非线性伺服系统的分段参数化摩擦建模与有限时间协同控制

The existence of friction phenomenon will directly affact the tracking performance of noninear servo systems, and even affect the system stability and reliability. Presently, most research approaches on friction modeling and compensation control for nonlinear servo systems mainly have two main issues: First, the existing friction models are proposed only for low-speed friction situations, however, the single low-speed friction modeling method will become not precise any more in the case that the servo system requires higher motion velocity and acceleration, and the velocity direction needs to be changed frequently. Secondly, model uncertainties and external disturbances always exist in nonlinear servo systems so that the friction identification and compensation accuracy will be much deteriorated. In this project, a novel piecewise parameterized friction model is proposed, and an unified identification model of the friciton dynamic is built for nonliear servo systems in high-speed and low-speed motion processes to improve the friciton modeling accuary. By employing the extended state observer (ESO) technique, the uncertainties existing in servo systems can be compensated and thus the dependence of control performance on modeling accuracy is reduced. Moreover, a finite-time synergetic controller is designed to improve the response speed and robustness of servo systems by combining the finite-time control method with the synergetic control theory. Therefore, the research of this project has important scientific research value and practical application prospect.

摩擦现象的存在将直接影响到非线性伺服系统的跟踪性能，进而影响到系统的稳定性和可靠性。目前，大多数非线性伺服系统摩擦建模和补偿控制研究主要存在两方面的问题：首先，现有的摩擦模型大多只是针对低速摩擦的情况而提出的，当伺服系统具有较高的运动速度和加速度要求、需要频繁进行速度换向时，单一的低速摩擦建模方法将会变得不够准确；其次，非线性伺服系统常常会受到模型不确定和外部干扰的影响，使得摩擦辨识和补偿精度大大降低。本项目拟研究一种新的分段参数化摩擦动态模型，建立非线性伺服系统在高、低速运动中摩擦动态的统一辨识模型，提高摩擦建模的精度。通过采用扩张状态观测器技术，补偿伺服系统的不确定因素，降低控制性能对摩擦建模精度的依赖性。此外，结合有限时间控制方法和协同控制理论设计有限时间协同控制器，提高伺服系统的响应速度和鲁棒性。因此，本项目的研究具有重要的科研价值和实际应用前景。

伴随着科技的快速发展，在许多民营工业甚至航天航空领域中广泛应用了伺服系统，进而要求其伺服的性能和精度达到更高的阶层。伺服驱动器直接控制电机，其性能的优劣大幅度影响整套控制系统，优越的伺服控制器响应快、误差小、控制范围宽、控制精度高，能够显著地提高产能效率和精度。但是伺服系统中不可避免的摩擦现象会危害到传动伺服系统的表现情况，甚至会降低伺服系统的性能品质和精度。因此，研究控制方法对摩擦进行补偿，减小摩擦力矩可能会对伺服系统正常运行带来的危害性影响，具有非常重要的理论意义和应用前景。本项目针对非线性伺服系统中的摩擦建模和补偿控制问题开展研究工作。主要贡献包括：采用萤火虫优化的方法对摩擦进行参数辨识，并在此基础上自适应非线性滑模控制器。通过采用扩张观测器技术，补偿伺服系统的不确定因素，降低控制性能对摩擦建模精度的依赖性。提出增强瞬态控制性能的有限时间受限控制方法，保证伺服系统的稳态跟踪误差在有限时间内快速收敛至平衡点附近。设计基于神经网络估计的有限时间滑模控制策略，并在转台伺服系统上完成实验验证。