Determination of object position, vortex shedding frequency and flow velocity using artificial lateral line canals

Abstract

The lateral line system of fish consists of superficial neuromasts, and neuromasts embedded in lateral line canals. Lateral line neuromasts allow fish to sense both minute water motions and pressure gradients, thereby enabling them to detect predators and prey or to recognize and discriminate stationary objects while passing them. With the aid of the lateral line, fish can also sense vortices caused by an upstream object or by undulatory swimming movements of fish. We show here that artificial lateral line canals equipped with optical flow sensors can be used to detect the water motions generated by a stationary vibrating sphere, the vortices caused by an upstream cylinder or the water (air) movements caused by a passing object. The hydrodynamic information retrieved from optical flow sensors can be used to calculate bulk flow velocity and thus the size of the cylinder that shed the vortices. Even a bilateral sensor platform equipped with only one artificial lateral line canal on each side is sufficient to determine the position of an upstream cylinder.

Keywords: artificial lateral line; biomimetics; flow sensor; mechanoreception; optical sensor.

Figure 1

Responses of sensor platform I to a human finger that moved alongside the ALLC with a velocity of about 20 cm/s. Arrows indicate direction of finger motion. Responses of AN_A (upper graph) and AN_B (lower graph) are shown. Note that the responses were inverted if the direction of finger motion was reversed.

Figure 2

A typical AN response to a dipole flow field. The sinusoidal voltage used to drive the mini-shaker (bottom left) and the voltage output of the AN exposed to the vibrating sphere (top 4 traces) is shown. Distances between the ALLC and the vibrating sphere were 1, 2, 3 and 4 cm. The voltage output of the AN is plotted on the same scale. Inset: Voltage output of the AN as a function of sphere to pore distance. The fitting function (dashed line) is f(x) = 42.66 x^−3.03. This indicates a third order decrease of sensor output with increasing sphere distance.

Figure 3

The response of an artificial CN exposed to the vortices shed from a stationary cylinder (diameter 2 cm) (A) and frequency spectrum of the response (B). Peak frequency (0.82 Hz) was similar to the calculated vortex shedding frequency (0.98 Hz). (C) Measured (black bars) and calculated (white bars) vortex shedding frequencies for cylinders of different diameters.

Figure 4

Responses of AN_A (upper line in each graph) and AN_B (lower line in each graph) (for AN_A and AN_B see Figure 6A) to the vortices caused by a cylinder (diameter 2 cm) placed at various positions upstream to the ALL. Note that response amplitudes decreased with increasing distance of the cylinder and that the responses were more pronounced for the AN that was ipsilateral to the cylinder.

Figure 5

Responses of AN₁ to AN₈ (cf. Figure 6B) to bulk water flow. Note that the most prominent peak visible in the response of AN₁ is systematically delayed from AN₁ to AN₈. The curves are not to scale due to the manual fabrication of each AN.

Figure 6

Horizontal (left) and vertical (right) cross sections of platforms III (A) and IV (B) that were used for the present study. Black squares indicate the positions of the ANs inside artificial lateral line canals. Platform III housed two and platform IV housed eight ANs. The length of pores (Lp) was 1 cm, the inter pore distance (Ls) was 1 cm and the diameter of the canal (Dc) as well as the pore diameter (Dp) was 3 mm in both devices.

Figure 7

Velocity calculated with data obtained from AN₁ to AN₈ as function of bulk flow velocity. Note that the fitting curve intersects close to the origin (b = 0.05). Its slope (m) is larger than 1. This could be due to vibrations or a crosstalk between adjacent ANs.

Figure 8

Scheme of an artificial CN. The drawing is not to scale.