基于切换策略的飞机蒙皮检测机器人运动控制研究

发布时间：2018-09-10 10:15

【摘要】：随着中国航空工业迅速发展,飞机已经与每个人的生活息息相关,因此飞机蒙皮表面损伤带来的安全隐患备受关注。然而飞机蒙皮检测仍采用人工检查的方式,存在劳动强度大、检测效率低、漏检率高以及成本费用高等问题,此外也不能够完全保障飞机维修人员的操作安全。本文针对双框架飞机蒙皮检测机器人在飞机表面做稳定步态运动,研究了双框架飞机蒙皮检测机器人的切换策略与运动轨迹控制。首先,介绍了一种真空吸附式双框架飞机蒙皮检测机器人的机械结构以及分析了在飞机蒙皮表面的运动步态;针对该机器人的吸盘吸附系统,描述了刚性切换和柔性切换两种切换策略;根据检测机器人在飞机蒙皮表面的受力分析,建立了机器人的动力学模型。其次,根据机器人自身的特点,设计了一种基于事件驱动的切换策略。针对机器人非线性系统,研究了基于支持向量回归的自适应控制器,利用干扰上界值设计辅助变量来估计系统中的外部干扰。仿真结果表明,本文所给出的基于支持向量回归的自适应控制律可以保证存在未知干扰的机器人系统有良好的轨迹跟踪性能和鲁棒性能。然后,利用切换系统描述飞机蒙皮检测机器人整体姿态变化。针对切换系统模型,利用Backstepping方法,研究了机器人轨迹跟踪控制方法,通过公共Lyapunov函数证明了切换系统的闭环稳定。当切换系统中存在未知干扰,利用Nussbaum干扰观测器进行了有效估计,并将非线性干扰观测器和基于Super-twisting算法的干扰观测器与Nussbaum干扰观测器进行性能比较。利用平均驻留时间法证明了切换系统的闭环稳定性,仿真结果表明在未知干扰的作用下,该控制器可以保证机器人切换系统运动控制的稳定性。接着,研究了具有输入受限、外部干扰和参数未知的飞机蒙皮检测机器人运动控制方法。针对机器人系统存在的输入受限问题,利用双曲正切函数将饱和输入转化为光滑的控制输入,结合Backstepping控制方法,新增的子系统可以有效抑制输入受限对系统的影响;针对机器人系统中的外部干扰和未知参数,设计了Nussbaum干扰观测器并利用Lyapunov函数证明了切换系统的闭环稳定性,仿真结果表明了该控制器是有效的。最后,开展了飞机蒙皮检测机器人的运动控制实验研究,将机器人分别吸附于平面和斜面上进行了切换运动,对机器人的刚性切换和柔性切换两种吸盘切换策略进行了相关实验分析,验证了检测机器人切换系统的可靠性和稳定性。
[Abstract]:With the rapid development of China's aviation industry, aircraft has been closely related to everyone's life, so the safety risks caused by surface damage of aircraft skin have attracted much attention. However, manual inspection is still used in the detection of aircraft skin, which has many problems, such as high labor intensity, low detection efficiency, high rate of missing detection and high cost. In addition, it can not completely guarantee the operation safety of aircraft maintenance personnel. In this paper, aiming at the stable gait movement of a double-frame aircraft skin detection robot on the plane surface, the switching strategy and motion trajectory control of the double-frame aircraft skin detection robot are studied. Firstly, the mechanical structure of a vacuum adsorbed double-frame aircraft skin detection robot is introduced, and the moving gait on the surface of the aircraft skin is analyzed. In this paper, two switching strategies, rigid switching and flexible switching, are described, and the dynamic model of the robot is established according to the force analysis of the detection robot on the surface of the aircraft skin. Secondly, according to the characteristics of the robot, an event-driven switching strategy is designed. An adaptive controller based on support vector regression is studied for nonlinear robot systems. The upper bound of disturbance is used to design auxiliary variables to estimate the external disturbances in the system. Simulation results show that the proposed adaptive control law based on support vector regression can guarantee the trajectory tracking and robustness of robot systems with unknown disturbances. Then, the switching system is used to describe the whole attitude change of the aircraft skin detection robot. Aiming at the switching system model, the trajectory tracking control method of the robot is studied by using the Backstepping method. The closed-loop stability of the switching system is proved by common Lyapunov function. When there is unknown disturbance in switched systems, the Nussbaum disturbance observer is used to estimate effectively, and the performance of nonlinear disturbance observer and Nussbaum disturbance observer based on Super-twisting algorithm is compared with that of Nussbaum disturbance observer. The closed-loop stability of the switching system is proved by the mean dwell time method. The simulation results show that the controller can guarantee the stability of the robot switching system motion control under the action of unknown disturbance. Then, the motion control method of aircraft skin detection robot with limited input, external interference and unknown parameters is studied. Aiming at the problem of input limitation in robot system, the hyperbolic tangent function is used to transform the saturated input into smooth control input. Combined with the Backstepping control method, the new subsystem can effectively suppress the influence of input restriction on the system. Aiming at the external disturbance and unknown parameters of robot system, the Nussbaum disturbance observer is designed and the closed-loop stability of the switched system is proved by using Lyapunov function. The simulation results show that the controller is effective. Finally, the motion control experiment of the aircraft skin detection robot is carried out. The robot is adsorbed on the plane and the oblique plane respectively. Based on the experimental analysis of rigid switching and flexible switching, the reliability and stability of the robot switching system are verified.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】