Asynchronous Data Fusion for AUV Navigation Using Extended Kalman Filtering

摘要

A truly Autonomous Vehicle must be able to determine its global position in theabsence of external transmitting devices. This requires the optimal integration of all available organic vehicle attitude and velocity sensors. This thesis investigates the extended Kalman filtering method to merge asynchronous heading, heading rate, velocity, and DGPS information to produce a single state vector. Different complexities of Kalman filters, with biases and currents, are investigated with data from Florida Atlantic's Ocean Explorer II surface run. This thesis used a simulated loss of DGPS data to represent the vehicle's submergence. All levels of complexity of the Kalman filters are shown to be much more accurate then the basic dead reckoning solution commonly used aboard autonomous underwater vehicles.