Project CETI (Cetacean Translation Initiative) aims to collect millions to billions of high-quality, context-sensitive vocalizations to understand how sperm whales communicate. But knowing where whales will surface in order to find them and capture data can be difficult. This makes it difficult to attach a listening device and collect visual information.
Today, the Project CETI research team, led by Stephanie Gil, assistant professor of computer science at Harvard’s John A. Paulson School of Engineering and Applied Sciences (SEAS), proposes a new reinforcement learning framework with autonomous drones to find and predict sperm whales. I did it. Where they will surface.
This study Science Robotics.
This new research explores the use of a variety of drones, such as the Project CETI aerial drone, with very high frequency (VHF) signal detection that emulates an ‘airborne antenna array’ by leveraging the signal phase along with the drone’s motion to estimate the directionality of pings received from CETI. Use a sensing device. Whale tag. This shows that it is possible to use these diverse sensor data, as well as predictive models for sperm whale diving behavior, to predict when and where whales will surface. Armed with this information, Project CETI can now design algorithms for the most efficient routes for drones to encounter or encounter whales on the surface. This also opens up the potential for conservation applications to help prevent ships from colliding with whales while on the surface.
Presenting an autonomous vehicle for whale tracking and rendezvous via remote sensing or AVATARS framework, this study jointly develops two interrelated components: autonomy and sensing. And there is a detection feature that measures the Angle-of-Arrival (AOA) on whale tags to inform the decision-making process. Measurements from autonomous drones to surface tags, acoustic AOA from existing underwater sensors, and whale behavior models from previous biological studies of sperm whales are provided as input to the AVATARS autonomous decision-making algorithm. whale.
AVATARS is the first product to co-develop VHF sensing and reinforcement learning decision-making to maximize robot-whale encounters at sea. A well-known application of time-critical rendezvous is with ride-sharing apps that use real-time sensing to record the dynamic routes and locations of drivers and potential passengers. When a rider requests a ride, you can assign a driver to meet them as efficiently and timely as possible. Project CETI’s case is similar in that it is tracking whales in real time with the goal of coordinating the drone’s rendezvous to meet the whales at the surface.
This study advances Project CETI’s goal of acquiring millions to billions of high-quality, highly contextualized whale vocalizations. Adding different types of data will improve location estimation and routing algorithms, helping Project CETI achieve its goals more efficiently.
“We are excited to contribute to the groundbreaking advancements of Project CETI. By leveraging autonomous systems and advanced sensor integration, we can address key challenges needed to track and study whales in their natural habitat. This is not only a technological advancement, but also a “This is an important step in helping us understand the complex communication and behavior of ,” said Gil.
“This research is a major milestone for the Project CETI mission. We are now one step closer to significantly improving our ability to collect high-quality and large-scale data sets on whale vocalizations and associated behavioral contexts, allowing us to better hear and translate what we hear.” You can go, the sperm whale is talking,” said David Gruber, founder and director of Project CETI.
“‘This research was an incredible opportunity to test our systems and algorithms in a challenging marine environment. Combining wireless sensing, artificial intelligence and marine biology, this interdisciplinary research will demonstrate how robotics can be part of the solution for future marine environments. “This is a prime example of what can be done,” said Ninad Jadhav, a PhD candidate at Harvard University and first author of the paper.
Sushmita said, “This project will help us in the field where robotics and artificial intelligence can enrich data collection and promote research into the broader sciences in language processing and marine biology, which can ultimately help protect the health and habitat of sperm whales. “It provides a great opportunity to test algorithms,” he said. Bhattacharya is a postdoctoral researcher in Gil’s REACT lab at SEAS.
Additional information:
https://www.projectceti.org/
