【作者】 黄飞杰；

【导师】 贾庆轩；

【作者基本信息】 北京邮电大学 ， 机械电子工程， 2013， 硕士

【摘要】 随着人类太空探索活动的快速发展,空间机械臂被广泛的应用于在轨服务、空间对接及星球探测等重要场所,特别是对于那些复杂繁重、要求准确性高而又环境恶劣的空间任务,以空间机械臂代替人工进行作业必不可少,空间机械臂高精度控制技术已成为当前机器人领域的重要研究方向。本论文来源于教育部博士点基金课题“太空柔性机械臂的低速高精度定位与快速振动抑制(20110005120004)”,以机械臂柔性关节为研究对象,重点对其动力学特性及高精度控制策略进行了深入的分析与研究,本文的主要工作如下：首先,针对当前机械臂研究应用状况,在对影响关节控制性能因素分析的基础上,建立柔性关节模型,给出关节动力学方程；其次,为了克服关节柔性、摩擦、间隙及扰动等因素对柔性关节高精度控制性能的影响,将自抗扰控制方法引入柔性关节控制器设计,完成柔性关节双位置闭环反馈的自抗扰控制系统设计,并进行控制策略的数值仿真和实验验证；再次,在分析柔性关节传统反演控制设计方法不足的基础上,提出基于扩张状态观测器的动态曲面控制策略。采用扩张状态观测器观测系统状态,全局采用Lyapunov函数的动态曲面方法设计了具有稳定性和鲁棒性的控制器,并给出系统稳定性分析及数值仿真和实验结果。所提出控制策略不仅克服了反演控制设计过程中带来的“计算膨胀”问题,且无需检测电机速度及关节速度。再其次,研究了柔性关节的神经网络控制策略,提出了柔性关节自回归小波神经网络-动态曲面控制策略。所提出的控制策略采用自回归小波神经网络对关节控制系统的不确定项进行逼近,进一步结合动态曲面设计方法,设计了具有全局稳定性的柔性关节自回归小波神经网络-动态曲面控制系统。所采用的神经网络具有快速的学习能力和收敛性,所设计的控制器能快速的调整自身输出以抑制各种不确定因素及扰动的影响,保证关节具有高精度位置跟踪性能。最后,搭建柔性关节实验平台,编写关节运动控制程序,对所提出的柔性关节控制策略进行实验验证。更多还原

【Abstract】 With the rapid development of the human space exploration activities, the space manipulators are widely used in the on-orbit servicing, space docking and planetary exploration and some other important places. Especially for those heavy and complicated tasks, meanwhile high accuracy is required and have a bad work environment, the manipulators are essential needed to replace human, the high accuracy and robust control technology of space manipulator has become an important research issue in the field of robotics. This paper comes from the Doctoral Fund of Ministry of education project "low speed space flexible manipulator high precision positioning and rapid vibration suppression", deeply focus on the dynamic and high accuracy control strategies of flexible joint of manipulator, the main work is as follows:Firstly, focus on the research and application of manipulator, based on the analysis of factors that affect the flexible joint control performance, the model of flexible joint is established, and the dynamic equation is given.Secondly, in order to overcome the impact of joint flexibility, friction, clearance and disturbance, the Active Disturbance Rejection Control (ADRC) is introduced to finish the double position closed loop feedback ADRC system, numerical simulation and experiment are designed to verify the control strategies.Thirdly, analyze the shortcoming of the traditional Backstepping control design of flexible joint, the Dynamic Surface Control(DSC) based Extended State Observer (ESO) control strategy is proposed. Apply the ESO to observed system status, the global stability and robustness controller is designed using DSC technique under the Lyapunov function, the stability analysis and numerical simulation and experiment are given. The proposed control strategy not only overcome the "calculation of explosion" in Backstepping design, but without the measurement of motor velocity and joint velocity.Again, study the neural network control strategy of flexible joint and the Self-Recurrent Wavelet Neural Network-Dynamic Surface Control strategy (SRWNN-DSC) is proposed. The strategy apply the SRWNN to approach uncertain items of flexible joint system, further combined with DSC design technique to accomplish the global stability SRWNN-DSC system of flexible joint. The applied network has a fast learning ability and convergence; the designed controller can quickly adjust its output to inhibit a variety of uncertainties and disturbance and ensure the high precision tracking performance of flexible joint.Finally, design a flexible joint experimental platform, write and download the movement control program to verify the proposed control strategies.更多还原