For net integrity inspection operation using Remotely Operated Vehicles (ROVs), it is very important to be able to accurately decide the position and orientation of the ROV within the net-cage. This information can for instance be used to avoid collision with the net and avoid entanglement of the tether; to accurately pin-point where in the net a hole was detected by the vision system; to know what part of the net has already been inspected; and to plan the motion of the ROV in order to cover the rest of the net structure.
Our solution for net-relative localization is based on onboard sensors only (yaw gyro, yaw compass, depth sensor, doppler velocity log, camera, Global Navigation Satellite System (e.g. GPS), and inclinometer or accelerometer.
By using a kinematic model of the ROVs position and orientation relative to the net, combined with information from all these sensors, we have obtained very promising results in simulations. The next step will be to use real sensor data in the algorithm, and verify the localization performance against other solutions (e.g. a solution using external beacons for position and orientation).