Perceiving like Pisces:
Lessons from Fish Sensing and Schooling

 

The lateral line found in most species of fish is a sensory organ without analog in humans. Acting in a similar fashion to an array of pressure sensors, the lateral line is used by fish for behaviors including tracking prey, detecting predators, avoiding obstacles, and schooling. While much progress has been made in marine vehicle design, sensory capabilities remain limited and expensive when operating in dark and turbid environments.

 


Figure 1. The lateral line is composed of two subsystems: superficial neuromasts and the trunk canal. The distribution of the lateral line on a Lake Michigan Mottled sculpin is shown above middle. The black dots denote the distribution of superficial neuromasts, while the trunk canal subsystem is outlined in red. Located on the skin of the fish, superficial neuromasts are exposed to the mean flow and act primarily as velocity sensors. Canal neuromasts are located within canals beneath the skin, and are connected to the external flow via a series of pores. Pressure gradients between pores induce localized flows which stimulate the neuromasts.

 

Smart Skin Pressure Sensor Arrays

In an effort to replicate the behaviors observed in fish that are enabled by the lateral line organ, work is ongoing to develop pressure sensor arrays fabricated from a carbon black-PDMS (CBPDMS) composite for use on marine vehicles.

Requirements for Hydrodynamic Sensing:
Flexibility: Pressure sensor arrays are intended to be surface mounted on curved surfaces typical of the hulls of marine vehicles.
Form Factor: Sensor pitch must be on the order of millimeters to properly resolve hydrodynamic stimuli.
Robustness: Sensor arrays are meant for sustained underwater operation.
Sensitivity: Sensors must be capable of detecting stimuli on the order of tens of pascals and at frequencies below ten Hertz.
Cost: Sensor arrays must be low-cost to allow for a high density of sensors.

 


Figure 2. Top Left: Version 1 conformal pressure sensor array utilizing solid CBPDMS strain gauges and strain enhancing diaphragms. The Version 1 sensor has been successfully field tested on an ASV. Top Right: Porous CBPDMS pressure sensor that has been encapsulated using insulating spray. The array utilizes a four-point probe arrangement to reduce contact resistance. Bottom Left: Household plastic wrap was found to be an excellent encapsulation method for the porous CBPDMS due to minimal impact on the Young’s modulus. Bottom Right: Completed two-dimensional array utilizing solid CBPDMS and copper foil traces. The entire array is encapsulated in pure PDMS.

 


A dynamic calibration of the porous CBPDMS sensor was performed by plotting the sensor voltage output versus water wave dynamic pressure. The sensor voltage was from a single sensor location on an array encapsulated using simple kitchen plastic wrap, which was found to have minimal impact on the Young’s modulus of the porous material. The mean of the voltage has been removed to consider only the dynamic response of the sensor. Using a linear best fit, a sensitivity of 11.7 mV/kPa was achieved.

Object Detection and Identification

In parallel to the development of pressure sensor arrays, work is done to understand the signals that such sensors would measure and how they can be used for underwater navigation.


Figure 4: Snapshots of changes in velocity (u-u0) and pressure coefficient (Cp-Cp0) as a NACA0012 foil passes near a cylinder at Reynolds number Re = 6 250. The displacement thickness along the airfoil is shown by a solid grey line. (b): Magnified view of the area enclosed within the dashed line of (a), showing the disturbance amplified by the boundary layer in the form a vortices.

For instance, we have shown that as a foil passes near a cylinder, it causes a disturbance that can be used by pressure sensors embedded in the foil to locate and identify the cylinder. This disturbance typically gets amplified by the boundary layer, making it even easier to identify the cylinder. We developed a methodology to calculate the unsteady pressure based on combining potential flow predictions with results from linear stability analysis of the boundary layer.