Conservation and Engineering Communities: How Do We Help?

Over the course of 2020, Hackster.io came together with tech leaders across the industry to help tackle something big — wildlife conservation. With the help of our partners and leadership from Smart Parks, the global elephant population was identified as particularly vulnerable wildlife that we could help grant relief to.

OpenCollar, a Smart Parks-funded initiative, has been working to engineer and deploy tracking collars on elephants since its launch in 2019. These tracking collars are potential data gold mines where conservationists can organize, decipher and report on live data feeds during monitoring and protection efforts in the field.

Together with Microsoft, Avnet, Edge Impulse, Smart Parks, and many others, we launched the ElephantEdge engineering challenge. We called upon Hackster community members and passionate developers around the world to build advanced machine learning programs with the goal of helping park rangers analyze the data coming from their elephant collars.

The results blew us away.

After three months, our challenge received 31 data analysis programs that use a combination of ML, AI, and IoT to do everything from tracking elephant migration to listening in on elephant noise patterns. Wildlife and tech experts from across the industry chose the top 10 solutions to test when the first version of the collars are made available.

It took just three months for these software solutions to go from concept to final-stage, ready-to-use programs.

The Hackster community has always had a way of working at warp-speed when presented with a challenge. In the last year they've designed technology-focused solutions to help detect and protect against COVID-19, alleviate symptoms of menopause, and fight wildfires with drones. Now the community has played a part in giving the global elephant population a fighting chance. If the Hackster community can make a difference in ways like these, it begs the question: what role do open source communities play in worldwide conservation efforts?

To answer this question, we must first understand why open source engineering communities exist, what they offer, and where they fit into the world of conservation. Arguably the most important reason for the existence of open source communities is the wealth of knowledge that is shared. Open source hardware and software allow for free distribution, without the demands of paywalls or voiding warranties (mostly). Access to a community-driven bank of source code, hardware specs, design instructions, and getting started guides means that anyone can learn how to use this type of technology.

In addition to knowledge, open source communities offer just that - community. Like-minded engineers, makers, and developers with a unified goal of sharing knowledge can consume information from their counterparts, ask questions and learn on a single platform.

Our ElephantEdge contest is an excellent example of how communities can harness open source technology to make the most significant impact. Traits of communities like these fulfill a few of the most critical conditions we see with global conservation efforts, including:

Speed: Ever-changing conditions of our planet require solutions that are deployed rapidly and can be easily updated as needs evolve. The ElephantEdge challenge brought a big cause to the community and solutions were finalized after three months. ML and data analysis tools like Edge Impulse and Avnet's IoTConnect platform helped community members build dashboards to support park rangers's efforts to understand data received from tracking collars. Speedy development, speedy deployment, incredibly powerful solutions for tasks that may otherwise take years. These solutions, once deployed with the elephant collars, can be updated and enhanced while deployed in the field. More data in conjunction with better quality analysis means we’ll understand the elephants better with every data point collected. It only gets better.

Low barrier to entry: Open source communities make it possible for anyone, regardless of age, country, education, or experience, to learn complex concepts like robotics and 3D printing. Community driven topics from beginner to expert means members can share and consume knowledge without being limited by skill set or experience - one of the oldest barriers to entry for this industry. An “all hands on deck” approach to wildlife conservation at this scale is critical and the lower barrier to entry communities offer is the best way to turn passionate people into developers who can help.

Collaboration: Tech industry leaders offer resources, educational opportunities and networks to help enable developers communities. Communities have the passion, skillset and curiosity required to innovate solutions for humanity’s greatest challenges. Why not create platforms for these parties to come together? Our contest was possible because of the support we received from over 10 industry partners. This level of collaboration in our industry is necessary to elevate conservation efforts across the globe.

These communities aren’t the turnkey solution to global conservation efforts but they offer a non-trivial path forward through small and meaningful collaboration. With more collaboration, awareness, involvement from engineering communities and industry leaders alike, and a continuation of the points above, real change could come — and the sooner the better.

I worked with a few of our ElephantEdge contest winners to get their perspectives on open source technologies, their inspiration for joining the competition, and more.

What about the ElephantEdge contest inspired you to compete?

Swapnil, UK (winning project): Most of the projects I have prepared were like a prototype but through ElephantEdge I had an opportunity to change that. ElephantEdge gave me the chance to develop something which could be used in a real product, which will solve a real problem and that too related to animal conservation.

What is the feature that sets you apart from other winners?

Dhruv, India (winning project): The acoustic models that I built included four different models. The first model included the acoustics that were used to understand the condition of elephants whether aggressive, defensive, assurance or soliciting, which set apart the acoustic data type to to understand the emotional behaviour of Elephants.

How did the learning models (Edge Impulse, IoTConnect) help you take your project from concept to prototype?

Dhruv, India (winning project): It was solely Edge Impulse which helped build the project from concept to prototype. EdgeImpulse not only helped to improve efficiency and reduce size of the model but also provided a complete platform to build the model using in-built tools to easily build acoustic, accelerometer and vision models. This platform provided a completely optimised and developed model with the dataset and included deployment for different microcontrollers to process the TinyML models.

How do you think what you built is scalable? Do you think it could be applied to other industries? If so, how?

Swapnil, UK (winning project): With more and different types of data we can increase the capability of our algorithm. It will require no to very little change in the network architecture. Our "Human Presence Detection" algorithm can be used in the detection of any specific sound, for example, gunshot detection, chainsaw detection in prevention of illegal tree felling, abnormality in a heartbeat using stethoscope, etc. The "Elephant Activity Monitoring" algorithm can be used in the detection of any action performed by anyone or anything, for example, handwritten word recognition, gait recognition, etc.

What type of role do you think open source communities, like Hackster, play in worldwide conservation efforts?

Ankur, India (winning project): I'd like to think that with open source communities, we accelerate the development of conservation solutions. Like now after the ElephantEdge challenge, we could take the best parts from each submission and put together an even better creation. Some advanced research areas- like to identify individual animals and species- are extra difficult and require us to collaborate.

What’s one piece of advice you would give other developers interested in conservation technology?

Sumit, India (winning project): From camera traps to satellites to drones, conservation technology is increasingly critical to ensuring that people and nature thrive. Whomever comes to the conservation technology space, their aim should be to maximise the effectiveness of technology for conservation, ecological research, and citizen science by providing experienced as well as new users with accessible and freely available, peer-edited guidance.