CN-121995419-A - Low-orbit satellite-borne GPS cycle slip detection method

CN121995419ACN 121995419 ACN121995419 ACN 121995419ACN-121995419-A

Abstract

The application is suitable for the field of satellite navigation signal processing, and provides a low-orbit satellite-borne GPS cycle slip detection method which comprises the steps of obtaining observation value data of a low-orbit satellite-borne GPS, calculating MW combined widelane ambiguity values of the satellite-borne GPS, optimizing ICEEMDAN algorithm set average times and noise standard deviation by adopting IAO algorithm, decomposing the widelane ambiguity values into a plurality of intrinsic mode function IMF components and a residual error component by adopting optimized ICEEMDAN algorithm, calculating permutation entropy values of all IMF components by adopting permutation entropy algorithm, screening out IMF components needing to be subjected to denoising, carrying out denoising on the IMF components needing to be subjected to denoising by combining a wavelet threshold denoising method, carrying out signal reconstruction on the IMF components subjected to denoising by adopting the wavelet threshold denoising method and the IMF components not needing to be subjected to denoising, obtaining denoised widelane ambiguity, calculating difference value of denoised widelane ambiguity between a current epoch and an adjacent epoch, and judging whether cycle slip exists according to a preset threshold. The application obviously improves the cycle slip detection accuracy and reliability of the low-orbit satellite in the high-dynamic and strong-noise environment, and reduces the misjudgment rate and the missed judgment rate.

Inventors

JI YUANFA
GAO LING
SUN XIYAN
JIA QIANZI
LIANG WEIBIN
LIU XINYI
FU WENTAO
LIANG WENBIN
LI JIE

Assignees

桂林电子科技大学

Dates

Publication Date: 20260508
Application Date: 20251224

Claims (10)

