Underwater Mining Robot
February 2017 – May 2017
I worked in a team to design an underwater mining robot for the class Design of Electromechanical Robotic Systems (2.017). We were required to create a robot that interacted with water, had some form of sensing, performed calculations off of those measurements, and then took action off of those calculations.
We chose to create a prototype sampling robot for deep-sea nodule mining. On certain portions of the sea floor, metallic nodules are abundant and rich in valuable metals. These nodules have roughly the size and shape of an average potato, so we used potatoes in our testing. We used a large nut gatherer in our robot which was able to pick up nodules of the typical size. Our prototype used a tether cable to supply power and communicate, but a field robot would have to be untethered.
The robot was kept neutrally buoyant by balancing flotation devices and masses, with its movement controlled by propellers. We designed our propellers to thrust downward strongly enough to force nodules into our gatherer. The propellers also had to be able to lift the robot, newly laden with nodules, up to the surface.
On the team, I was in charge of hardware and testing. I helped to design the robot body, and I validated that our propellers would be adequate for the job. I built a special testing apparatus to measure propeller force and took measurements in the Sea Grant water tank. I showed that our theoretical model was correct and that the propellers would be adequate.
Our robot was controlled by a Raspberry Pi, with multiple sensors. We had a sonar to detect the distance to the sea floor and the nodules, along with two cameras which could identify nodules from the sea floor. The Raspberry Pi would use this information to try to maintain a steady distance above the sea floor, with finer control being done manually over the tether.
This project proved to be too large in scope for the space of a single semester. Although the class finished before we were able to test the whole system together at once, we were able to show that each subsystem worked individually.
This is the final assembled robot.
This is the propeller testing apparatus I designed. The sensor needed to be kept out of the water, and the torque balance amplified the signal of interest.
Here the test fixture is shown mounted in the Sea Grant water tank.
Here's the SolidWorks assembly model of the robot, shown from the front. The cameras are in blue, the thrusters are in black, the flotation devices are orange, the nut gatherer is on the red rod, and the electronics are contained in the gray box at the back.
Here's the SolidWorks assembly model of the robot again, this time shown from the back. In this view you can see where the sonar sensor, shown in teal, is mounted.
This is the time evolution of force generated by a single propeller at 99% power. The measured steady-state value was 6.059 N with a standard deviation of 0.088 N. This agrees with the theoretical model which predicted a force of approximately 6 N.