Cycle slip detection and restoration method based on LSTM neural network

Abstract

The invention provides a cycle slip detection and restoration method based on an LSTM neural network. Regression prediction of feature data of a time sequence carrier phase measurement value is realized through an LSTM, and the detection and restoration of the cycle slip of a GNSS carrier phase measurement value are realized through the assistance of Doppler information. The method comprises the following steps: (1), acquiring a carrier phase measurement value and a Doppler value, and preprocessing the data; (2), designing an LSTM neural network used for predicting characteristic data of a carrier phase measurement value; (3), processing the data set by using an LSTM neural network; (4), detecting and repairing cycle slip by using an output result of the LSTM neural network; and (5), updating epoch information, and repeatedly executing the steps (1) to (4) until all epochs are processed; and finally obtaining a cycle slip detection result and a cycle slip restoration result. After verification, the method can effectively detect and repair cycle slips longer than 0.3 cycle.

Description

technical field [0001] The invention relates to a cycle slip detection and repair method based on an LSTM neural network, which belongs to the field of satellite navigation. Background technique [0002] The carrier phase measurement value is output by the GNSS receiver tracking loop carrier NCO, which plays a key role in the precise positioning process. Under the influence of factors such as receiver tracking loop loss, satellite signal interruption, low carrier-to-noise ratio, and high dynamics, the receiver will track and change circuits, resulting in discontinuous carrier phase measurement values, that is, "cycle slip". It will seriously affect the accuracy of precise positioning, so the detection and repair of cycle slips has become a key research issue in GNSS precise positioning data processing. Traditional cycle slip detection and repair methods mainly include high-order difference method, integral Doppler method, ionospheric residual method, polynomial fitting meth...

AI Technical Summary

Problems solved by technology

Under the influence of factors such as receiver tracking loop loss, satellite signal interruption, low carrier-to-noise ratio, and high dynamics, the receiver will track and change circuits, resulting in discontinuous carrier phase measurement values, that is, "cycle slip". It will seriously affect the accuracy of precise positioning, so the detection and repair of cycle slips has become a key research issue in GNSS precise positioning data processing

Traditional cycle slip detection and repair methods mainly include high-order difference method, integral Doppler method, ionospheric residual method, polynomial fitting method, etc., each of which has its own advantages and disadvantages. For example, the ionospheric residual method is not suitable for single-frequency data , the high-order difference method and the polynomial fitting method require a high sampling rate, and the integral Doppler method is difficult to detect small cycle slips, etc.

At present, there are not many literatures that apply neural network to cycle slip detection and repair, such as literature [1] Wu Dong, Hu Wusheng. New method of GPS cycle slip detection based on neural network [J]. Surveying and Mapping Engineering. 2008, 17( 6): 67-70. It applies BP neural network to cycle slip detection, by predicting the carrier phase measurement value, and comparing the predicted value to judge whether a cycle slip occurs, this method directly processes the carrier phase measurement value, resulting in The detection accuracy drops, and the BP neural network does not take into account the timing characteristics of the data; literature [2] Yi Tinghua, Li Hongnan, Yi Xiaodong, etc. GPS cycle slip detection and repair based on wavelet and neural network [J]. Sensing technology Acta Sinica. 2007, 20(4):897-902. It also uses BP neural network, and extracts wavelet features from carrier phase, and detects and repairs cycle slips by predicting features. Although this method uses wavelet methods The detection accuracy is improved, but there is also the problem of not considering the timing of the data

Images