针对执行器不确定非光滑约束特性的模糊建模和控制方法

With the development of control technique in manufacturing fields including multi-axis precision machining, piezoelectric actuator and high precision industrial robots, the requirements for sensitivity and control accuracy of actuators are becoming more stringent. When suffered from the nonlinear actuator constraints, the control signals are severely distorted such that the control performances are dramatically threatened and thus the oscillation or instability is generated. This project will study control problems from nonlinear actuator constraints in the following three aspects. Firstly, the fuzzy modeling technique is carried out to divide the non-smooth nonlinear constraints into single-valued and multi-valued mapping cases. In what follows, the novel identification methods are employed to each case based on the learning data sets. Secondly, under the recognized conditions, the feasible inverse function technique and hybrid intelligent adaptation mechanism, as well as modified Lyapunov control tools, are proposed to achieve the satisfying transient and steady performances. Thirdly, without the recognized conditions, the nonlinear decomposition technique is constructed to solve the control difficulties form unknown control gain and control singularity. In this research, the effective fuzzy modeling techniques and identification methods will be presented and robust controller will be provided to tackle the key scientific problems from nonlinear asymmetrical actuator constraints. This project not only goes in great depth with theoretical contributions, but also is of great application prospects.

现代高精密执行部件中往往存在非光滑非线性约束，当控制信号通过非线性约束环节时，其相位、频率、形状或幅值等特性将随之发生变化，甚至出现异常波动、振荡甚至失稳现象，将对控制性能造成极大的负面影响。本项目将围绕执行器非光滑约束探讨相关控制问题: 1）充分考虑执行器未知非光滑约束的动态和参数不确定性，分别针对具有单值映射和多值映射的非光滑非线性约束提出与之匹配且新颖的模糊建模和辨识方法。2）当执行器非光滑约束模型参数可辨识时，提出了光滑逆补偿控制方案并进一步研究基于平滑设计的逆模型补偿方法、基于解耦的分层自适应机制和平衡稳/暂态性能的主动受限型Lyapunov设计；3）当参数难以辨识时，围绕执行器未知非光滑约束设计线性/非线性分解策略，并解决分解过程产生的未知控制增益和控制奇异等相关控制问题。总之，在考虑执行器非光滑约束的情况下实现有效的模糊建模和控制设计是控制理论面向应用亟待解决的关键科学问题。

本项目围绕执行器非光滑约束探讨相关控制问题: 1）充分考虑执行器未知非光滑约束的动态和参数不确定性，分别针对具有单值映射和多值映射的非光滑非线性约束提出与之匹配且新颖的模糊建模和辨识方法。2）当执行器非光滑约束模型参数可辨识时，提出了光滑逆补偿控制方案并进一步研究基于平滑设计的逆模型补偿方法、基于解耦的分层自适应机制和平衡稳/暂态性能的主动受限型Lyapunov设计；3）当参数难以辨识时，围绕执行器未知非光滑约束设计线性/非线性分解策略，并解决分解过程产生的未知控制增益和控制奇异等相关控制问题。4）将本项目所提出的算法应用到双臂机器人系统、双边遥操作系统和视觉伺服机械系统等实际控制对象，成功解决上述实际系统机械臂具有非光滑约束而导致的控制精度不精确难题。近4年共发表录用论文84篇，SCI收录的国际期刊论文71篇，其中《IEEE Trans.》和《Autotmatica》36篇；会议论文13篇。申请发明专利17项。项目实际取得成果已超出预定目标。