Aiming at the control problem of nonlinear uncertain systems with asymmetric saturated actuators and u nknown external disturbances, a composite control method integrating dynamic surface control (DSC), ad aptive neural network estimation, and a nonlinear saturation compensation mechanism is proposed. In the scenarios of ship course and trajectory tracking, the system faces multiple challenges such as symmetric and asymmetric actuator saturation, as well as unknown external disturbances. Radial basis function (R BF) neural networks are utilized for online approximation of unknown nonlinear functions and external d isturbances. Combined with dynamic surface technology, the problem of "explosion of complexity" in tra ditional backstepping control is eliminated. A nonlinear function with inverse correlation to error gain is designed to dynamically adjust the control gain, balancing the requirements of tracking accuracy and sat uration suppression. Furthermore, a Gaussian error function is introduced to construct a continuously diff erentiable asymmetric saturation model. An auxiliary dynamic system is integrated to compensate for the saturation nonlinear effect, achieving smooth amplitude limitation of rudder angle commands. Comparati ve MATLAB simulation results demonstrate that the course tracking error is reduced by 1°, the fluctuati on amplitude of the rudder angle is decreased by approximately 50%, the number of rudder angle satura tion events is reduced by about 60%, and the error convergence time is shortened by roughly 30%. The proposed composite control method effectively addresses the issues of asymmetric saturation and externa l disturbances, significantly enhancing the accuracy and robustness of the ship course control system.

Abstract
1. Introduction
2. Problem Analysis
    2.1. Gaussian error function smoothing of rudderangle saturation
    2.2. Neural networks approximate unknownnonlinear functions
3. Controller Design
    3.1. Definition of tracking error
    3.2. Design a virtual control law
4. Stability Analysis
5. Simulation experiment
6. Conclusion
References

• ZHAO Zhijun(School of Navigation and Shipping, Shandong Jiaotong University, China)
• DING Shengda(School of Navigation and Shipping, Shandong Jiaotong University, China)
• LI Xingyi(School of Navigation and Shipping, Shandong Jiaotong University, China)
• HU Yancai(School of Navigation and Shipping, Shandong Jiaotong University, China) Corresponding author