| Summary: | 博士 === 國立臺灣大學 === 工程科學及海洋工程學研究所 === 101 === An intelligent autonomous underwater vehicle (AUV) simulator with improved navigation, autopilot and guidance systems that constrained by maneuvering models was developed. The dynamics system of the AUV’s navigation on the base of 3-D Euler-Lagrange equations of motion in six degrees of freedom (6-DOF) was integrated with Euler-Rodriguez quaternion, Euler-angle and Euler-axis methods to represent singularity-free AUV’s attitude intuitively; the fourth-order Runge-Kutta method was a time-marching model in the simulator. In this study, the simulator with a quaternion-based control system was used to test motion performance, maneuverability both of a mini AUV and a large-scale AUV, including tracking performance, path motion stability, course changing, course keeping and diving abilities, etc. For validation of the simulation codes, experimental results of the ISiMI AUV open-loop tests, including turning test and zigzag test, were used to compare with simulation results of the AUV simulator. Comparisons of the results implied the adopted methods and hydrodynamics reasonable. The proposed guidance system in the simulator includes a 3-D line-of-sight (LOS) algorithm, a quaternion-based proportional-Integer-derivative (PID) controller and a path generator, which automatically generates continuous-curvature paths of cubic B-spline class constrained by AUV maneuverability, 3-D Euler-Lagrange formulation and waypoints. Gauss-Legendre method was applied to calculate length of objective paths. A 3000-T AUV was used to test the guidance system. Comparisons of linear and cubic path-planning strategies were discussed, including a straight line and a conventional cubic spline method, three parametric methods for planning cubic B-spline paths, and an iterative method for improving and expanding the function of the path generator. Simulation results of the tracking performance tests show that the AUV can precisely approach targets using the proposed method. The improvement in the cross-tracking error was approximately 80%, whereas reduction in travelling time was 5% in this case. In addition, a Web-based 3-D AUV visualization system was developed to render 3-D models, attitudes and position of AUV with Web-based interactive function experienced.
|
|---|