1. The low-orbit satellite-borne GPS cycle slip detection method is characterized by comprising the following steps of: S101, acquiring observation value data of a low-orbit satellite-borne GPS, wherein the observation value data comprises carrier phase observation values and pseudo-range observation values corresponding to two different carrier frequencies, and MW combination widelane ambiguity values of the satellite-borne GPS are calculated according to the two different carrier frequencies and the corresponding carrier phase observation values and pseudo-range observation values; S102, optimizing ICEEMDAN the aggregate average times and the noise standard deviation of an algorithm by adopting an IAO algorithm so as to minimize the fitness function; s103, decomposing the wide-lane ambiguity value into a plurality of intrinsic mode functions IMF components and a residual error component by adopting an optimized ICEEMDAN algorithm; S104, calculating the arrangement entropy values of all IMF components by adopting an arrangement entropy algorithm, screening out IMF components needing to be subjected to denoising, and denoising the IMF components needing to be subjected to denoising by combining a wavelet threshold denoising method; S105, carrying out signal reconstruction on the IMF component subjected to denoising treatment by a wavelet threshold denoising method and the IMF component not subjected to denoising treatment to obtain denoised widelane ambiguity, calculating the difference value of denoised widelane ambiguity between the current epoch and the adjacent epoch, and judging whether cycle slip exists or not according to a preset threshold.
2. The method for detecting satellite-borne GPS cycle slip of low-orbit satellite according to claim 1, wherein S101 further comprises preprocessing the observed value data and eliminating abnormal values; The calculation formula for calculating the MW combined widelane ambiguity value of the satellite-borne GPS according to the two different carrier frequencies, the corresponding carrier phase observed values and the pseudo-range observed values is as follows: Wherein, the For the MW combined widelane ambiguity values, Respectively two different carrier frequencies, Respectively carrier frequencies The corresponding carrier phase observations are made, Respectively carrier frequencies The corresponding pseudorange observations are then processed, For the combined wavelength of the MW, Is the propagation velocity of the electromagnetic wave in vacuum.
3. The method for detecting satellite-borne GPS cycle slip of low-orbit satellite according to claim 1, wherein S102 comprises the steps of: Performing self-adaptive optimization on key parameters of ICEEMDAN algorithm by adopting an IAO algorithm, determining iteration steps, population scale and parameter searching range of the IAO algorithm, taking the set average number NR and noise standard deviation NStd of the ICEEMDAN algorithm as optimization variables, and minimizing permutation entropy as a fitness function; Decomposing an original signal into a plurality of IMF components and residual items by using preset ICEEMDAN algorithm parameters, and carrying out fitness evaluation on each IMF component one by one to obtain a fitness value of each IMF component; The IAO algorithm is used for simulating four strategies of the hawk hunting behavior, namely high-altitude flying expansion search, low-altitude flying shrinkage search, low-flying grabbing and walking capturing, the average number NR and the noise standard deviation NStd of the collection are iteratively updated through the strategies to minimize the fitness function, the optimal individual is reserved based on the elite selection strategy, finally, whether the algorithm is converged or not is judged, iteration is continued, and when the fitness value change of the IMF component is smaller than a preset threshold value or reaches the maximum iteration number, the algorithm is considered to be converged and the parameters of the optimized ICEEMDAN algorithm are output as the global optimal solution.
4. The method for detecting satellite-borne GPS cycle slip of low-orbit satellite according to claim 3, wherein the method for obtaining the fitness value of each IMF component by decomposing the original signal into a plurality of IMF components and residual items by using the parameters of a preset ICEEMDAN algorithm comprises the following steps: Inputting MW combined wide lane ambiguity values as signal values, initializing ICEEMDAN algorithm parameters and IAO algorithm parameters, and enabling an adaptability function to be minimum in permutation entropy; decomposing the signal value by using the current ICEEMDAN algorithm set average times NR and the noise standard deviation NStd to obtain a plurality of IMF components and residual items; and calculating the fitness value of each IMF component according to the fitness function.
5. The low-orbit satellite-borne GPS cycle slip detection method according to claim 1, wherein the IAO algorithm is an improved hawk optimization algorithm, and based on the hawk optimization algorithm, the IAO algorithm mainly performs two-point optimization, namely, adopting a Tent chaotic mapping method to complete population initialization; The method for completing population initialization by adopting the Tent chaotic mapping method specifically comprises the following steps: Generating an initial population by adopting Tent chaotic mapping For the j-th dimensional variable of the i-th individual, the iterative formula of the Tent map is: Wherein, the K=2 is a chaotic control parameter, and the system is in a complete chaotic state at the moment; Mapping the chaotic variable to the actual search space is: Wherein, the For the j-th dimensional position of the i-th individual, And The lower bound and the upper bound of the j-th dimension are respectively adopted, and N is the population scale; the self-adaptive weight updating global optimal solution specifically comprises the following steps: the adaptive weight function is: wherein T is the current iteration number, T is the maximum iteration number, In order to adapt the weight coefficient of the model, Nonlinear increment along with the iteration times; calculating fitness of each body ; Self-adaptive adjustment is carried out on the global optimal position to obtain a global optimal solution And an optimal fitness value : If it is Then , Wherein, the For the globally optimal location obtained before the t-th iteration, For the position adjusted by the adaptive weight, Is an objective function; For the exploration phase , A first strategy, high altitude soaring expansion search, namely high altitude vertical diving, hawk recognition of hunting areas, and selection of the best hunting area through vertical bending high altitude soaring; Wherein, the For the globally optimal location obtained before the t-th iteration, Representing the mean value of the positions of the current solutions of the connection at the t-th iteration, A random number in the interval of [0,1], For a time decay factor, decreasing with iteration to avoid premature convergence, The position of the ith individual at the t-th iteration; the second strategy is that low-altitude soaring shrinkage search, namely contour flight and Levy flight, hawk is explored in a divergent search space through contour flight of short gliding attack, and global searching capability is enhanced by combining with a Levy flight mechanism; Wherein, the The method is a Levy flight step length vector in a D-dimensional space; for a randomly selected individual location, , Searching track parameter vectors for the spiral; the Levy flight step size is calculated as follows: Wherein, the =1.5 Is the stability index, For Gamma function, u and v are random vectors obeying normal distribution, Representation of Obeying mean value 0, variance 0 Is a normal distribution of (2); Representation of Obeying a standard