Requirement modeling and simulation verification of ship course-keeping control system based on MBSE

Objectives To address the dispersion of requirement information, the lack of unified parameter management, and the limited traceability of simulation results to performance requirements in the design of ship course control systems, a model-based systems engineering (MBSE) method for requirement modeling and integrated simulation verification is proposed. Methods Following the requirement–function–logical–physical (RFLP) forward design process, a set of systems modeling language (SysML) models is developed, including the mission profile, requirement decomposition, measure of effectiveness (MoE), system context, internal interface, closed-loop control activity, and integrated simulation activity models. Value properties and object flows are used to establish traceable relationships among requirement thresholds, system parameters, MATLAB variables, and verification metrics. A Norrbin-type nonlinear Nomoto model is employed to describe the ship yaw dynamics, and a fixed-weight radial basis function (RBF) neural network-compensated sliding-mode controller is used for closed-loop verification. Simulations are conducted under step heading commands, sinusoidal heading commands, straight-line course-keeping disturbances, and model parameter perturbations. Results Under the 90° step heading command, the settling time is 169.20 s, the maximum overshoot is 0.84°, and the steady-state heading error is 0.37°. For the sinusoidal heading command with an amplitude of 30° and an angular frequency of 0.01 rad/s, the tracking root mean square error is 1.23°. Under composite disturbances during straight-line course keeping, the steady-state heading error is 0.77°. When the ship model parameters are significantly perturbed, the system remains capable of tracking a 60° step heading command, with a maximum overshoot of 2.41° and a steady-state heading error of 1.15°. The rudder angle remains within the prescribed limit of ±35° in all simulation cases, The rudder usage levels under all four operating conditions satisfy the steering economy constraint. Conclusions The proposed method provides an integrated representation of course control requirements, system architecture, interface behavior, model parameters, and simulation metrics. It supports requirement satisfaction assessment under multiple operating conditions while improving model consistency and result traceability in the design verification of ship course control systems.