Considerations of real-time control problems for an Autonomous Underwater Vehicle (AUV) are addressed in this research. Among these problems is the ability to control the submersible given its highly nonlinear operation environment. In order to account for these variations, robust control techniques must be used. In particular, Variable Structure Control (VSC) with Doyle-Stein Observer has proven to produce optimal results while maintaining a high degree of robustness. This led to the development of a real-time error detector using the robust observer to provide system redundancy through software. The culmination of this work is a real-time autopilot written in the "C" language which is ready for implementation and testing in the Naval Postgraduate School AUV prototype. We also address the aspect of real-time signal processing and conditioning in terms of Synchro-toResolver Conversion and anti-aliasing filters. The synchro problem involves converting a nonpotentiometric directional gyro output to a natural binary format which calls for an intricate design of power transformers, analog-to-digital converter, and passive element components. Lastly, the use of Generalized-Immittance Converter circuitry in the design of very low frequency anti-aliasing filter applications is developed and tested.

Approved for public release; distribution is unlimited.

http://archive.org/details/amicrocomputerba1094530711

Lieutenant, United States Navy