normal distribution with a mean value of 0 and a variance of 1; the spiral search trajectory parameters are as follows: Wherein, the For the initial radius to be the same, In order to be a coefficient of the spiral growth, In order to be able to achieve an angular velocity, For the initial phase position, And Representing the spiral shape in the search, Is the first The polar diameter of each point is equal to the polar diameter of each point, Is the first Polar angle of each point; for the development stage , The third strategy is low fly grabbing, namely, after the specific position of the target is successfully locked and necessary landing and attack deployment work are completed, the hawk executes a vertical descent strategy; Wherein, the The position mean value of the population at the t-th iteration; To develop parameters, controlling the amplitude of approach to global optimum; To develop parameters, the random walk range is limited, UB and LB are the upper and lower bounds of the search space, respectively; random numbers in the interval of [0,1 ]; the fourth strategy is walking capture, the hawk is used for final attack by hiking and grasping the prey and using a mass function, and the formula is as follows: Wherein QF is a quality function for balancing the search strategy at the t-th iteration, As a parameter of the rate of flight, Is a motion direction parameter; after each new solution is generated, greedy selection is performed based on the fitness value: Wherein, the For passing through strategy Or (b) The new solution to be generated is provided, The mechanism ensures that the population adaptability is improved monotonically.
6. The method for detecting satellite-borne GPS cycle slip of low-orbit satellite according to claim 1, wherein S103 comprises the steps of: Decomposing the widelane ambiguity values by adopting an optimized ICEEMDAN algorithm, and adaptively adding Gaussian white noise to the widelane ambiguity values in each screening process to obtain non-stable widelane ambiguity values The method comprises the steps of dividing the IMF components into a plurality of IMF components and a residual component, arranging the IMF components from high to low in frequency to extract local features of the signals and oscillation components with different frequencies, wherein the residual component represents the overall trend or low frequency component of the signals.
7. The method for detecting cycle slip of GPS on board a low-earth satellite according to claim 6, wherein said adaptive addition of Gaussian white noise to said widelane ambiguity values results in a non-stationary widelane ambiguity values The decomposition into a plurality of IMF components and a residual component specifically comprises the steps of: the formula for adaptively adding Gaussian white noise to the widelane ambiguity value is as follows: Wherein, the For the original signal, i.e. the widelane ambiguity value, For the jth gaussian white noise realization, Is the first IMF of the white noise, The noise coefficient is N, and the integration times are N; Calculating a local mean and obtaining a first residual component The formula is as follows: Wherein, the Representing a local mean operation; Extracting first IMF component The formula is as follows: adding the q-th stage self-adaptive Gaussian white noise, and adopting the following formula: Wherein, the For the noise figure of the q-1 stage, The q-th IMF component of white noise; Calculation of the qth residual component The formula is as follows: recursive formula of use Extraction of the q-th IMF component ; Extracting until the residual component After the termination condition is met and the decomposition is completed, the original signal is decomposed into q IMF components and a residual component, and the final decomposition result formula is as follows: Where y is the original widelane ambiguity value, free of artificially added noise, For the q-th IMF component, For the final residual component, K is the total number of IMF components.
8. The method for detecting satellite-borne GPS cycle slip of low-orbit satellite according to claim 1, wherein S104 comprises the steps of: Calculating the permutation entropy values of all IMF components by adopting a permutation entropy algorithm, and judging that the IMF components need to be subjected to denoising treatment if the permutation entropy values are larger than a preset denoising threshold value; for IMF components needing denoising treatment, a wavelet threshold denoising method is adopted to eliminate noise; the method for calculating the permutation entropy of all IMF components by adopting the permutation entropy algorithm comprises the following steps of: For a signal sequence X with the length of N, namely, q IMF components obtained by decomposing a widelane ambiguity value by adopting an optimized ICEEMDAN algorithm, carrying out phase space reconstruction by adopting a Takens delay embedding theorem, and carrying out phase space reconstruction according to a preset embedding dimension m and delay time Embedding dimension=3, delay time=1, and obtaining matrix Y as: Wherein, the For the phase space reconstruction matrix, each row of the matrix represents one m-dimensional reconstruction vector, and the total number of the m-dimensional reconstruction vectors is K, and the number of the reconstruction vectors is ; The elements within each reconstruction vector are arranged in ascending order as follows: Wherein, the A column representing elements within the reconstruction vector; after the elements in each reconstruction vector are arranged in ascending order, a symbol sequence is formed by the position indexes before the arrangement, and the symbol sequence can represent each reconstruction component as follows: Calculating the probability of each permutation mode by adopting a frequency estimation method: based on probability distribution of the arrangement mode, calculating an arrangement entropy value according to definition of Shannon information entropy, wherein the arrangement entropy value is as follows: to facilitate comparison and normalization at different embedding dimensions, normalized permutation entropy values are defined: According to a preset denoising threshold value =0.7, Then determine The IMF component of (c) is subjected to denoising processing.
9. The method for detecting satellite-borne GPS cycle slip of low-orbit satellite according to claim 8, wherein the wavelet threshold denoising method is to perform threshold processing on wavelet coefficients to realize signal denoising, and adopts db4 as a wavelet base, wherein the number of wavelet decomposition layers is 3; The method comprises the steps of adopting VisuShrink as a threshold value and adopting a soft threshold value as a denoising strategy, shrinking wavelet coefficients according to a preset threshold value, and reconstructing signals by using the updated coefficients so as to obtain a denoising result; The denoising formula is as follows: Wherein, the Represents the kth wavelet coefficient of the jth layer, T is VisuShrink threshold, when When the wavelet coefficients are considered to contain useful signals, the wavelet coefficients are preserved but the T units are shrunk toward zero when When the wavelet coefficients are considered to be mainly noise, the wavelet coefficients are set to zero.
10. The method for detecting cycle slip of low-orbit satellite-borne GPS according to claim 1, wherein determining whether a cycle slip exists according to a preset threshold is specifically: Wherein, the Is the standard deviation of noise, when the current epoch i and the previous epoch De-noised widelane ambiguity between If the difference of the following epoch exceeds the preset threshold With the current epoch De-noised widelane ambiguity of (2) If the difference of (2) is smaller than 0.1, it is determined that cycle slip occurs, otherwise it is determined that the difference is rough.

