| 梁捷,陈力.柔性关节-柔性臂空间机器人的神经网络自适应反演控制及双重柔性振动抑制[J].计算力学学报,2014,31(4):459~466 |
|
| 
码上扫一扫！ |
| 柔性关节-柔性臂空间机器人的神经网络自适应反演控制及双重柔性振动抑制 |
| Neural network adaptive backstepping control and double flexible vibration active hierarchical suppression of space robot with flexible-joint and flexible-link |
| 投稿时间：2013-05-21 修订日期：2013-08-31 |
| DOI：10.7511/jslx201404008 |
| 中文关键词: 柔性关节柔性臂空间机器人 动力学模拟 高斯径向基函数神经网络 反演控制算法 关节柔性补偿控制器 |
| 英文关键词:space robot with flexible-joint and flexible-link Gaussian radial basis function neural network backstepping control algorithm joint flexibility compensating controller |
| 基金项目:国家自然科学基金（11372073，11072061，10672040）；福建省自然科学基金（2010J01003）资助项目. |
|
| 摘要点击次数: 1932 |
| 全文下载次数: 1569 |
| 中文摘要: |
| 空间机器人系统的柔性主要体现在空间机器人的臂杆和连接各臂杆之间的铰关节。由于空间机器人系统结构的复杂性，以往研究人员对同时具有柔性关节和柔性臂的系统关注不够。为此探讨了参数未知柔性关节-柔性臂空间机器人系统的动力学模拟、轨迹跟踪控制算法设计和关节、臂杆双重柔性振动的主动抑制问题。首先，采用多体动力学建模方法并结合漂浮基空间机器人固有的线动量和角动量守恒动力学特性，推导了系统的动力学方程。以此为基础，考虑到空间机器人实际应用中各关节铰具有较强柔性的情况，引入一种关节柔性补偿控制器解决了传统奇异摄动法应用受关节柔性限制问题，导出了适用于控制系统算法设计的数学模型。然后，利用该模型，基于反演思想在慢时标子系统中设计神经网络自适应控制算法来补偿系统参数未知和柔性关节引起的转动误差，实现系统运动轨迹跟踪性能；针对快时标子系统，设计了鲁棒最优控制算法抑制因柔性关节及柔性臂引起的系统双重弹性振动，保证系统的稳定性。最后，通过仿真对比实验验证了所设计控制算法的有效性。 |
| 英文摘要: |
| The flexibility of space robot system is mainly embodied in links of space robot and their connected hinge joints.For the reason of complexity of space robot system structure, researchers usually pay less attention on the system which both has flexible joint and link.So it's discussed in this paper for dynamics simulations, algorithm design of trajectory tracking control, and active suppression problem of joint and link double flexible vibration for space robot system with flexible-joint and flexible-link under the situation of parameter uncertain.Firstly, the dynamics equation of system is deduced by modeling multi-bodies dynamics method and combing with connatural dynamics characters of linear momentum and angular momentum of free-floating space robot.And on basis of above, as each joint hinge has strong flexibility in practical use of space robot, a joint flexibility compensation controller is introduced to solve such problem that application of traditional singular perturbation method is limited by joint flexibility, so mathematical model that be suitable for the design of control systems is established.Then using the mathematical model, a neural network adaptive control algorithm that based on the idea of backstepping in the slow subsystem is designed to compensate system parameter unknown and rotation error caused by flexible joint, thus system trajectory tracking performance is fulfilled.For the fast subsystem, a robust optimal control algorithm is presented to suppress the system double elastic vibration that caused by flexible joint and flexible link, thus the stability of the system is guaranteed.Finally, the availability of said designed control algorithm is testified by the simulation comparison experiment. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
