Professor Christopher Clark is a new addition to the Harvey Mudd College Engineering Department. While not teaching, his research interest is underwater robotics. Clark’s robots have done everything from testing the health of the waters in Norway to exploring ancient, underwater cisterns in Malta to tracking leopard sharks in California.
In addition to shark tracking locally, Clark is involved in The International Computer Engineering Experience (ICEX) program. In ICEX, a team of students works on a system to allow a robot to accomplish its stated task. For the project in Norway, the robot had to patrol the waters and test for the amount of dissolved air. In Malta, the team lowered the robot into ancient corridors below the city that were full of water. The robot then autonomously made its way through the halls while mapping them.
Despite the variety in such projects, there is a common theme. Clark works to allow his robotic units to function together.
For instance, in tracking sharks with multiple robots, the robots must not only stay away from the shark so as not to disturb it, but also must be able to work efficiently with each other. This means that the robots must keep track of each other’s position with regard to themselves and the shark. The shark being tracked has been tagged so that the robots have a sense of its location.
The device does not provide a perfect fix. The robots have to make guesses at the location of the shark. By staying evenly spaced in a circle around the shark, the robots can collectively get the best fix on its actual location. The ability to do so is a complex example of a multi-robot system.
Before the robots can perform such tasks, Clark must find ways to get his gear where it needs to be. This can be a tricky task considering, as he points out, that “electronics don’t like water.” The robot must be able to survive in the water before it can attempt to accomplish its task—and accomplishing a task in water is much more difficult than accomplishing it on land.
“Underwater, there are things in the way,” Clark said. “You can’t see where you’re going.”
I have had moderate experience with the complexities of underwater vehicles myself, although with a much lower degree of complexity and with depressingly lower success. In high school, a friend and I made a remote-operated, underwater vehicle for an engineering class. It was just a cube of PVC tubing with four water pumps on it. We wired up all the pumps to an old Atari joystick so that it would, in theory, be able to move in two directions. We waterproofed the wiring by stringing it through a gardening hose all the way back to the control.
When we put the thing in the water, the water cooled down our hose, which had been sitting coiled up for a couple weeks. This completely prevented the vehicle from moving. When it came time for us to present on our project, we simply showed the class that when we moved the joystick, the correct pump clicked.
As Clark said, “The underwater world is difficult.”