Research On Beidou/GNSS Cycle Slip Detection And Repair Method Based On LM-BP Neural Network

The high-precision Beidou /GNSS(Global Navigation Satellite System)plays an important role in development and construction process of China.In the carrier phase measurement of Global Navigation Satellite System(GNSS),if the cycle slip problem can not be found and dealt with in time,it will seriously affect the positioning result.Therefore,based on the in-depth study of cycle slip theory,this paper constructed a series of cycle slip detection and repair methods of Beidou /GNSS single frequency,dual frequency and triple frequency data based on LM-BP neural network.The main research contents of this paper are as follows:(1)The domestic and foreign research progress of cycle slip detection and repair theory was described in detail.By comparing different types of cycle slip detection and repair ideas and methods,the applicable conditions and effects of different cycle slip detection and repair methods were summarized.Based on the analysis of cycle slip theory and the application of neural network in data adjustment,this paper analyzed the related theory of cycle slip and the application of neural network in data adjustment,and expounded the necessity and theoretical basis of cycle slip detection and repair based on LM-BP neural network algorithm.(2)The common cycle slip detection and repair methods of single frequency data were analyzed,and a cycle slip detection and repair model based on LM-BP neural network is proposed.Specific idea: Firstly,LM-BP neural network was used to fit and predict the double difference sequence of satellite carrier phase,then the threshold was set to determine whether the residual sequence composed of predicted value and actual value had cycle slip,and finally the cycle slip value was separated and the carrier phase observation value was repaired.The method is validated by using the measured data of Beidou.The results show that this method can detect and repair cycle slips in static double-difference data and improve the overall timeliness of cycle slips detection and repair.(3)This paper analyzed the common cycle slip detection and repair methods of dual frequency data,studied the shortcomings of the traditional TurboEdit method,and proposed an improved TurboEdit cycle slip detection and repair method based on LMBP neural network,which makes use of the characteristics of LM-BP neural network that can quickly and accurately fit nonlinear data.It was combined with TurboEdit method to realize continuous small cycle slip detection and ensure the stable positioning accuracy.At the same time,LM-BP algorithm was used to process the carrier phase single difference observation,and a new cycle slip detection was constructed to realize the fast cycle slip separation.(4)The cycle slip detection and repair methods of three frequency data were analyzed.Based on LM-BP neural network algorithm,an improved LM-BP accelerated search method based on pseudo-distance phase combination combined with simultaneous ionospheric residual error combination cycle slip detection and repair method was proposed,and the method was simulated and verified by Beidou triple frequency data.Experimental results show that this method can detect and repair various types of cycle slips in three frequency data,and the speed is obviously improved.