论文部分内容阅读
本文首次提出一种传感器故障检测和信号恢复的改进鲁棒自联想神经网络新方法。文中阐述了改进鲁棒自联想神经网络的结构和算法，总结和归纳了传感器的六类故障模式.仿真了这些传感器故障模式的信号表现形式，并采取改进鲁棒自联想神经网络检测和恢复此六类传感器故障。本方法具有易于实时实现，结构简单的优点，计算机实验表明本方法是行之有效的。
For the first time, this paper presents a new robust self-associated neural network method for sensor fault detection and signal restoration. In this paper, the structure and algorithm of improved robust ARF neural network are summarized, and six kinds of fault modes of the sensor are summarized and summarized.The signal manifestation of these sensor failure modes is simulated and the improved Robust ARN neural network is used to detect and recover this Six types of sensor failure. The method has the advantages of easy real-time implementation and simple structure, and the computer experiment shows that the method is effective.