Model-free Adaptive Heading Control of Hovercraft Based on Data-driven

This paper addresses the heading control challenge of hovercraft, considering their unique navigational characteristics such as a minimal underwater physical field, limited resistance to wind and waves, and unstable regions. To tackle these issues, an enhanced data-driven control approach has been utilized to investigate the heading control issue in hovercrafts. Initially, the dynamic uncertainty model of the hovercraft is linearized into an equivalent data model using pseudo-partial derivatives. Subsequently, a model-free self-adaptive PD control algorithm for hovercraft heading is developed in conjunction with the PD feedback control law. Additionally, an incremental PID controller is designed to regulate the hovercraft's speed to manage the numerous unstable areas. The efficacy of the proposed data-driven heading control algorithm is substantiated through simulation tests. A distinctive aspect of this research is the development of a hovercraft heading and speed control method that operates without relying on any model information, suitable for control systems that are highly nonlinear and strongly coupled, where establishing an accurate model is challenging. Notably, the control system design is based solely on the input and output data such as the hovercraft's pitch, rudder angle, speed, and heading.