Development of automatic gain-tuning algorithm for heading control using free-running test data

Abstract

This study proposes an automatic gain-tuning algorithm for ships. The proposed algorithm is designed to tune the gains of the ship controller automatically, rather than using trial and error. The forward speed and steering models were derived by linearizing and simplifying the 3-degrees of freedom (DOF) nonlinear equation of motion of the ship. The initial control gains were calculated using an error dynamics model constructed by combining the steering and system models of the controller. The maneuvering simulations and sensitivity analysis of the control performance at various control gains were performed for gain-tuning. System identification was conducted based on derived dynamics models and free-running test data. The tests verified that the gain-tuning algorithm corrects the gains more accurately and rapidly than trial and error. In addition, the algorithm reduced overshoot by 85% compared to the initial control gains.