Animal tracking is among the oldest forms of science in recorded history, its origins dating back tens of thousands of years. Despite its importance to understanding the world around us, trackers have long struggled with issues like incomplete records or poor connectivity in the field. These trackers are collecting precious data that can help conservation efforts, and, as such, they need access to real-time information and historical records.
This information is particularly pertinent, given the rate at which climate disasters are threatening the survival of at-risk species. Trackers need to be able to quickly identify signs of animal life, like paw prints, droppings, or trail paths, and seamlessly communicate with others in the field. This real-time connectivity is critical to the future of this ancient art form.
Researchers from Hewlett Packard Labs, HPE Aruba Networking, HPE Ezmeral, and HPE GreenLake Cloud Services Solutions have partnered with South African expert tracker Alex van den Heever to help preserve the ancient art of animal tracking. Together, the team is using technology to help connect people with nature, raising awareness on issues like how climate is impacting the local ecology.
A primer on animal tracking
It sounds simple, but what actually is animal tracking?
Tracking is more than just watching animals move from place to place. It requires the knowledge and skill to recognize, interpret, and follow signs of animal life. Most crucially, trackers must be able to know what the animals are doing without seeing them. Trackers are looking for three key signs: What does this animal eat? What habitat does it prefer? Where is it going and why?
There are two key components of tracking. There is "track & sign identification," which is the study of signs left by invertebrates, amphibians, birds, reptiles, and mammals. The second component is what's known as "trailing": following a set of tracks (trail) through the landscape over varying substrates for extended time periods, until the animal is found.
These foundational elements have been at the heart of tracking since its inception. Animal tracking began in Africa, with early humans relying on their memories of where they’ve seen animals to track their movements. This marked an early occasion where science was required to solve a necessity in life – finding food.
These early trackers displayed the remarkable abilities still needed today to be effective in the field – a good set of eyes, physically fit, endless curiosity, and a desire to think critically to solve a problem.
Through the advancements of modern-day technology like GPS tagging and digital photo sharing, trackers have become much more technologically competent. Today, trackers must be able to understand sophisticated analytic programs and use these systems to conceive new scenarios for animal behavior.
The field does come with its own set of unique challenges. Oftentimes signs like tracks or droppings can be extremely similar among species and hard to differentiate. Trackers also run into issues of connectivity – causing challenges for recording species’ signs. Given that many animal tracking efforts are around endangered wildlife, trackers are often working in remote areas with little to no network access. Without that connection, trackers can’t record and share information on the species in real time or communicate with others in the field.
Solving tracking needs
This need has been top of mind for van den Heever since he reached out to Hewlett Packard Labs in September of 2022.
Van den Heever is the general manager of Tracker Academy in South Africa, a not-for-profit organization that trains disadvantaged rural people in traditional skills of wildlife tracking. Tracker Academy needed a digital solution that trackers can learn to use to record data from wildlife in the field. Despite technological advances, communities of traditional trackers have been declining over the years, and van den Heever needed a truly effective tool that would help spark interest for anyone interested in nature.
Despite taking his idea to other agencies and businesses, van den Heever could not find a partner who could help him reach his dream.
“Animal tracking represents the earliest form of using deductive and analytic reasoning, and learning it has been a lifelong journey for me,” said van den Heever. “I needed a partner who could address the goal of creating a mobile application for tracking to share with the public, including those tasked with protecting and monitoring endangered species. A mobile application will give the public an opportunity to engage with these ancient skills and be able to interpret animal’s sign without actually seeing it.”
Van den Heever and Dobias van Ingen from HPE Aruba Networking discussed this solution multiple times over the last years. They believed that by combining recources from HPE Aruba Networking, HPE Ezmeral and Hewlett Packard Labs it would be possible to tackle the two highlighted issues – using AI/ML for analytics, and for connectivity and positioning. Together they kicked off a team that included skillsets from the broader HPE ecosystem. Labs’ researchers brought in HPE Ezmeral to assist with the analytics piece, and HPE Aruba Networking to provide advice on connectivity and positioning. Together, the team met with van den Heever to outline how to execute on the vision.
