基于神经网络的MBP算法及应用

The MBP Arithmetic and Application Based On A Neural Network

利用神经网络的自我学习特性,以MBP算法的PID控制器取代常规PID调节器,介绍了MBP的算法推导及其控制器结构,通过仿真后得到两种控制算法的曲线比较,并应用于机载压电陀螺稳定平台中,可明显提高低频段的带宽,改善低速运动时的稳定性。

With the self-learning property of a neural network, we take advantage of the PID controller based on the MBP arithmetic to replace the rtaditional one. It also introduces the inference and the structure of the MBP arithmetic controller. By way of the simulation, we get the two curves compared with the two arithmetics. From the application to the aerial-laoding piezoelectric gyroscope stabilized platform, the results show that the bandwidth and the stability are improved.