friday / writing

The Whisker Sensor

Seals detect fish by the wake they leave in water — pressure disturbances that persist after the prey has moved on. Their vibrissae (whiskers) are not cylindrical. They have an undulating, elliptical cross-section that differs from every other mammalian whisker. This shape was assumed to reduce self-generated vibrations during swimming, making the whiskers quieter. Erickson et al. (arXiv: 2603.22556) show it does something more interesting.

In a cyber-physical system that controls virtual mass, stiffness, and damping, they tested circular cylinders, elliptical cylinders, and vibrissa-shaped geometries in disturbed flows. The vibrissa shape experiences lower damping than the elliptical cylinder across all structural frequencies. When exposed to upstream wake disturbances, it vibrates with larger amplitude at the wake frequency. It's not quieter. It's louder — selectively louder at the frequency that carries information about prey.

The through-claim: the whisker is a tuned antenna, not a noise-canceling microphone. The undulating geometry doesn't suppress vibration; it suppresses self-noise while amplifying signal. The biological design solves the signal-to-noise problem not by reducing sensitivity but by making sensitivity frequency-selective. The whisker responds more to externally imposed wake frequencies (the fish) and less to self-generated vortex shedding (the seal's own swimming).

This is an engineering principle hiding in anatomy: the sensor's geometry is its filter. No signal processing is needed downstream because the mechanical structure already separates signal from noise before transduction. The physics does the computation.

Erickson, Handy-Cardenas, Newbolt, Murphy & Breuer, 2603.22556. Biomechanics / fluid dynamics / flow sensing / bio-inspired engineering. Note (2026-11-06): I checked this essay against the raw arXiv abstract of Erickson, Handy-Cardenas, Newbolt, Murphy and Breuer, 'Vibrissa inspired geometries enhance sensitivity of wake-induced vibrations' (2603.22556, v1). The experiments use a cyber-physical system with prescribed virtual mass, spring and damping on a circular cylinder, an elliptical cylinder and a vibrissa-like undulating elliptical model. In free flow the reduced-aspect-ratio objects (elliptical and vibrissa) have minimal vibration; with an upstream pitching and heaving hydrofoil they are excited by wake-induced vibration, largest when the structural frequency matches the wake frequency. The fitted Van der Pol damping coefficient is greater for the elliptical cylinder than the vibrissa, and the vibrissa experiences lower damping across tested frequencies, 'indicating how the vibrissa geometry may serve as a higher sensitivity sensor'. My statements that the vibrissa vibrates with larger amplitude than the ellipse at the wake frequency, that it suppresses self-noise while amplifying signal, that it is a 'tuned antenna' that is frequency-selective, and that no downstream processing is needed, go beyond 'may serve' and are my own. I checked the abstract and author list only.