MIT CSAIL just revealed footage of SoFi, the lab’s robotic fish, which looks right at home swimming amongst the coral reefs of Fiji. The project is an attempt to create an autonomous underwater vehicle that looks as close to a real fish as possible, in hopes of studying marine life without disturbing them in the process.
The system is built around a soft robotic muscle, designed to operate similarly to a real-life fish tail. “We developed a system that takes silicone elastomer and placed hollow cavities in such a way that can equally distribute pressure on the skin of the body,” the study’s lead author Robert Katzschmann told TechCrunch. “We have two balloon chambers and flow water back and forth. That change in pressure causes the tail to undulate back and forth.”
It’s a principle that works similarly to existing soft robotics, many of which utilize shifting pneumatics to create motion in their joints. Here, it allows for the fish to be in constant motion, emitting less sound as it travels through the water.
The team did, however, use sound in other ways. A diver, equipped with a waterproofed Super Nintendo employed a custom acoustic system to help guide SoFi from afar.
“One challenge is that radio signals are absorbed really quickly in water, so something like WiFi or Bluetooth would only work within a few feet,” explained grad student, Joseph DelPreto. “Sound travels really well underwater, so we used that instead. The remote control sends out sounds that are too high-pitched for humans to hear, but the robot can decode them. Using this, we can send high-level commands to the robot.”
For now, the system is a cool video, but the team hopes access provided by Sofi’s on-board camera and fisheye lens could ultimately give marine biologists unprecedented access to their subjects.
“The fish could potentially do extraordinary things for our understanding of whales,” expand CSAIL head Daniel Rus, adding that whale births have been an extremely difficult phenomenon to capture on video. “Imagine using our fish as a non-threatening observer that is able to capture images and scenes that have never been seen before. We can learn so much about marine life.”[gallery ids="1610233,1610232,1610234,1610235,1610236,1610237,1610238"]