论文部分内容阅读
针对BP算法在进行系统辨识时存在着速度慢、非平滑内插、受噪声影响很大、逼近精度不高，特别是对强干扰不具有鲁棒性等问题，提出一种BP网络的鲁棒算法，该算法直接利用样本点对样本点的分布特征进行估计，并采用带有损失因子的与误差分布有关的二次型能量函数，并用于动态系统辨识，仿真结果表明了算法不仅对白噪声具有鲁棒性，而且对强干扰也具有鲁棒性．
Aimed at the problems of slow and non-smooth interpolation of BP algorithm in system identification, which is greatly affected by noise, low approximation accuracy and especially not robust to strong interference, a BP network robust The algorithm directly uses the sample points to estimate the distribution characteristics of the sample points, and uses the quadratic energy function with loss factor related to the error distribution for the dynamic system identification. The simulation results show that the algorithm not only has the function of white noise Robust, but also robust to strong interference.