Font Size: a A A

AUV Dead-Reckoning Navigation Based On Neural Network

Posted on:2018-04-28Degree:MasterType:Thesis
Country:ChinaCandidate:Y X XieFull Text:PDF
GTID:2348330515474042Subject:Computer technology
Abstract/Summary:PDF Full Text Request
AUV is one of the most important tools to explore the ocean.Since the accuracy of navigation system determines whether the AUV can safely work and return,AUV navigation system is required more and more accurate.Currently,the commonly used ways of navigation are Dead-Reckoning(DR),Acoustic navigation,Geophysical navigation,etc.Acoustic navigation methods(LBL,SBL,USBL)use sonar array to estimate the AUV's range and bearing,and it has a great accuracy.However,this method's disadvantage lies in its limited distance?Dead-reckoning(DR)is one of the most commonly used methods of AUV navigation,which primarily uses the known initial position,speed and direction to calculate the next location.The traditional DR relies on the inertial sensors such as gyroscope and accelerometer,gyroscope is used to estimate the direction of AUV relative to reference coordinate system,and the accelerometer is used to estimate the distance of navigation.A major challenge for DR navigation is from measurement error of the inertial sensors(gyroscope,accelerometer,etc.),especially when the AUV is near or at the ocean surface,it's motion will be affected by the ocean waves,and the pitch angle of AUV will change rapidly with the waves.This rapid change and offset will lead a great noise to the direction measured by gyroscopes,and then lead a large error to the DR navigation.In this paper,we propose an algorithm,combined with acoustic navigation and dead reckoning method to positon the AUV.Firstly,we locate the AUV accurately in a short period of time in the beginning of the navigation by acoustic navigation.Secondly we use the positon information and sensor data to estimate the pitch angle of AUV,and then use the neural network to explore the time-varying relationship between acceleration measurement and orientation measurement.Thus,we can use acceleration to calculate pitch angle through this time-varying relationship.Thus in the future of the navigation,we can complete AUV's DR navigation only use the acceleration without gyroscope.Most importantly,we can improve the accuracy of AUV navigation through avoiding DR errors caused by gyroscope noise at the sea surface.
Keywords/Search Tags:dead-reckoning, acoustic navigation, neural networks, accelerometer
PDF Full Text Request
Related items