基于视觉检测板球系统的RBF-PID控制研究

Research on RBF-PID Control Based on Visual Detection Ball and Plate System

板球系统是一个典型的多变量、不确定、强耦合的非线性系统,一直被国内外许多学者作为研究非线性控制系统的实验平台。文中通过动力学分析得到板球系统的数学模型和相应的简化模型,针对系统震荡较严重的问题,在RBF神经网络参数调整中加入动量因子,减小系统震荡;对系统响应速度慢的问题,采用LM算法代替梯度下降法对PID参数进行在线优化。实验结果表明,该控制算法可使控制误差缩小在2 mm以内,控制时间减少了3 s且具有较好的鲁棒性。

Ball and plate system is a typical multivariable,uncertain and strongly coupled nonlinear system. It has been used by many scholars at home and abroad to study the experimental platform of nonlinear control system. In this paper,the mathematical model of the ball and plate system and the corresponding simplified model are obtained by the kinetic analysis. In the case of system shock,the momentum factor is introuduced to the RBF neural network parameter adjustment to reduce the system oscillation. In the view of system slow response,the LM（ Levenberg-Marquardt） algorithm is adopted to optimize the PID parameters in instead of the gradient descent method. The experimental results show that the control algorithm can reduce the control error within 2 mm,the control time is reduced by 3 s and has good robustness.