柔性滤波驱动机构的自适应控制方法研究

The flexible filter drive mechanism has the characteristics of time varying, uncertain parameters, multi-field coupling, multivariable, complicated friction behavior and etc. Most of the existing intelligent control methods which can control for ideal model are finding difficult to achieve the high quality control of the flexible filter drive mechanism. In order to gain a better control, this research program is divided into four steps. Firstly, for the executive component—servo motor existing control problem with chaos, input and output constraint, unknown parameter etc., the adaptive control method is put forward; Secondly, aimed at the servo motor control problem with the state which is difficult to accurately measure and time-varying delay, an observer-based adaptive control method is proposed to overcome the position or velocity signal immeasurabilty and unknown parameters, and extend the limitation on time-varying delay and its upper bound; Thirdly, to avoid the shortcomings of mechanical structure for the harmonic drive mechanism, the design of the flexible filter drive mechanism can be improved, and mathematical model is established involving in time-varying meshing stiffness, flexible deformation, dynamic and static friction and transmission error etc., then a adaptive control method is put forward based on LuGre friction mode; Finally, considering the influence of the actuator fault and unpredictable state, an adaptive fault-tolerant control method is designed for the flexible filter drive mechanism in the fault state. The project has important scientific significance and application value on promoting development of high quality control for the flexible filter drive mechanism (including servo motor) and improving system performance.

柔性滤波驱动机构具有时变、参数不确定、多场耦合、多变量和摩擦行为复杂等特点。现有智能控制方法大多是针对理想模型进行控制，难以实现柔性滤波驱动机构的高品质控制。本项目拟首先针对执行元件—伺服电机存在的混沌、输入输出约束、参数未知等控制问题，提出自适应控制方法；然后，针对具有状态不易精确测量、变时延的伺服电机控制问题，提出基于观测器的自适应控制方法，克服位置或速度信号不易测和参数未知的困难，放宽对变时延及其上界的限制；其次，为避免机械结构上谐波驱动机构的不足，改进柔性滤波驱动机构的设计，建立涉及时变啮合刚度、柔性变形、动静态摩擦及传动误差等因素的数学模型，提出基于LuGre摩擦模型的自适应控制方法；最后，考虑执行器故障和状态不可测的影响，设计柔性滤波驱动机构在故障下的自适应容错控制方法。本项目研究对促进柔性滤波驱动机构（包括伺服电机）高品质控制发展、提升系统的性能具有重要的科学意义与应用价值。

柔性滤波驱动机构具有时变、参数不确定、多场耦合、多变量和摩擦行为复杂等特点。现有智能控制方法大多是针对理想模型进行控制，难以实现柔性滤波驱动机构的高品质控制。针对执行元件—伺服电机存在的混沌、输入输出约束、参数未知等控制问题，建立具有时变时滞的永磁同步电机模型，将在处理未知扰动和系统动态上具有逼近功能的单权值RBF神经网络应用到控制器设计中，采用快速终端滑模面来提高跟踪精度，在融合快速终端滑模面、一阶低通滤波器和RBF神经网络的情况下，提出自适应滑模控制方法；然后，利用分数阶微积分理论对被控伺服电机系统动力学特性更为精确的表征，采用频率分布模型得到真实状态空间方程，构建一阶滤波器解决传统反演法计算膨胀问题，设计高增益观测器来在线观测不可测的状态信号，基于回声状态神经网络在反演法的框架中提出自适应镇定方案；其次，建立涉及时变啮合刚度、柔性变形、动静态摩擦及传动误差等因素的数学模型，提出基于LuGre摩擦模型的自适应控制方法；最后，将自适应控制技术推广应用于微机电系统和机械离心式飞轮调速系统，提出了具有扩展状态观测器的微机电系统自适应动态面控制技术，通过自适应动态面控制技术设计了机械离心式飞轮调速系统的混沌控制器，深入研究了分数阶微机电系统的自适应控器中控制律设计、稳定性证明、参数整定等过程。本项目研究对促进柔性滤波驱动机构（包括伺服电机）高品质控制发展、提升系统的性能具有重要的科学意义与应用价值。