A High-tech Early Warning System for Biodiversity Research

Environmental changes are now reaching historic levels in intensity and speed: drastic declines in biodiversity and climate change are increasingly destabilizing our ecosystems with unforeseeable consequences for humans and animals. More and more animals are being displaced from their natural habitats in search of food and are invading human living spaces.

To counter these developments and protect biodiversity, we have joined forces with the Leibniz Institute for Zoo and Wildlife Research (IZW) in the GAIA Initiative. Together with our project partners, Vectronic Aerospace GmbH and Rapid Cubes GmbH, we developed a high-tech early warning system that detects ecological changes and critical events in the environment.

By using state-of-the-art camera tags with sensor-based artificial intelligence and satellite-based IoT communication, the behavior of wild animals can be precisely analyzed and comprehensively documented. This allows suitable measures to be taken at an early stage to promote biodiversity and prevent negative developments in the long term.

Innovative animal tags: minimum size, maximum efficiency

Small and energy self-sufficient

The tags must neither hinder nor disturb the animals. The use of animal tags extends over long periods of time and long distances, which imposes restrictions in terms of weight, size and energy budget.

To process data on local hardware, the computing power must be adequately dimensioned. However, this is challenging for mobile tags on small animals. Vultures can only carry a small part of their own body weight, which limits the size and weight of the hardware. The execution of complex AI algorithms on animal tags is therefore only possible to a limited extent, as the required hardware would be too heavy.

Clever data
aggregation

Sensors enable the measurement of data on animals and their environment, from which scientists can derive important information about the condition of ecosystems. However, the challenge lies in the enormous amount of data, which often exceeds the transmission capacity.

By locally pre-processing sensor data, the amount of data can be reduced before transmission, thereby decreasing the required data rate. At the same time, training neural networks requires large amounts of data and high computing power. Since the needed datasets for wildlife research are often insufficient and not reproducible, the computation-intensive algorithms pose a challenge for execution on resource-limited, portable embedded systems.

Minimal
latency

Animal swarms mainly live in remote areas that are beyond the reach of terrestrial communication infrastructure.

In poorly connected regions, large amounts of data can often only be transmitted within days or even weeks, resulting in high latency. The delay in data transmission can also significantly hinder the evaluation of the data. This can be extremely problematic, especially in critical situations where rapid action is required. Animal welfare activists and researchers are often dependent on up-to-date data to make informed decisions. Slow data transmission can therefore impair the ability to react immediately.

IoT meets Artificial Intelligence

Sensor-based data
evaluation

The data collected with animal transmitters is difficult to interpret and cannot be sent in its entirety due to the energy-intensive transmission.

That's why we rely on sensor-based AI, which reduces the amount of data to essential information. A decision can be made in real time as to which data is sent to the satellite, thereby optimizing the utilization of the radio channels. The image recognition AI integrated in the transmitter provides additional support by recognizing relevant animal behavior patterns and only triggering the camera sensor when necessary.

Data processing with Extreme Edge Computing

Due to the limited computing power of the portable animal transmitters, we use the concept of extreme edge computing.

This means that complex algorithms are not only processed centrally, but also distributed across several animal transmitters that work together in a swarm. By setting up ad hoc networks, we can efficiently distribute the computing load and thus increase the performance of the data analysis.

Satellite-based data
transmission

To achieve nationwide IoT connectivity, it is essential to support satellite-based networks in remote areas.

With our communication system, which is based on mioty® technology, we enable direct, bidirectional data transmission between animal transmitters and satellites. This not only ensures that the collected data is transmitted securely and energy-efficiently, but also enables real-time control of the transmitter nodes.

A synergy of innovative projects

The GAIA Initiative currently consists of the two sub-projects GAIA-Sat-IoT and SyNaKI. They were funded by the Federal Ministry for Economic Affairs and Climate Protection based on a decision by the German Bundestag.

 

GAIA-Sat-IoT

In the “GAIA-Sat-IoT” project, intelligent camera tags with integrated AI and satellite-based IoT communication were developed to collect ecological data and for the early detection of wildlife diseases.

Duration: 01.01.2022 to 31.12.2024

Funding reference number: 50YB2201B

 

SyNaKI

In the “SyNaKI” project, a virtual swarm intelligence was developed that uses the interaction of animals and microprocessors in digital networks to achieve more efficient data analysis through distributed artificial intelligence. The aim was to improve the data transmission and analysis of animal transmitters, inspired by the swarm intelligence of vultures and their interaction with other predators.

Duration: 01.01.2022 to 30.06.2024

Funding reference: 50YB2202B

You may also be interested in

GAIA Initiative

Artificial Intelligence and satellite communication for wildlife research and conservation

 

Satellite communication

Go back to the business area “Satellite communication“