欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对扫描隧道显微镜工作台高精度控制的需求，设计了一种基于神经网络理论的控制系统，其微位移工作台由压电陶瓷驱动器和柔性铰链机构组成。在对驱动器结构进行分析的基础上，建立了工作台的数学模型。以三层神经网络自学习PID控制器代替常规PID控制器，实现了样本的在线采集和优化，有效地克服了神经网络控制器需要离线训练的缺点；采用BP算法对神经网络进行在线训练，增强了系统的实时控制性能。实验结果表明，10μm下的过渡时间从3．25s缩短到1．6s，稳态误差从2．78％减小到1．39％。
英文摘要：
According to the requirements of high precision control for the stage of STM, a new control scheme based on neural network was proposed. The micro-displacement stage was composed of a piezoelectric ceramic actuator and a flexure hinge. The principle of the actuator was analyzed and the mathematical model was set up. The traditional PID controller was replaced by a 3-layer neural network self-learning PID controller. The on-line collection and optimization of sample were realized. The disadvantage of controller＇s off-line training was overcome. The neural network was trained by the BP algorithm online. The performance of the stage real time control was enhanced. The experimental results show that the response time of 10μm is shortened from 3.25s to 1.6s and the steady error is reduced from 2. 78%to 1.39%.
同期刊论文项目
同项目期刊论文
期刊信息
位置：