Description

Low-orbit satellite-borne GPS cycle slip detection method Technical Field The application belongs to the field of satellite navigation signal processing, and particularly relates to a low-orbit satellite-borne GPS cycle slip detection method. Background Low Earth Orbit (LEO) satellite-borne GPS technology has become an important way to achieve high-precision Orbit determination and real-time navigation and positioning. Compared with GPS observation acquired by a ground station, the low-orbit satellite has cycle slip phenomenon in carrier phase observation due to high-speed motion and complex space environment influence, so that phase ambiguity resolution and orbit precision are obviously influenced. Cycle slip refers to the whole phase mutation caused by factors such as signal interruption, multipath effect, receiver noise or high dynamic condition in the carrier phase observation process. To maintain continuity of phase observations, cycle slips must be accurately detected and repaired. Therefore, the research of the satellite-borne GPS cycle slip detection method suitable for the dynamic characteristics of the low-orbit satellites is significant. With the deep application of the global navigation satellite system (Global Navigation SATELLITE SYSTEM, GNSS) in LEO precise orbit determination, researchers at home and abroad propose various cycle slip detection strategies for dual-frequency and multi-frequency GPS signals. Currently common methods include a high-order differential method, a pseudo-range-phase combination method, a geometry-independent combination method, an ionosphere residual method, a wavelet transform method, a MW (Melbourne-Wunnema) combination method, and the like. The MW combination method is a method for performing cycle slip detection by using pseudo-range observation values and carrier phase observation values of a double-frequency signal. The method completes the detection of cycle slip by constructing a wide lane combination of carrier phase values and a narrow lane combination of pseudo-range values. The MW combination method can effectively eliminate the influence of geometric distance and ionosphere delay, and has simple structure and wide application. However, because it relies on pseudo-range observation, the result is highly sensitive to pseudo-range observation noise, if the observation noise is large or the signal quality is poor, a small cycle slip of 1-2 cycles may be covered by noise, resulting in erroneous judgment and missed detection, and it is difficult to meet the requirement of low-orbit satellite real-time precise orbit determination. Disclosure of Invention The application aims to provide a satellite-borne GPS cycle slip detection method for a low-orbit satellite, which aims to solve the problems that the MW combination method depends on pseudo-range observation, the result is highly sensitive to pseudo-range observation noise, if the observation noise is large or the signal quality is poor, small cycle slips of 1-2 cycles can be covered by noise, misjudgment and omission are caused, and the real-time precision orbit determination requirement of the low-orbit satellite is difficult to meet. The application provides a low-orbit satellite-borne GPS cycle slip detection method, which comprises the following steps: S101, acquiring observation value data of a low-orbit satellite-borne GPS, wherein the observation value data comprises carrier phase observation values and pseudo-range observation values corresponding to two different carrier frequencies, and MW combination widelane ambiguity values of the satellite-borne GPS are calculated according to the two different carrier frequencies and the corresponding carrier phase observation values and pseudo-range observation values; S102, optimizing ICEEMDAN the aggregate average times and the noise standard deviation of an algorithm by adopting an IAO algorithm so as to minimize the fitness function; s103, decomposing the wide-lane ambiguity value into a plurality of intrinsic mode functions IMF components and a residual error component by adopting an optimized ICEEMDAN algorithm; S104, calculating the arrangement entropy values of all IMF components by adopting an arrangement entropy algorithm, screening out IMF components needing to be subjected to denoising, and denoising the IMF components needing to be subjected to denoising by combining a wavelet threshold denoising method; S105, carrying out signal reconstruction on the IMF component subjected to denoising treatment by a wavelet threshold denoising method and the IMF component not subjected to denoising treatment to obtain denoised widelane ambiguity, calculating the difference value of denoised widelane ambiguity between the current epoch and the adjacent epoch, and judging whether cycle slip exists or not according to a preset threshold. In the application, key parameters (the number of times of collection average and the standard deviation of noise) are adaptive