博士 / 國立成功大學 / 系統及船舶機電工程學系 / 105 / This dissertation proposes integrating an H∞ controller with an artificial potential field method (APFM) to solve collision avoidance and navigation issues in an autonomous underwater vehicle (AUV). We applied APFM in designing three types of control algorithms—altitude, depth, and heading; the AUV used the proposed control algorithms to navigate in unknown three-dimensional static environment and avoid collisions with obstacles. The depth-control algorithm involved a safe altitude above the seafloor to prevent the AUV from colliding with the seafloor or vertical obstacles. The altitude-control algorithm involved a maximum safe depth below the surface to prevent the AUV from following terrain beyond the maximum pressure of vessel strength. The heading-control algorithm was based on an improved APFM to solve the local minimum problem; it combined APFM with an obstacle boundary following method (OBFM) to solve the problems of the AUV falling into the U-shaped trap and repetition path hovering.
For simulation analysis, we modeled a self-developed National Cheng Kung University Autonomous Underwater Vehicle (NCKU-AUV) as a device under test. Using a planar motion mechanism (PMM) test, we obtained the hydrodynamic force coefficient of the vehicle. We applied the system simulation parameters for building the mathematical model used to design the H∞ controller. We used the loop shaping method to adjust the weight function, obtaining the optimal controller. The simulation results showed that the controller had robustness and anti-interference properties, met system performance requirements, and provided stability. Additionally, we tested the control algorithms for altitude, depth, and heading, simulating seafloor with different depths, different terrain, and various types of obstacles; the simulation results showed that the proposed collision avoidance algorithm was effective and guided the NCKU-AUV to avoid the underwater obstacles safely and correctly. Using the simulated H∞ controller, the simulated NCKU-AUV was able to accurately navigate to the appointed target.
For practical testing, physical sensors including a Doppler velocity log, depth gauge, altimeter, and inertia measurement unit were installed on the physical NCKU-AUV to measure its velocity, altitude, depth, and attitude angle. In particular, five sonar sensors were installed at the bow end of the vehicle to detect horizontal and vertical obstacles; these sonar sensors were able to measure the distances between the vehicle and obstacles. AUV trials were carried out in the towing tank at NCKU. Obstacles were placed at the bottom of the towing tank, and the walls of the tank acted as a horizontal U-shaped obstacle. All control tests (fixed altitude, fixed depth, and navigation) were conducted and completed successfully. Preliminary results of the towing tank test validated the feasibility and effectiveness of the proposed H∞ controller with APFM; real sea testing could be conducted in future to prove the practicality of this system.
| Identifer | oai:union.ndltd.org:TW/105NCKU5345044 |
| Date | January 2017 |
| Creators | Shun-MinWang, 王舜民 |
| Contributors | Ming-Chung Fang, Cheng-Neng Hwang, 方銘川, 黃正能 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | en_US |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 135 |
Page generated in 0.0128 seconds