Puneet Sharma, director of Hewlett Packard Labs’ Networking & Distributed Systems Lab, assembled a team for the project, including young researchers who were excited to be a part of it.
“This project gave us a chance to really help the greater good,” said Shivang Aggarwal, a Labs research scientist who worked on the team. “We were so excited the moment we saw just how big of an impact something like this could have.”
Van den Heever’s goal for the project was to create a digital application that end users could use to take a picture of an animal’s track or sign and then instantly be able to identify it. Van den Heever met with HPE team members to share his vision and hear how they can help him achieve it.
The backbone of an application like this is the analytic infrastructure it’s built on. Van den Heever stated that the application needed to act as a digital repository of crucial information and be accessible to anyone curious about animals beyond trackers, like citizen scientists, tourists, and students. The app needs to address the core challenge of identifying animals in Africa without actually seeing them. Despite attempts by many over the last decade, no tracking app has been able to support this need.
To tackle this, HPE Ezmeral Unified Analytics was tapped to handle the analytic framework of the solution. Ezmeral’s fabric offered a way to operationalize a wide array of cloud platforms in a data fabric/mesh layer. This enabled support for traditional and advanced analytics needed by van den Heever’s trackers in the field. This offering was the first solution van den Heever found in his years of searching capable of meeting his analytic goals.
Building upon these analytics, van den Heever wanted users to be able to take pictures of indicators like footprints or droppings, and instantly be able to identify them. Labs was tapped to address this through the use of machine learning.
“We’ve seen the success of apps like eBird that leverage AI models to help users identify bird species, and that technology is going to shine a spotlight on our ancient artform,” van den Heever said. “AI and ML are going to help create a new wave of citizen scientists to bring tracking to the mainstream for the first time.”
Labs researchers worked to develop a model that would extract as much information as possible from the image, reference it with the data analytics provided from HPE Ezmeral, and then return a result to the user. The ML model can also help trackers differentiate between indicators that look similar among species, like paw prints.
“If I know my research team well enough, they are basically technical geeks, like Sherlock Holmes – the more layers and challenges, the better,” Sharma said. “They really enjoy the technical challenges and finding the best way to the end results.”
To solve the challenge remaining of using these features in hard-to-reach areas, the team tapped HPE Aruba Networking’s connectivity expertise. Having already worked with Aruba on previous tracking projects, van den Heever knew the HPE unit could solve connectivity issues in complex environments. By leveraging HPE Aruba’s networking capabilities, animal trackers are able to upload photos share data to the data repositories and get the identification, all without a live web connection.
“This app is about so much more than just footprint identification and keeping tabs on where animals are,” said van den Heever. “It’s really about understanding our entire ecosystem and how it connects to broader issues like the impacts of climate change.”
The ancient art form of animal tracking needs modern-day digital support to continue on for years to come. Despite declining numbers, van den Heever and the HPE team are hopeful that projects such as this animal tracking app will help regrow interest and preserve the practice.
The opportunity to work on this project has not been lost on Labs team members, who feel a personal connection to it.
“If there’s anything we can do to help the broader community, that’s a great opportunity for us,” said Lianjie Cao, a senior research scientist, who works on the project. “It’s not every day that you see something you work on have a direct impact in the world.”
HPE Aruba Networking’s van Ingen sums it up well.
“In the age of digital transformation and with the power of AI/ML on HPE Ezmeral platform and HPE Aruba Networking technology we unveil nature’s mysteries and protect its magnificent creatures,” he said. “Guided by an amazing team of brilliant minds consisting of Hewlett Packard Labs and Alex van den Heever, we harness the capabilities of artificial intelligence and leverage advanced networking solutions to embark on a transformative journey of animal tracking. Let us unite in our mission to safeguard the beauty and diversity of out animal kingdom with technology and teamwork paving the way.”
