AI & ML
An Autonomous Sea Turtle Robot that Follows Marine Animals
arXiv cs.RO0 views
arXiv:2602.21389v2 Announce Type: replace Abstract: Autonomous robots can transform how we observe marine ecosystems, enabling vision-based focal follows (animal tracking) at scale, but close-range operation in reefs and other cluttered habitats remains difficult. Vehicles must maneuver safely near animals and fragile structures while coping with currents, variable lighting and limited sensing. Previous approaches simplify these problems by leveraging soft materials and bioinspired swimming designs, but such platforms remain limited in terms of deployable autonomy. Here we present an approach to vision-based autonomy that advances the state of the art in biomimetic marine robots and we demonstrate our approach on a sea turtle-inspired underwater robot. The robot combines robust depth-heading stabilization with obstacle avoidance and target-centric control, enabling it to track and interact with moving objects in complex terrain. We validate the approach in controlled pool experiments and in an artificial coral reef at the New England Aquarium, demonstrating stable operation and reliable tracking of fast-moving marine animals. To the best of our knowledge, this is the first integrated biomimetic robotic system, combining novel hardware, control, and field experiments, deployed to autonomously follow real marine animals. With all perception and control running onboard, the robot can hold a commanded depth, navigate around obstacles (91% success rate in the aquarium exhibit with complex lighting and visual features including glass panels) and successfully identify and follow sea turtles fully autonomously. An additional operator-initialized tracking mode extends beyond just sea turtles, including fast-moving rays, barracuda, and human divers. Together, these results establish a practical route toward bioinspired underwater robots capable of autonomous exploration and close-range animal monitoring in sensitive ecosystems.
Source:https://arxiv.org/abs/2602.21389PreviousSimVLA: Attributing Gains in VLA Models Through Controlled AblationNextLearning Object-Centric Spatial Reasoning for Sequential Manipulation in Cluttered Environments
Related posts

Tối ưu hóa kiến trúc LLM với Perplexity Orchestrator API: Multi-Cloud Failover và Drop-in Integration
0dev.to
From Checks to Decisions
0dev.to
Test your Android app with Claude Code on a cloud phone
0dev.to
If you can't explain something in simple terms, you don't understand it (2017)
0Hacker News
The AI free tier is shrinking: five tools that changed pricing in 2026
0dev.to

Hyphen vs En Dash vs Em Dash: Differences, Examples, and When to Use Each
0dev.to