Abstract
To improve the performance of tracking control for dynamically positioned (DP) vessels, this paper presents the design and implementation of an adaptive robust backstepping controller without using model parameters in the presence of unknown time-varying disturbances. The bias radial basis function neural networks (RBFNN) are employed to compensate for the uncertainties and disturbances of ship dynamics. The introduction of bias term of RBFNN helps to avoid the designing of additional integrator of backstepping. The output of the RBFNN can be non-zero when the inputs of the RBFNN deviate significantly from the center points because of the bias term, which improves the robustness of the adaptive RBFNN controller. The reference trajectories of this paper are generated by linear quadratic regulator (LQR) to diminish the overshoot and save the energy consumption. The system matrices used in LQR are identified through the data-driven technique. Uniform boundedness of the closed-loop system is proved by the Lyapunov stability theory. Simulation results illustrates the performance of suggested control scheme.
