Ever wonder how fish can find their way around so easily in murky water? Well, most of them use something called their lateral line – a row of hair cells down either side of their body that detect changes in water pressure caused by movement, or by water flowing around objects. Now, scientists from Singapore’s Nanyang Technological University and MIT have copied the lateral lines of the blind cave fish, in a man-made system designed to allow autonomous underwater vehicles (AUVs) to navigate more accurately and efficiently.
Ordinarily, AUVs use cameras, sonar, or an underwater acoustic positioning system. Cameras aren’t much use in murky water, however – and a lot of the world’s water bodies are murky. Sonar and acoustics are better in such situations, but the hardware can be expensive, and taxing on the AUV’s batteries.
By contrast, the blind cave fish-inspired sensors can reportedly be made for under US$100 per array, and use little power. Additionally, they don’t produce audible sonar “pings” that can be harmful to aquatic animals.
Instead, the system utilizes an array consisting of two rows of five sensors (yes, we know there are four rows in the picture – it's presumably two joined arrays). Each sensor measures just 1.8 x 1.8 mm and consists of a microscopic sensory pillar surrounded by hydrogel, that bends with changes in the water pressure. Combined with a computer vision system, the arrays reportedly allow Nanyang’s AUVs to create 3D images of nearby objects, and to map their surroundings.
Besides their use in AUVs it has also been suggested that the sensors could find use on military submarines, that don’t wish to announce their presence to enemies via sonar pings.
The scientists now hope to create a piezoelectric version of the system that generates all of its required operating electricity from the movement of the water. They are also investigating ways in which the sensors could detect submerged objects even when there’s little or no current.
Another later line-inspired underwater sensing system is being developed by scientists from the University of Illinois and Northwestern University.
Source: Nanyang Technological University
Maybe something we could see if we can talk to them and have them send us a sensor to test, or send them an OpenROV to test the sensor.
This is pretty cool. There was some similar work a while back by robotics researchers using electric field sensing fish as the model. Basically, the researchers built arrays of electric field sensors to do the same thing (ie. "tomography" -- determining which space is occupied). I wrote about some of it a while back:
More recently... some Italian robotics researchers presented work at ICRA / IROS (some of the top robotics conferences), using the same electric field tomography in underwater ROVs to detect obstacles in murky water. I don't know the reference off hand (buried in email)... but I could probably look it up if you're curious.
I can try emailing...
Also, Eric always talks about trying to incorporate optical flow sensing. But if this sensor is really cheap...