OLS算法在无人机传感器故障诊断中的应用

在故障诊断领域,神经网络故障诊断方法以其优良的特性正得到越来越广泛的应用。本文针对无人机的特点提出一种基于RBF神经网络故障诊断方法,通过建立神经网络预测器来实现无人机机载传感器的故障诊断,其中网络学习算法的选取将直接影响神经网络故障诊断的性能。正交最小二乘算法(OLS)以其在设定网络参数方面的优点常用来作为RBF神经网络学习算法。本文将介绍OLS算法的原理和实现步骤,通过VC 6.0编程实现OLS算法,并利用无人机机载传感器数据来验证OLS算法的有效性。

Application of OLS Arithmetic to Fault Diagnosis for Airborne Sensors in Unmanned Aerial Vehicles

In the field of sensor fault diagnosis,neural network is applied more and more widely because of its excellent characteristics.In this paper,a RBF neural network based fault diagnosis approach is proposed by the development of a neural network predictor to diagnose sensor faults for Unmanned Aerial Vehicles.The choice of network learning algorithms will directly affect the performances of the neural network predictor,In view of its advantages in the design of network parameters,the OLS arithmetic is often used as a neural-network learning algorithm.The OLS arithmetic principle and implementation using VC 6.0 program is presented,and the OLS arithmetic is validated by the use of data from airborne sensors for unmanned aerial vehicles.