Deciphering the connection between upstream obstacles, wake structures, and root signals in seal whisker array...

Area of Science:

• Bio-inspired sensing
• Fluid dynamics
• Machine learning

Background:

• Seal whiskers are crucial for navigating complex underwater environments.
• Understanding the mechanisms of whisker-based sensing is vital for bio-inspired robotics.
• Current methods lack the ability to fully integrate fluid dynamics with sensory signal analysis.

Purpose of the Study:

• To develop a novel computational framework for analyzing seal whisker sensing.
• To elucidate the fundamental mechanisms linking fluid flow, signal patterns, and obstacle detection.
• To enhance the prediction accuracy of whisker-based sensing models.

Main Methods:

• Combined computational fluid-structure interaction (FSI) modeling with interpretable deep learning.
• Developed a model to correlate signal patterns with flow characteristics and obstacle attributes.
• Generated temporal and spatial importance values to identify significant signal patterns.

Main Results:

• Accurately predicted the location and orientation of upstream obstacles (circular plate).
• Identified significant temporal-spatial signal patterns critical for accurate predictions.
• Correlated identified signal patterns with specific flow structures relevant to sensing.

Conclusions:

• The novel method effectively deciphers seal whisker sensing mechanisms.
• Insights gained can inspire advancements in underwater sensing and robotic systems.
• Demonstrated the power of integrating fluid dynamics and deep learning for biological sensing research.