含噪声扰动的板球系统控制研究综述

Review of research on control of ball and plate system with noise and disturbance

板球系统是一个非线性、欠驱动、强耦合二维拓展平面,对其开发研究成果可以应用到各种平台上。回顾板球系统的非线性、不确定性、摩擦力和噪声等扰动影响因素,从系统模型、控制方法以及噪声扰动补偿策略等角度切入,分别在动力学模型和视觉伺服模型2个层面对国内外的研究进展进行综述。当前运动控制研究多集中于较为理想的简化数学模型,为实现板球系统的精确控制,运动控制研究须从常规工况向极限工况拓展,但是极限工况下板球系统的非线性和强耦合特征显著增强,对系统建模以及控制算法的自适应性和鲁棒性的要求提高。同时,为应对复杂数学模型下的包含噪声扰动等不确定性扰动问题,考虑干扰观测器和滤波器扰动补偿策略需要深入研究。扰动补偿可以抑制或消除不确定性扰动或未建模动态模式的影响,而滤波器则可以有效处理测量噪声和过程噪声估计动态系统的状态。运动控制的实现依赖于高效先进的控制算法,可见开发高效控制算法亟需解决。最后归纳了板球控制系统目前存在的主要难点,结合小波神经网络对板球控制系统存在的问题及研究方向进行了展望。

Ball and plate system(B&P system)is a nonlinear,under-actuatedand strongly coupled two-dimensional expansion plane.Research results of its development can be applied to various platforms.We reviewed the research issues on the impact of disturbances such as nonlinear,unknown uncertainty,friction and noise of the B&P system,from the perspective of system models,control algorithms and noise disturbance compensation strategies,the research progress at home and abroad was reviewed at two levels of dynamic model and visual servo model.Current research on motion control mostly focuses on the ideal simplified mathematical model.In order to realize the precise control of the B&P system,the research on motion control must be expanded from the normal operating conditions to the extreme operating conditions.However,the nonlinear and strong coupling characteristics of the B&P system under extreme operating conditions have been significantly enhanced.The requirements for adaptability and robustness were further improved.In order to deal with uncertain disturbances including noise disturbances under complex mathematical models,the disturbance observer and filter disturbance compensation strategies need to be studied in depth.Disturbance compensation can suppress or eliminate the influence of uncertain disturbances or unmodeled dynamic modes,while the filter can effectively handle measurement noise and process noise to estimate the state of the dynamic system.The realization of motion control relies on efficient and advanced control algorithms,so the development of efficient control algorithms needs to be solved urgently.Finally,we summarized the main difficulties currently existing in the B&P control system,combining the wavelet neural network and looking forward to the problems that still exist in the B&P control system.