中文摘要：

由于网络数据多元化以及网络非法入侵的干扰,当前设计出的网络稳定控制器往往响应时间过长,提出基于机器学习的网络稳定控制器设计方法。设计的控制器主要由开发板、控制电路和机器学习模块组成。机器学习模块利用特定的学习方式对网络被控对象进行监督,监督结果将被传送到控制电路进行多种学习行为的虚拟控制。开发板对虚拟控制结果进行接收,筛选出对网络被控对象的最优控制策略。机器学习模块对最优控制策略进行评价后,向网络被控对象实施稳定控制。实验结论证明,所设计的控制器可在维持对网络有效控制的同时,获取优异响应时间,响应能力较强,可较好地对设计目标进行实现。

英文摘要：

Because of interference of network illegal invasion and network data diversification, and the too long response time of the currently-designed network stability controller, a design method of network stable controller based on machine learn- ing is put forward. The controller is mainly composed of development board, control circuit and machine learning module. Ma- chine learning module using the specific learning way supervises the network controlled object. The supervision results are trans- mitted to the control circuit for virtual control of a variety of learning behaviors. The development board receives the virtual con- trol results, and screens out an optimal control strategy of the network controlled object. The machine learning module carries out stable control to the network controlled object after evaluation for the optimal control strategy. The experimental conclusions show that the controller can obtain the excellent response time while maintaining effective control to the network, has strong re- sponse ability, and can achieve the design goal well.