| Abstrakt: |
The objective of the navigation process is to determine the most efficient route for the hovercraft and regulate its movement along that route without any oscillation. The primary goal of this work is to determine the most efficient routes for a hovercraft operating in a global environment. The proposed approach, called Artificial Bee Colony Self Perception Particle Swarm Optimization (ABC-SPPSO), is utilized to accomplish this objective. The main benefit of utilizing the ABC-SPPSO algorithm is its ability to design the most efficient route while preventing collisions with stationary obstructions. In addition, a suggested controller that consists of a Feedforward Numerical Inverse Dynamic Controller (FNIDC) and a Feedback Neural Network Radial Basis Function (FNNRBF) control technique with the Grey Wolf Optimization (GWO) algorithm will be utilized to guide the hovercraft along predetermined paths. This suggested controller is designed to regulate the nonlinear dynamics of the hovercraft system in order to efficiently and rapidly generate the forces exerted by the starboard and portboard fans. These forces are utilized to control the orientation and position of the hovercraft. Moreover, the utilization of the suggested controller effectively reduces the differences between the desired and the actual positions in both the X-axis and the Y-axis. Additionally, the controller nearly eliminates any deviation in orientation and ensures a stable response without any oscillation. Specifically, the controller ensures that the hovercraft will promptly and accurately adhere to its intended trajectories. Ultimately, we verify the accuracy of the numerical simulation outcomes of the suggested control approach by contrasting them with those of other controllers, specifically in relation to the highest level of error improvement in the X-position and the Y-position. In particular, when comparing the proposed controller to the Improved Quasi-Velocities (IQV) controller, the results show that the suggested controller decreases the error rate of tracking on the X-position by 22% and enhances the tracking error rate on the Y-position by 14.8%. Furthermore, the suggested controller was evaluated against the terminal sliding mode controller (TSMC), and the results of the comparison indicate that the proposed controller enhances the error rate of tracking on the X-position by 50.7% and on the Y-position by 64.5%. Additionally, the suggested controller was evaluated against the nonlinear cascade controller, and the comparative analysis demonstrates that the suggested controller enhances the X-position and the Y-position tracking error rates by 25.9% and by 33%, respectively. Finally, from a comparative study with the neural network-based adaptive dynamic inversion controller, the proposed controller enhances the error rate of tracking on the X-position by 51% and on the Y-position by 42.9%. [ABSTRACT FROM AUTHOR] |
