一种基于卡尔曼滤波的板球系统RBF-PID控制研究

板球系统在经典控制对象球杆系统的基础上进行了扩展,是一个多变量、强耦合、非线性控制对象。针对板球系统采用PID控制时小球震荡比较严重及PID控制存在参数镇定依靠经验调整且在控制过程中系统的抗干扰能力差等缺点,利用径向基函数神经网络(radial basis funtion)具有自学习、自适应能力及非线性映射能力,对PID参数进行智能优化,使系统对阶跃信号的响应速度加快了2.8 s。利用卡尔曼滤波实现对控制干扰和测量噪声信号进行抑制,通过MATALB绘制出定点控制实验中小球的轨迹,实验结果表明了此算法相对于传统的PID控制可以有效的提高板球的系统的抗干扰能力以及动态性能,能够使小球的稳定时间由26 s减少到7 s,超调量减少了11.4 mm,X轴方向的控制精度提高了0.7 mm,Y轴方向的控制精度提高了1.7 mm。

Research on RBF-PID control of ball and plate system based on Kalman filter

Ball and plate system is extended on the basis of the classic control object ball and beam system, which is a multi-variable, strong coupling and non-linear control object, which is extended on the basis of the classical Ball-Bar system. ball and beam system, which is a multivariable, strong coupling and nonlinear control object. Aiming at the serious spherical oscillation when using PID control in ball and plate system and the shortcomings of parameter stabilization depending on experience adjustment and poor anti-interference ability of the system in the process of control, etc. The radial basis funtion(RBF) neural network has the ability of self-learning, self-adapting and non-linear mapping, and intelligently optimizes the parameters of PID, which accelerates the response speed of the system to step signals by 2.8 seconds. Kalman filter is used to suppress the control interference and measurement noise signal. The trajectory of the ball in the fixed-point control experiment is plotted by MATALB. The experimental results show that the algorithm can effectively improve the anti-interference ability and dynamic performance of the cricket system compared with the traditional PID control, and can make the ball stable. From 26 seconds to 7 seconds, the overshoot is reduced by 11.4 mm, the control accuracy of X-axis direction is improved by 0.7 mm, and that of Y-axis direction is improved by 1.7 mm.