Coordination of Artificial Robotic School for Localization and Surveillance

Abstract

The involvement of autonomous underwater vehicles to monitor submerged structures, mines detection, or field exploration can benefit the economy by reducing the risk and the cost of the mission leading to more frequent monitoring. Reliable navigation and localization systems are fundamental for Autonomous Underwater Vehicles (AUVs) missions. The challenges associated with the underwater environment can heavily affect the reliability of the AUV. It is certain that electromagnetic signals are very limited underwater and thus the use of the Global Positioning System (GPS) is not feasible. Solutions to overcome the underwater challenges can be achieved with expensive methods making the AUV solution unfeasible for most applications. Real-time position is mostly obtained by combining expensive sensors such as the Doppler Velocity Log (DVL) or acoustic ranging and integrating them through sensor fusion techniques. The goal of this thesis is to develop an inexpensive underwater navigation system that does not rely on DVL and takes advantage of inexpensive Inertial Measurement Unit (IMU) and exploiting data from the dynamic model. In this work, an inverse dynamic model to estimate the velocity of the AUV is established, tested with realistic underwater simulation, and compared with similar state-of-the-art approaches. Results are integrated with an Extended Kalman Filter (EKF) to estimate the state of the robot and compared to the state-of-the-art approach. Finally, a collaborative approach to improve the state estimation of each robot is proposed and validated through simulation.

Date of Award	Dec 2022
Original language	American English
Supervisor	Federico Renda (Supervisor)