详情

原文传递 Modeling of a Dynamic Wave Environment for Unmanned Surface Vessel Control.

题名：	Modeling of a Dynamic Wave Environment for Unmanned Surface Vessel Control.
作者：	Malia, J. F.
关键词：	Control systems, Simulations, Robotics, Ocean waves, Unmanned surface vehicles, Computer programs, Models, Verification, Test and evaluation, Degrees of freedom, Equations of motion, Graphical user interface, Seakeeping, Maneuvers, Scenarios, Ros (robotic operating system), Aegir time-domain seakeeping code, Usv (unmanned surface vessels), Navasim software, Wam-v (wave-adaptive modular vessels), Gazebo simulation environment, Dof (degrees of freedom), Eom (equations of motion), Gui (graphical user interface), Urdf (unified robot description format), Rtf (real-time factor), Ndw (non-dimensional wavelength)
摘要：	Modeling and simulation methods for the seakeeping dynamics of surface vessels vary widely. Higher fidelity models often demand high computational complexity that limits application to offline computation and are not applicable to all development cycles. These models are often based on solving nonlinear fluid flow equations, which cost time and computational power. Simplified models can reduce complexity, supporting rapid development. For developing control and autonomy of small unmanned surface vessels, choosing model fidelity requires a tradeoff between the accuracy of results and complexity of simulation. This thesis compares two methods for modeling and simulating the seakeeping of a small unmanned vessel: Gazeboan open source, real-time robotics simulator, with an extension that integrates a model for the hydrodynamic wave forces into the rigid-body dynamics engine, and AEGIRa nonlinear motion solver that leverages a high order boundary element method and numerical methods of motion integration. The forces, motions and runtimes are compared for various wave cases. The results suggest that the simplified models generate vessel motions of acceptable fidelity for the development of autonomy, but that for certain wave environments, the differences are significant. The results provide guidance for how the simplified Gazebo extension could be improved without adding significant computational load.
报告类型：	科技报告