论文部分内容阅读
针对在线参数辨识计算工作量大,造成难以实时给出参数估值的问题,利用Hopfield型网络的快速优化计算能力,通过对Hopfield网络改进,推出了一种全并行递推神经网络参数辨识方法,使计算量较传统的参数辨识方法大大减小。同时由于神经网络的互连作用,增强了辨识的鲁棒性,为实时给出参数估值提供了可靠的保障。
Aiming at the problem of large computational workload of on-line parameter identification and calculation, it is hard to estimate the parameters in real time. By using the Hopfield network’s ability of rapid optimization and calculation, an all-parallel recursive neural network parameter identification method is proposed by improving Hopfield network. So that the calculation of traditional parameter identification method is greatly reduced. At the same time, due to the interconnection of neural networks, the robustness of identification is enhanced, which provides a reliable guarantee for parameter estimation in real time.