An intelligent navigation system for an unmanned surface vehicle

Abstract

A multi-disciplinary research project has been carried out at the University of Plymouth to design and develop an Unmanned Surface Vehicle (USV) named ýpringer. The work presented herein relates to formulation of a robust, reliable, accurate and adaptable navigation system to enable opringei to undertake various environmental monitoring tasks. Synergistically, sensor mathematical modelling, fuzzy logic, Multi-Sensor Data Fusion (MSDF), Multi-Model Adaptive Estimation (MMAE), fault adaptive data acquisition and an user interface system are combined to enhance the robustness and fault tolerance of the onboard navigation system. This thesis not only provides a holistic framework but also a concourse of computational techniques in the design of a fault tolerant navigation system. One of the principle novelties of this research is the use of various fuzzy logic based MSDF algorithms to provide an adaptive heading angle under various fault situations for Springer. This algorithm adapts the process noise covariance matrix ( Q) and measurement noise covariance matrix (R) in order to address one of the disadvantages of Kalman filtering. This algorithm has been implemented in Spi-inger in real time and results demonstrate excellent robustness qualities. In addition to the fuzzy logic based MSDF, a unique MMAE algorithm has been proposed in order to provide an alternative approach to enhance the fault tolerance of the heading angles for Springer. To the author's knowledge, the work presented in this thesis suggests a novel way forward in the development of autonomous navigation system design and, therefore, it is considered that the work constitutes a contribution to knowledge in this area of study. Also, there are a number of ways in which the work presented in this thesis can be extended to many other challenging domains.