A Multispecies Data Logger for Wildlife Research and Conservation

Animals communicate with signals that range from simple visual cues to complex vocalizations and gestures. Intraspecies and interspecies communication has been studied extensively in birds, dolphins, dogs, primates, and other animal species.

The field is beleaguered with challenges. Data is hard to obtain, and animal communication signals are often difficult to interpret objectively. However, with advances in machine learning and artificial intelligence, we can begin to make sense of the chirps, calls, whines, and other signals animals communicate.

The Interspecies Internet is an ongoing effort to create a foundation for communication and interaction between humans and animals. The internet connects people across countries, states, and communities. Could it be used to connect other sentient beings?

Studying human-animal communication requires a substantial amount of data. The Data Logger Project by Interspecies Internet is an initiative to unite individual data collection efforts in field science and animal conservation. The goal is to develop a modular, integrated architecture for logging and analyzing sensor data from different species and environments for field research, wildlife monitoring, and conflict mitigation.

According to Interspecies Internet co-founder, Neil Gershenfeld, the need for a multi-species data-logging platform first came up during the Interspecies Internet Workshop in 2024. The workshop’s proceedings showed significant effort duplication in the animal translation space, especially for data logging.

Gershenfeld proposed that the ideal data logger would be open source and collaboratively developed, modular with multimodal “sensory” interfaces, accessible to non-technical users, and integrated with a scalable data platform, to be suitable for widespread use.

In a roundtable discussion, leading figures and researchers in open hardware design, wildlife communication, and wildlife conservation showcased their work to identify possible points of collaboration.

Wolves use a wide range of signals to coordinate hunting, maintain social bonds, and defend territory. Jeff Reed, founder of the Cry Wolf Project at Yellowstone National Park, said the three-year research effort had acquired 200,000 hours of wolf call types, barfs, farts, and snores via battery-operated autonomous recording units (ARUs) deployed around the national park.

Reed and his team created a multi-sensory recording unit (MRU), Grizcam, that uses 360° video, multi-directional audio, and edge compute, for surveillance, security, and wildlife monitoring. Reed says they are working to add upgrades like real-time gunshot detection and sound localization.

Patrick Chwalek’s CollarID is a lightweight tracking device that collects environmental data (acoustics, temperature, humidity, and gas) and animal movement data (via GPS and accelerometer) at the same time. By combining both classes of data, researchers can gain valuable insights into endangered populations and answer fundamental questions about their interactions. The CollarID is still in the prototype stage and is being tested on different animals in Botswana, Chile, Kenya, and the United States.

Catherine Hobaiter is a professor at the University of St Andrews and heads the Wild Minds research lab, studying how wild animals, especially apes, behave and communicate in their natural environment. Data collection, although crucial to understanding animal communication, is labor-intensive and time-consuming. So, Cat and her lab created DeepWild for behavior tracking in wild chimpanzees and bonobos.

Like Hobaiter, Joris Komen, founder of humaneLabs, is working on behavior analysis for wild animals, specifically focusing on elephants. His lab explores differences in wildlife behavior in urban areas and the wild to predict and mitigate human-elephant conflict. They produced a 39-keypoint pose-estimation tool, Pachyset, and a data visualization/analytics platform for behavioral scientists, AniVis.

FluentPet is a soundboard that enables pet owners to communicate with their dogs and cats through button presses, creating the world’s largest interspecies communication dataset. FluentPet is affiliated with the Massive Interspecies Communication Observatory (MICO) and has developed an extensible JSON format for documenting interspecies interactions.

Seeed Studio CEO Eric Pan spoke about his company's readiness to assist with the design, development, engineering, and commercialization of the prospective multi-species data logging device.

The roundtable discussion highlighted the need for collaboration in interspecies communication. There are concurring projects to advance the field, but development is done in silos, making the projects hard to scale.

Hackster.io is the world’s largest hardware developer network and the leading platform for hardware collaboration. We’d love to see more interspecies communication projects hosted on the Hackster platform. The ongoing Edge AI Earth Guardians contest would be a great place to spotlight some of these innovations.