Noise Reduction Of The Strapdown Inertial Navigation System Based On The Neural Network And Flying Parameters Solving

In the strapdown inertial navigation system,the data from accelerometer and gyroscopehave noise.The noise impacts the precision of the data and the precision of the computedvalue,and reduces the precision of the navigation.The algorithm of the neural network has agood ability to close to unfamiliar function.In advance,it has no necessary to know themathematical model.It reduces the error from the mathematical model.So it is better thantraditional filter method. Aimming at getting rid of the noise of the data of accelerometer andgyroscope in the strapdown inertial navigation system,this paper mainly studies the followingcontent:The common noise of the data from accelerometer and gyroscope is studied by Allanvariance.This paper uses Allan variance methods to analyze the data of accelerometer andgyroscope,and gets the common noise of the data of accelerometer and gyroscope. And it usesAllan variance methods to analyze estimate the data from the neural network.This paper has researched the effect of using BP neural network and RBF neural networkto handle the data from accelerometer:for handling the data from accelerometer,it analyzedthe effect of BP neural network and RBF neural network in time and precisionimportantly.According to the conclusion,this paper uses newrb function in RBF neuralnetwork to handle the data of accelerometer and gyroscope,and we can see that neuralnetwork has effectiveness in eliminating noise.This paper has researched a kind of the strapdown inertial navigation system made up bythree accelerometers and three gyroscopes.It uses four elements methods and Runge-Kuttaandmethods to simulat the motion state of aircraft.In the end,it uses the data of accelerometersand gyroscopes to gain the information of location and speed.