人工神经网络思想优化EPA动态通信调度

Optimizing EPA Dynamic Communication Dispatching Performance by Utilizing the Thoughts of Artificial Neural Network

研究了工业以太网国际标准EPA的通信调度方法和性能,并分析了工业EPA应用层采用的令牌/生产者/消费者实时通信模型,指出该模型忽略了系统工作过程中周期性变量和非周期变量之间的相互影响。针对参量访问的优先级分配不够周全和灵活的问题,提出了利用人工神经网络的思想优化EPA实时动态通信调度,给出了一个新的智能实时通信调度应用模型。过程控制系统中EPA应用的仿真结果显示,该智能通信调度模型表现良好、优点明显。

The communication dispatching method and performance of International Standard EPA for industrial Ethernet is researched, and the token/producer/consumer real-time communication model of EPA application layer is analyzed. It is proposed that this model ignores the interaction between periodical and un-periodical variables, and the parameter access priorities have not allocated considerately and flexibly. By adopting the concept of artificial neural network, the EPA real-time dynamic communication dispatching is optimized, and the new intelligent real-time communication application model is given. The simulation result of EPA application in process control system indicates that this intelligent communication dispatching model works better and features obvious merits.