应用人工神经网络构建的VPDN用户故障检测系统

VPDN Users Fault Measuring System Based on Artificial Neural Network

讨论一个应用人工神经网络设计的CDMA网络VPDN数据用户故障检测系统。该系统应用反向传播神经网络,利用理想状况下的用户数据训练网络,而用带有随机性的模拟实网数据检测故障判断的准确性。计算机仿真结果表明在噪声的干扰下神经网络基本可以给出具有参考价值的判断。通过实验充分证明利用神经网络完成故障判断的可行性,而设计、分析、检验等多个研究步骤为神经网络的此类应用提供了可供参考的蓝本。

This paper discusses a malfunction estimation system based on artificial neural network for COMA VPDN data subscribers. This system applies back propagation neural network, uses network data in ideal conditions to train the neural net-work,and then utilizes simulated network data with randomness to verify the accuracy of the estimation. Simulation results suggest that the neural network can give valuable estimations even under the perturbation of noise. Investigations in this paper confirm the feasibility of using neural network to detect malfunction, and design, analysis, verification and other research procedures provide us with a prototype for similar neural network applications.