Navigating in deep water is challenging due to limited access to GPS and flow maps. Research from Kanso Lab demonstrates that robots can autonomously learn to navigate unsteady flows from a robot-centric perspective, as long as they can sense local flow velocities and gradients.
Ocean monitoring is essential for understanding ecosystem functioning, marine biodiversity and the ocean’s carbon cycle, particularly in the face of our rapidly changing climate.
To expand the current capabilities of underwater robots for long-term ocean surveillance and monitoring, we need effective control strategies that enable robotic swimmers to seamlessly navigate through shifting currents.
Rather than being controlled by an external agent, on-board sensors have the potential to enable robots to swim autonomously and learn to respond to unsteady flows.
But is it feasible for robots to learn underwater navigation autonomously without any land references? Navigating underwater environments presents unique challenges because of the dynamic nature of flow currents and the absence of global positioning signals (GPS).
New research from Kanso Bioinspired Motion Lab at USC, published in Nature Communications, expands underwater robot-centric learning, helps explain why aquatic organisms have arrays of flow sensors that detect gradients, and provides physics-based guidelines for translating land-based principles to unfamiliar and diverse flow environments.
Read the paper here.
Published on April 8th, 2025
Last updated on April 8th, 2025
