Improved UUV positioning using acoustic communications and a potential for real-time networking and collaboration

Abstract

Due to the lack of underwater navigation aids, a system capable of helping a submersible vehicle maintain position accuracy, permitting it to stay underwater for longer periods, is extremely desirable. The accuracy and reliability of the state-of-the-art techniques are compromised in multipath environments. In this dissertation, environments where multipath represents a problem are explored. A more accurate estimate for underwater distances by using acoustic modems and an acoustic ray tracing code to model the environment is proposed for real-time applications. To be able to establish the submersible vehicle's position, a tracking algorithm relying on Kalman filtering (KF) techniques was developed to fuse all available data. The lack of directional information from the acoustic modems employed produced nonlinearities that were treated using the extended Kalman filter (EKF) and the unscented Kalman filter (UKF). The developed algorithms were initially tested using synthetic data, and the results showed the importance of a smoothing algorithm to produce realistic trajectories. This analysis also suggested that faster convergence and better behavior in the presence of noise was achieved by the UKF approach. Real data collected during sea trials confirmed the robustness of the UKF which, despite few and inconsistent measurements, was able to provide reliable submersible vehicle positioning. Better tracking results were also achieved when underwater distances were estimated by modelling the environment.