Hardware Design of an Improved ROV
Pictured above: renders of some of the components designed for the ROV
|
Background
Underwater vehicles are extremely important for oceanographic research-climate change researchers need them to get measurements under disappearing ice sheets, marine biologists need them to catalog deep-sea life where humans can’t survive, and geologists need them to study hydrothermal vents on the seafloor. But navigating and controlling these vehicles from afar-and even autonomously-presents unique problems that most terrestrial robots don’t face: GPS signals can’t penetrate underwater, and cameras are ineffectual in the dark and featureless conditions. My lab, the Dynamical Systems and Controls and Lab, develops alternative systems for controlling and navigating underwater vehicles. Of course, ship time for testing these systems is prohibitively expensive, and testing these algorithms on vehicles at sea is prohibitively risky, so we do testing on analogous robotic testbed vehicles in a tank at Hopkins. My own work was focused on improving the design of our current testbed ROV, the JHU ROV. This new testbed is designed with a number of improvements:
|
Paroscientific Pressure Sensor ROS Node
Another project I took on during my time at the DSCL was to write a Robotic Operating System (ROS) node for a pressure sensor on the vehicle, part of the lab's effort to standardize all robot operations using ROS. Important for calculating depth for use in navigation and control algorithms, the sensor publishes serial messages indicating pressure readings. My ROS node thus reads these messages, parses them for relevant data, and then publishes both the raw pressure data and the depth, as calculated using Fofonoff's eponymous formula (Saunders and Fofonoff, 1976). Follow the link below to view the source code.