Hybrid adaptive control of autonomous underwater vehicle

Abstract

Hybrid adaptive control of autonomous underwater vehicle (AUV) is investigated. Dynamics of AUV vary by change in operating conditions and even theoretically or experimentally driven dynamical coefficients reflect an approximate to the exact ones. Adaptive control technique is employed to handle the uncertainty problems in the system dynamics. In the applied hybrid adaptive control, the system is simulated in a continuous domain while the control and identification sections are discrete. The discrete model and position of zeros of sampled data unstable system are addressed. Convergence rate of parameter estimation is crucial in the stability of closed loop system particularly when open loop unstable system passes its initial states or is entangled by radical changes in the dynamics. Adaptive normalization is suggested which improves the rate of convergence and conserves stability. The results of modified direct, indirect and linear quadratic Gaussian (LQG) adaptive control are presented.