针对脉宽调制(pulse width modulation，PWM)逆变电源控制系统，提出一种基于(diagonal recurrent neural network，DRNN)在线自整定PID控制和改进重复控制相结合的新型综合控制策略，给出PID参数在线自整定的控制算法和改进重复控制器的设计参数。利用改进重复控制改善系统的稳态性能,利用对角递归神经网络在线自整定PID控制提高系统的动态性能,既能克服常规控制逆变器波形跟踪性能差的不足，又能极大改善重复控制逆变器动态响应滞后的问题。实验结果表明，该综合控制策略能实现逆变器的快速动态响应和高精度稳态输出波形。

Aimed at the pulse width modulation (PWM) inverters control system, a integrated control strategy was put forward based on diagonal recurrent neural network (DRNN) adaptive PID and the modified repetitive control (MRC) technique, an algorithm of PID parameters self-tuning on line and parameters of MRC was presented. The dynamic behavior of the inverter system was enhanced by DRNN adaptive PID control , and the steady state characteristics of the system were also improved by MRC. Experiments show that the integrated control strategy can achieve high-dynamic response and near perfect output voltage waveform.