Virtual meetup highlights networked sensor technology for parks

Technology users and tech developers don’t always talk directly to each other, but they should.

The relatively small size of the scientific/conservation community and its need for a particular suite of features, such as ruggedness, waterproofing, long battery life, small size, ease of use, and sometimes even camouflage — at a low cost — can impede companies’ development of suitable systems. This, in turn, restricts the use of technology in the field.

To encourage communication between users and developers of technology for conservation, the WILDLABS platform began a series of virtual meetups earlier this month. The webinars feature engineers and other tech developers presenting some aspect of their conservation-related work, and listeners are encouraged to offer questions and comments. Anyone with internet access can join the meetups or watch recordings of them on Zoom that include notes on key points from the discussion.

The first meetup on November 8, 2018 introduced three engineers who described their projects using networked sensors for improving security and reducing human-wildlife conflict. The speakers discussed the project aims, the technologies they’ve developed and deployed, and the challenges of working in remote conservation areas.

Engineer Eric Becker spoke about World Wildlife Fund’s Wildlife Crime Technology Project (WCTP) and some general lessons he has learned in deploying technology in remote areas.

Through their testing, the WCTP team quickly recognized that drones, their initial focus, were not the solution to detecting poachers at night, “mainly due to the cost of the platforms, the [limited] flight time, and the size of the areas we need to cover,” Becker said.

They shifted their focus to rangers’ need for long-range night vision, deploying a series of networked thermal cameras, which scan 360 degrees and “see” heat emitted by people and animals, along reserve boundaries.

Becker said that an anti-poaching team would ideally want to respond to an alert before their target leaves the sensor’s field of view. Some of their video cameras can distinguish humans using artificial intelligence, so WCTP partners have revisited the use of drones to quickly catch up with and follow an intruder until rangers can approach.

Becker recommended that engineers interested in designing for wildlife conservation create systems that are wildlife- and weather-proof, have the backing of rangers and park management, and can be used and maintained by people “who may have never used a computer in their life.” He also suggested bringing tools, spare parts, and power sources to any deployment.

Laurens de Groot, co-founder of the non-profit SmartParks, agreed with Becker that drones were “a super-nice tool” but had limitations for certain wildlife protection-related tasks. “Maintenance is horrible,” de Groot said, “the terrain where you’re working is horrible [and sometimes too windy to fly], especially for fixed-wing drones, and they are expensive to run, so we thought this is not going to be a sustainable solution for conservation.”

He co-founded SmartParks to improve park management by gathering information from sensors set out to monitor the locations of cars, rangers, and animals, as well check the functioning of electric fences, to separate elephants and people, and other tasks.

Before deploying technology at a park, his group conducts a feasibility study to prioritize what and where sensors are most needed and then builds the network to meet those needs. With poor internet connections at many parks, the group has created closed long-range radio (LoRa) networks in reserves in Rwanda, Malawi, and Tanzania, and India. The group partners with African Parks, which has the infrastructure de Groot said was needed to implement a networked park and respond rapidly to electronic alerts.

de Groot said he and his team want to build a sustainable business, rather than rely on grants; he estimated a cost of 150,000 to 300,000 euros to build a smart park, depending mainly on the number of sensors and the area’s terrain.

Jan-Kees Schakel, founder and CEO of the non-profit Sensing Clues, works in a similar way though he said his team could “start small,” spending $5,000 or $10,000 with smaller parks just beginning to deploy connected field technology and up to $100,000 in areas with more infrastructure. Schakel’s team works with companies producing “best-of-breed” technologies to adapt their sensors and make them available for conservation needs. The team also connects these companies with conservation-related law enforcers and wildlife researchers to help create effective sensing networks.

The group has developed three innovative sensors of its own, each designed to distinguish a clear sign of human activity in an area. The Trespasser detects radio signals from smartphones; SERVAL detects sounds, such as motors, gunshots, shouting, chopping wood, or cattle, that are not expected in a protected area and classifies them using machine learning; and a third sensors detects light.

Wildtech will cover the second meetup on Next-Generation Wildlife Tracking, which took place on November 20^th, in a separate article.  The third meetup, on big data in conservation, is planned for December 12 from 3-4pm GMT / 10-11am EST. You can register for it here.

FEEDBACK: Use this form to send a message to the editor of this post. If you want to post a public comment, you can do that at the bottom of the page.