AI-assisted Bio-logging

We have developed AI-enabled bio-loggers to improve runtime and precision.
Published in Ecology & Evolution
AI-assisted Bio-logging

Share this post

Choose a social network to share with, or copy the shortened URL to share elsewhere

This is a representation of how your post may appear on social media. The actual post will vary between social networks

How can we enhance biological research with AI and IoT technologies? I am an information scientist specializing in mobile and ubiquitous computing. In our project “Systems Science of Bio-Navigation,” which is supported by the Japan Society for the Promotion of Science, we have studied the “algorithms” of animal navigation in collaboration with researchers from the engineering, information science, ecology, and neuroscience research domains.

The research groups of information scientists have helped in achieving the goal of the research by providing: i) assisting technologies for collection of animal behavior data through machine learning, and ii) supporting computational technologies for animal behavioral analysis.

The bio-logger equipped with a camera developed by us (left; latest version) and animal behavior observation with bio-loggers (right)

In our study related to data collection, we focused on bio-loggers (animal-borne data loggers) and developed AI-enabled bio-loggers. Bio-logging enables us to observe several aspects of animal life. However, because the mass of a bio-logger is restricted, the bio-logger has short runtimes when collecting data from resource-intensive sensors such as cameras.

To address this problem, we proposed the concept of AI-assisted bio-loggers, which we will refer to as AI on Animals (AIoA). These bio-loggers can use low-cost sensors, such as accelerometers and GPS, to automatically detect behaviors of interest in real-time, allowing them to conditionally activate high-cost sensors, such as cameras, to target those behaviors.

An activity (behavior) detection process is run on the processor of the bio-logger in real-time. In general, we debug the software on a computer (PC) while running it. However, it is difficult to do this in our case because the computer is on a bird! To debug the software and fine-tune the parameters, we also attended fieldwork on animals (seabirds) conducted by our research collaborators.

Climbing down cliffs at Niigata, Japan (left) and Peninsula Valdes, Argentina (right) to arrive at colonies of seabirds

Typically, information scientists prefer to stay at home or in laboratories. In addition, in many cases, our research activities in the laboratories are completely safe. I never expected to climb a dangerous cliff in Argentina in my research activities when I started my carrier as an information scientist! However, after frequent visits to the fields, I became more interested in biological studies and in the potential of AI and IoT technologies for understanding animal ecology. 

Peninsula Valdes, Argentina (left) and Kabushima, Aomori, Japan (right)

Here is an example video recorded by our bio-logger:

We have also published our YouTube channel that introduces videos recorded by our loggers. Please check and register it. For more details about our loggers, please refer to our paper.

For supporting technologies of animal behavioral analysis, we developed a deep learning method for the comparative analysis of animal trajectories. The method, called DeepHL, assumes a comparative analysis of two animal groups (e.g., male and female seabirds) and automatically highlights trajectory segments that are characteristic of one group, enabling biologists to focus on important segments of various trajectories.

Screenshot of the DeepHL web interface comparing the trajectories of a female (left) and male (right) streaked shearwater. Characteristic segments of the trajectories are highlighted in red. This example indicates that female birds prefer areas close to the coastline.

This study has also recently been published in Nature Communications (link):

Takuya Maekawa, Kazuya Ohara, Yizhe Zhang, Matasaburo Fukutomi, Sakiko Matsumoto, Kentarou Matsumura, Hisashi Shidara, Shuhei J. Yamazaki, Ryusuke Fujisawa, Kaoru Ide, Naohisa Nagaya, Koji Yamazaki, Shinsuke Koike, Takahisa Miyatake, Koutarou D. Kimura, Hiroto Ogawa, Susumu Takahashi, and Ken Yoda: Deep Learning-assisted Comparative Analysis of Animal Trajectories with DeepHL, Nature Communications 11 (5316), Oct. 2020.

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in

Subscribe to the Topic

Life Sciences > Biological Sciences > Ecology

Related Collections

With collections, you can get published faster and increase your visibility.

Materials and devices for separation, sensing, and protection

In this Collection, the editors of Nature Communications and Communications Materials welcome the submission of primary research articles that highlight the development and application of functional materials in the areas of separation, sensing, and protection.

Publishing Model: Open Access

Deadline: Jun 30, 2024

Applied Sciences

This collection highlights research and commentary in applied science. The range of topics is large, spanning all scientific disciplines, with the unifying factor being the goal to turn scientific knowledge into positive benefits for society.

Publishing Model: Open Access

Deadline: Ongoing