The Internet of Things (IoT) now offers sophisticated ways of safeguarding wildlife: communication networks can cover wild places; sensors can monitor animals, humans and equipment within them in (near) real time, and control centres can collate multiple data streams, increasingly with the aid of machine learning (ML) and artificial intelligence (AI), leading to timely management interventions.
Smart Parks is a Netherlands/UK-based social enterprise that provides technology solutions for wildlife protection, with a focus on keystone species that are essential to maintaining the diversity and functionality of ecological communities (grey wolves in Yellowstone, for example, or wildebeest in the Serengeti).
« We started three years ago on a small scale in Mkomazi, Tanzania where we were the first to put LoRa sensors inside the horn of a black rhino, » Smart Parks co-founder Laurens de Groot told ZDNet. « Since then, we have built Smart Parks in Malawi, Tanzania, Rwanda, Kenya, Congo, India, Zambia, Namibia and the Netherlands. We are working with partners such as African Parks, Peace Parks, WWF and many others. We’ve also supplied our sensor technology to several existing parks with their own LoRaWAN-infrastructure, such as Hluhluwe Imfolozi in South Africa. At the moment we have over seventy rhinos and hundreds of other species under our Smart Parks supervision. »
In 2019, Smart Parks launched the OpenCollar initiative, which aims ‘to attract and inspire talented students, researchers, and tech-savvy conservationists to develop tracking systems that are more customizable and a better fit for use on different animals’. The first deliverable from this project was the modular OpenCollar Elephant Tracker (details are available at opencollar.io, GitHub and WildLabs.net), which is currently under test in Liwonde National Park, an African Parks reserve in Malawi. The open-source hardware route may not suit all organisations, however, so Smart Parks also sells a range of devices, created in partnership with product development specialist IRNAS.
The latest Smart Parks/OpenCollar project involves LoRaWAN-based GPS collars fitted to lions and cheetahs, also in Liwonde. Using LoRaWAN connectivity for predator collars allows high-frequency GPS tracking, along with battery life measured in years rather than months.
Liwonde National Park manager Craig Reid said: « The new LoRa GPS transmitter collar units, which feed into our Wide Area Network, have had numerous benefits — not only do we get near live time tracking, but this comes at a far lower cost than traditional satellite GPS collars. We see a future where we complete the transition to this new technology and have the ability to monitor our priority species accurately, reliably and for a long period of time so as to inform our law enforcement and management strategies. »
The lightweight predator trackers can handle harsh outdoor conditions and can be fitted to existing collars or provided with belting and counterweights, Smart Parks said. The trackers can be configured remotely and provide detailed status updates on temperature, battery usage and movement (via an accelerometer). This allows settings to be optimised for particular species’ behaviour patterns, ensuring that GPS positioning is only attempted when the tracker is suitably oriented to resolve a location from the satellites, which can prevent unnecessary battery drain.
The ‘brains’ behind these next-generation trackers is the LR1110 LoRaWAN transceiver with GNSS and Wi-Fi from Semtech. Thanks to its low average time to a position fix (4.5s) and low average power consumption (27μWh), an LR1110-based tracker offers greater configuration flexibility and extended battery life. For example, a large mammal module could run for 100 months taking a position fix every four hours, or for 24 months with a 10-minute fix interval. By comparison, previous-generation devices based on the U-Blox ZOE M8G module typically deliver four-hourly fixes with just 12 months of battery life.
LR1110 tracker modules can be compact and lightweight enough to fit small mammals and birds, while their rapid position-fixing time allows them to cope with behaviours that provide intermittent access to GPS signals. Creatures that emerge from dense vegetation or surface at sea only occasionally are examples of such behaviour.
Smart Parks’ de Groot is positive about the role of IoT technology in wildlife conservation: « With the rapidly evolving industry around the Internet of Things, we are now also capable of providing parks with the capability to detect gunshots using acoustic sensors and recognize intruders with smart cameras that work with machine learning. It’s only a matter of time before we can measure animal health status remotely as well. »
On AI and machine learning, de Groot says: « As we collect more and more data, machine learning allows us to discover more behavioral patterns, providing more situational awareness that leads to actionable intelligence — where do I deploy my rangers best? It also allows for more technology to be deployed and work more efficiently. Smart cameras that can analyze on the device itself will significantly reduce the time between a poacher accessing a park and the discovery of the intruder. »
Further down the line, implants may replace tracking collars, says de Groot: « A lot of the wildlife park owners would really like this as the current collars are not very pleasant on the eye and can be inconvenient for the animals. It’s our mission to make this happen one day, but it comes with caveats at the moment. »
Lire la suite: www.zdnet.com