PID神经网络的VHDL实现

Realization of PID Neural Network Based on VHDL

PID神经网络模块是单变量或多变量非线性PID神经网络控制器的核心部分。从分析PID神经网络的设计原理入手,给出了PID神经网络实现的环形电路结构,采用自顶向下的设计方法,利用VHDL语言设计和实现了3层PID神经网络模块。仿真结果表明该模块功能完全正确。综合结果表明,该网络的实现仅使用了175个LE和9个嵌入式硬件乘法器,最高时钟频率可达116 MHz。

The core of nonlinear PID neural network controller with one variable and multi - variable is PID neural network module. The paper analyses the design theory of PID neural network,provids a ring circuit configuration to realize PID neural network module,designs and realizes 3 -layer PID neural network module with top- down design methodology based on VHDL. The simulation suggests the module operates well. The synthesizing result shows that the realization of network only uses 175 LEs and 9 embedded hardware multipliers,the maximal frequency can reach 116 MHz.