I work for a technology development company (Cambridge Consultants) and every academic year our interns run their own projects in small teams to develop a new technology proof of concept. I want to put forward a challenge for them in the conservation tech space, so I wanted to ask this community if you have any ideas along the lines of: 'if someone could invent a system that lets me ... , that would be a game changer'?
The best projects for us would be ones that need a combination of the physical and digital, something that needs cutting-edge innovation across maybe: Sensing, AI, Wireless comms, Robotics, or even Bioinnovation.
I know this probably isn't the best way to uncover the big challenges that most need this kind of innovation, but many heads are better than one, so any suggestions are much appreciated :)
31 August 2022 2:37pm
Oh my, I have about a million things... lol I should also say none of what I'm mentioning is novel, and there are groups working on each, but there's nothing ready and out there.
Super dream scenario that I'm not entirely sure humans even possess the technological capacity to do - some way of making a static sensor (e.g., camera trap, acoustic recorder) "permanent" in the sense that you only need to put it up and it will go for as long as you want, that doesn't include solar panels. I work in the rainforest, and the only way you get solar panels to run is if you climb above the canopy which is laborious, requires experts, extra costs/equipment, etc. Wind? Water (if near a river for example)? Piezo type stuff? A ridiculous battery? I don't have the answer but I'm hoping someone does eventually And in that same sense of not having to go back after deployment, using edge computing to process data on-device & transmit pertinent data back to a database as well as general stuff like "is it still running." This would be particularly useful with ML models tuned to e.g., gunshots, chainsaws/ax hits, etc. There are some one-off one-stop-shop "permanent" station projects, but they're all fairly bespoke and not necessarily off-the-shelf. Check out the conservation tech directory for some existing solutions (Conservation X's Sentinel, for example). The places I work in in Madagascar are incredibly remote, like multiple days of hiking just to get to the site. So minimizing how often we'd have to go back would save SO much time, effort, money, etc.
More (and better) no-code or little-code solutions that are easily implementable by non-AI people. There is more and more work being done on this but its in its infancy. Ecologists are used to R, and the field teams we work with often don't have coding backgrounds so something easily implementable, with multi-language support, etc. that can be easily adopted (SMART is a good example) and run across OS's, hardware, etc.
I know there has also been talk from the 'Tracking Progress' project on developing some sort of tag database. There is also the always-mentioned 'making tags smaller' argument coupled with 'making tags last longer'. There is of course a compromise there but getting as close to the best of both worlds (bonus points for affordability) would be a game changer. GPS collars even for large animals are also still fairly prohibitively expensive for large-scale deployment.
Also, the perpetual request for applying the "audiomoth model" to camera traps. E.g., a trail cam made by & for wildlife ecologists/conservationists rather than the hunting/security cams we use now. Again, groups are working on this but I haven't seen anything on the market.
Underwater camera-trap systems are also much needed!
Happy to chat with you if you want me to make better sense of what I've just blabbered on about! Feel free to DM me.