Active disturbance rejection-based robust course control for ships with uncertain steering dynamics

Abstract

The complex and uncertain steering dynamics of ships present significant challenges to traditional model-based control strategies, particularly under environmental disturbances and mechanical degradation. This study proposes an active disturbance rejection control scheme for ship course keeping, integrating a modified Smith predictor and an ESO to estimate and compensate for total disturbances in real time. Sea trial data from the training vessel Yukun indicate that experienced operators naturally employ discrete steering actions at 7.90–14.87 s intervals, contrasting with conventional autopilots that continuously output control signals for fine-tuning adjustments. To replicate this discrete control behavior, this study integrates a ZOH mechanism with holding time $\theta$ into the ADRC framework, reducing control frequency without compromising accuracy. Testing across Sea States 6–9 demonstrates robust performance. With optimal holding time $\theta$  = 10–15 s, the controller maintains ±5° heading accuracy while reducing steering operations by 43%–58% compared to high-frequency control. The controller maintains performance under ±50% parameter variations without gain scheduling. Software-only implementation enables immediate deployment on existing vessels, providing significant advantages for commercial fleet operations.