The applications for robotic technology get inspiration from a number of sources. The latest seems like the crossing of a submarine with the Terminator.
About 25 years ago, the waters off of Florida’s Atlantic Coast were introduced to an invasive species called the lionfish. Native to the Indo-Pacific region, the fish were probably the result of aquarium owners dumping their tanks. The problem is that lionfish are voracious predators and breeders.
Equipped with venomous dorsal fins, they not only devour huge juvenile populations of commercially vital grouper, snapper, and shrimp, but females can lay as many as 30,000 eggs every four days. This potent combination has allowed the lionfish to infest waters from Florida to the Bahamas and negatively impact fishing all along the way.
In response to the lionfish problem, students at Worcester Polytechnic Institute in Massachusetts have designed an autonomous underwater robot for hunting the fish. The unit has two special systems. The first is an artificial intelligence-driven computer vision platform that uses a neural software network tied to a database with thousands of images of lionfish other marine inhabitants, like humans.
This platform allows it to know what to shoot.
Once a lionfish is detected, a revolving spear holder that works similar to a revolver firearm rotates the next spear into firing position. An elastic band is released to propel the metal spear, which floats. So once it connects with the target, the spears are buoyant enough to float the tip and the lionfish to the surface for collection. This last part is extremely important. Because although it’s a tremendous pest, Lionfish are also considered extremely tasty, with market price fetching upwards of $20/lb. The combined autonomous nature of the robot and floating tips allow for accessing deeper areas than spearfisherman can access, which are also the more popular hiding spots for lionfish.
Initial testing has only taken place in a pool, but ocean experiments are projected for 2019.