
  • Open source camera trap made from standard components
  • Open source acoustic sensor for doing acoustic surveys (AudioMoth)
  • Open source small wildlife tracker (< 1g)
  • Open source medium sized wildlife tracker (what would be the constraints?)
  • Open source large sized wildlife tracker (OpenCollar ?)

I would also say that being open source isn't enough. There would need to be enough information via video and written tutorials to teach people how to program and customize them. I think putting together a toolkit like this (with some devices already in place) would be really interesting and exciting. Any comments, additions, or feedback?

This would be amazing! And it'll be a huge undertaking to get them to a product quality that just works. While I am with you that all the open source tech needs proper documentation that includes written and video tutorials, for a lot of conservationists it should just work out of the box without any soldering or coding (even Arduino). At most what they're willing to do is configure it with a few switches/knobs or an app or a website.

Also this toolkit must be developed with close interaction at every stage with the end users, the conservationists. And many of them. We've gotten into the trap of developing tech in our own silo and then showcasing them.

We've quite some experience developing open source animal sensing tech for camera trapping. And we're working on raising funding for developing a camera trap for conservationists.

 

Hi PrithviRajNarendra. Yes, I agree. That's why I think it's good to get the conversation started now. There are a lot of wildlife researchers and there are a lot of techies. What is the intersection between what's needed and what's possible? And the intersection between what has to be out of the box functionality, and what can be customizable. It probably won't be right on v1.0 but if we start talking now and iterating, by v3.0, we should have something pretty decent :)

Akiba