This paper focuses on the attitude tracking control of the autonomous underwater vehicle (AUV) using control moment gyros (CMGs) with the lumped nonlinearities including model uncertainty, coupling dynamic property, external disturbance, and input saturation. To describe the attitude of the AUV without singularities, quaternions are used to describe its translational and rotational motion. A finite-time convergent extended state observer (FTCESO) in conjunction with the sliding mode control (SMC) approach is exploited to design the tracking controller for the closed-loop system with the finite-time convergence. Meanwhile, in order to release the burden of the observer, the anti-windup compensator is utilized to handle the nonlinearity of input saturation. A switch function is considered to make a switch between the robust controller and the constant-rate reaching law. Subsequently, with consideration of the inherent singularity problem of actuator dynamics, the constrained steering logic is implemented to avoid this issue without introducing other torque errors in theory. Finite-time stability of the attitude tracking system is guaranteed by the Lyapunov-based approach. Finally, the simulation results validate the attitude tracking performance of the CMG-based AUV with the proposed control strategy, when it is subject to the stated nonlinear uncertainties. INDEX TERMS Autonomous underwater vehicle, control moment gyros, attitude tracking control, extended state observer, input saturation.

Abstract-A permanent magnet synchronous motor (PMSM) drive control system for drive machine tools includes nonlinearities, uncertainties and external perturbations that should be considered in the design of control laws. A quantification factor self-tuning fuzzy PID controller for drive system is suggested by analysis the deficiency and characteristics of conventional PID controller and the limitations of conventional fuzzy PID controller. The simulation results show that the controller can map the error to the fuzzy domain in the whole speed reference range, and slow the system down when the error is small. There is almost no overshoot in both haste and slow process. The proposed controller is superior to the conventional fuzzy PID controller in dynamic stability performance and speed tracking power, and the quantification factor self-tuning fuzzy PID controller has strong robustness to external disturbance.