Design of control system for autonomous airship and experiment

飞艇是一种独特的工具，可以实现安全和长时间的飞行和站位保持。近年来，人们对无人飞艇的兴趣与日俱增。许多有趣的项目正在进行中，如平流层平台项目，在该项目中，平流层飞艇被用作高速无线通信的移动和低成本平台，而不是卫星。位置和姿态控制是最重要的部件技术之一，本文研究的是欠驱动飞艇的平面位置和姿态控制问题，飞艇是一个独立输入比控制自由度少的非完整系统。考虑了欠驱动飞艇动力学存在有界不确定性的情况下的鲁棒全局镇定问题。利用光滑的时变坐标变换将飞艇问题简化为线性系统的问题。该变换与系统的初始状态无关，并且是非奇异的。获得了以下优点。首先，在动态已知的情况下，提出了一种新的控制器来保证系统在任何初始状态下的指数稳定性。基于著名的线性控制理论的设计方法简单明了。其次，研究了参数未知但有界的不确定飞艇的鲁棒镇定问题。新问题归结为不确定线性系统的二次镇定问题。设计方法简单，且涉及到一定的代数Riccati方程。最后，通过实验验证了所提出的控制系统的有效性。

Airship is a unique vehicle which enables safe and long-duration flight and station keeping. Recently, the interest to unmanned airships has been increasing. Many interesting projects are in progress such as Stratospheric Platform Project, in which the stratospheric airship is utilized as a mobile and low-cost platform for high-speed wireless communication instead of satellites. The position and attitude control is one of the most important component technologies.This study deals with the problem of controlling both the planar position and orientation of an under-actuated airship, which is a nonholonomic system with fewer independent inputs than degree of freedom to be controlled. A robust global stabilization problem is considered for the case where the dynamics of the underactuated airship are subjected to bounded uncertainties. A smooth and time-varying coordinate transformation is utilized to reduce the problem of the airship to that of a linear system. The transformation is independent of the initial state of the system and is nonsingular. The following advantages are obtained. First, the case is dealt with where the dynamics are known and presents a new controller to guarantee exponential stability of the system for any initial states. The design method based on the well-known linear control theory is simple and straightforward. Second, the robust stabilization problem is considered for the uncertain airship with unknown but bounded parameters. The new problem is reduced to the well-known quadric stabilization problem for an uncertain linear system. The design method is simple, and it involves a certain algebraic Riccati equation. Finally, experimental results show the effectiveness of the proposed control system.