基于FFT的概率神经网络故障诊断模型

无人机的执行器和传感器系统受到材料与环境等诸多因素的影响, 容易发生各类故障, 严重时甚至会造成坠机, 因此实现无人机早期故障的有效诊断对预防飞行事故具有重要意义. 本文以六旋翼无人机Simulink模型作为研究对象, 针对飞行器电机和角速度传感器的早期故障, 提出了一种基于快速傅里叶变换（Fast Fourier Transform, FFT）的概率神经网络故障诊断模型. 首先, 在Simulink平台上对六旋翼无人机进行飞控模型的建立; 然后采用FFT对数据进行有效的时频分析; 最后基于MATLAB设计并建立概率神经网络模型, 利用FFT数据进行故障分类, 实现无人机的故障诊断.

Probabilistic Neural Network fault diagnosis model based on FFT

The actuator and sensor system of the UAV is affected by many factors such as materials and environment, the UAV is prone to various types of faults and even may crash in severe cases. Therefore, the effective diagnosis of the early fault of the UAV is of great significance in preventing flight accidents. The paper chooses the Simulink model of the six rotor UAV as the research object, a probabilistic neural network fault diagnosis model based on Fast Fourier transform (FFT) is proposed for the early fault of aircraft motor and angular velocity sensor. The flight control model of the six rotor UAV is established on the Simulink platform. Then the FFT is used to analyze the data in an effective time frequency analysis. Finally, the probabilistic neural network model is implemented with MATLAB, and the FFT data is used to classify the faults to realize fault diagnosis of UAV.