Decoding Propulsive Wakes using Biomimetic Seal Whiskers

Area of Science:

• Biomimetics and Sensor Technology
• Fluid Dynamics and Hydrodynamics
• Aquatic Ecology and Monitoring

Background:

• Optical and acoustic methods for underwater detection are limited.
• Passive hydrodynamic sensing, inspired by seals, offers a non-disruptive alternative.
• Biomimetic whiskers can detect hydrodynamic wakes from moving objects.

Purpose of the Study:

• To quantify the ability of a force-instrumented biomimetic whisker to discriminate four properties of a propulsive source simultaneously.
• To assess the effectiveness of a convolutional neural network in interpreting whisker responses to hydrodynamic wakes.
• To establish the hydrodynamic force signature as a basis for non-disruptive aquatic monitoring.

Main Methods:

• Towing tank experiments using a 30x scaled biomimetic seal whisker with a six-axis force/torque transducer.
• Analysis of whisker response to wakes generated by a flapping NACA 0012 airfoil under 1,050 conditions.
• Development and training of a 1D convolutional neural network on raw six-channel time-series data.

Main Results:

• The biomimetic whisker's force response accurately reflected source frequency, pitching amplitude, angle of attack, and downstream distance.
• The convolutional neural network achieved 94.1% average test accuracy in simultaneously recovering all four source attributes.
• Frequency and angle of attack were recovered with high precision, while pitching amplitude and distance showed minor classification errors.

Conclusions:

• A single force-instrumented biomimetic whisker can serve as a compact sensor for characterizing propulsive wake sources.
• This technology shows significant potential for non-disruptive detection and tracking in fisheries protection and aquatic monitoring.
• Machine learning interpretation of hydrodynamic signatures enhances the capability of biomimetic sensors for underwater applications.