论文部分内容阅读
根据形状记忆合金(SMA)直线驱动的特点,提出了基于BP神经网络的控制方法,探讨了该方法在控制系统中的应用。通过分析研究SMA的主要特性,获得了形状记忆合金性能参数,对人工神经网络方法进行了研究,并结合MATLAB神经网络工具箱展开编程工作。结合SMA丝硬件驱动器,完成了人工神经网络算法程序,获得训练后应变量与控制量之间的拟合关系。实验结果表明,人工神经网络训练效果良好,吻合度高,我们可以利用所得曲线,在已知所需应变量的情况下,反向控制PWM波的占空比,达到精确控制的目的。
According to the characteristics of linear drive of shape memory alloy (SMA), a control method based on BP neural network is proposed, and its application in control system is discussed. By analyzing and studying the main characteristics of SMA, the performance parameters of SMA are obtained, and the method of artificial neural network is researched. The programming work is carried out by combining MATLAB neural network toolbox. Combined with SMA wire hardware driver, the artificial neural network algorithm program was completed, and the fitting relationship between the amount of response and the amount of control after training was obtained. The experimental results show that the artificial neural network has a good training effect and high coincidence degree. We can use the obtained curve to control the duty cycle of the PWM wave in the reverse direction under the condition of known required amount of strain, and achieve the purpose of precise control.