A method, device and terminal for improving RTK integer ambiguity fixing rate

Abstract

A method, device and terminal for improving the fixing rate of RTK integer ambiguity are provided. The method introduces a double reliability check and a fixed solution degradation multiplexing mechanism. Specifically, after fixing the wide-lane ambiguity, a double check of the residual and ADOP (ambiguity dilution of precision) is performed simultaneously. The high-precision wide-lane fixed solution obtained through the check is used as a constraint condition to participate in the subsequent floating-point solution filtering. Thus, the pollution of the state space by the wrong wide-lane solution is eliminated from the source, and the purity of the narrow-lane ambiguity solution environment is ensured. When the narrow-lane ambiguity fixing fails, a degradation multiplexing strategy is started. The wide-lane fixed solution of the current epoch is evaluated in real time to determine whether it meets the preset accuracy threshold. If yes, the wide-lane fixed solution is directly output as the final positioning result. The present application not only realizes the effective use of the wide-lane ambiguity integer solution, but also provides a reliable suboptimal solution with an accuracy between the narrow-lane fixed solution and the floating-point solution for the user in a complex environment.

Description

Technical Field

[0001] This invention belongs to the field of satellite navigation and positioning technology. Specifically, this invention relates to a method for improving the fixation rate of RTK integer ambiguity (i.e., the number of missing integer cycles in carrier phase measurement). More specifically, this invention relates to a method, apparatus, and terminal for improving the fixation rate of RTK integer ambiguity. This invention significantly broadens the usable boundary of RTK technology in complex observation environments while ensuring the absolute reliability of positioning results, and effectively improves the overall fixation rate of the system. Background Technology

[0002] The key to achieving centimeter-level positioning in real-time kinematic (RTK) technology lies in the accurate solution of integer ambiguity in carrier phase observations. However, in complex scenarios such as urban canyons and tree-lined roads, multipath effects and signal obstruction severely impact observation quality, making it difficult to fix integer ambiguity and degrading positioning accuracy to decimeter or even meter-level floating-point solutions. To address this challenge, the industry generally adopts a step-by-step fixing strategy combining wide-lane and narrow-lane ambiguities. In complex scenarios, the wider-lane ambiguity, with its longer wavelength (approximately 86 cm) and easier fixability, is used to assist in solving the narrow-lane ambiguity.

[0003] In existing wide-lane ambiguity fixing techniques, various patents have proposed different optimization schemes. For example, CN121634171A (A Wide-Lane Ambiguity Fixing Method) introduces external ionospheric correction parameters to correct the combined ionospheric error of wide-lane under long baselines, thereby improving the accuracy of the long-baseline wide-lane fixing solution. In addition, CN115685272A (A Method for Determining Ambiguity Fixing Solutions) proposes a hierarchical fixing strategy for wide-lane combined distribution, aiming to improve the success rate of wide-lane fixing in PPP-AR.

[0004] However, the method mentioned in CN121634171A requires the introduction of external corrections and is only applicable to long baseline scenarios, not to short baseline RTK scenarios. While the wide-lane fixing strategy used in PPP-AR in CN115685272A can improve the success rate of wide-lane fixing, it cannot be directly used in RTK.

[0005] Furthermore, existing technologies for improving RTK fixation rates using wide-lane ambiguity suffer from two major drawbacks. First, they lack effective verification of the reliability of the wide-lane fixation solution itself. Conventional methods, after obtaining the wide-lane fixation solution, use it as the precise observation equation to constrain the narrow-lane floating-point ambiguity. If the wide-lane solution itself is incorrectly fixed due to residual errors, these errors will be introduced into the filtering process, polluting the entire state space. This not only fails to assist narrow-lane fixation but may also cause filter divergence, missing potentially successful positioning epochs. Second, they fail to fully utilize the accuracy value of the wide-lane fixation solution. In epochs where narrow-lane ambiguity cannot be fixed, existing technologies typically abandon or downgrade to floating-point solutions, ignoring the fact that the wide-lane fixation solution may still possess a much higher positioning accuracy than the floating-point solution, resulting in a waste of observation resources. Summary of the Invention

[0006] To address the aforementioned shortcomings, this invention proposes a method, apparatus, and terminal for improving the integer ambiguity fixation rate in RTK. Its core innovation lies in the introduction of a "dual reliability check" and a "fixed solution degradation reuse" mechanism.

[0007] Specifically, after fixing the wide-lane ambiguity, this invention simultaneously performs dual checks on its residuals (the difference between the fixed observed values ​​and the model predictions) and ADOP (Ambiguity Dilution of Precision). Only high-confidence wide-lane fixed solutions that pass the checks are allowed to participate as constraints in subsequent floating-point solution filtering. This mechanism eliminates the pollution of the state space by erroneous wide-lane solutions from the source, ensuring the purity of the narrow-lane ambiguity solution environment.

[0008] Furthermore, when narrow-lane ambiguity fixing fails, this invention does not simply revert to a floating-point solution, but instead initiates a degradation reuse strategy: it evaluates in real time whether the wide-lane fixing solution of the current epoch meets a preset accuracy threshold. If it does, the wide-lane fixing solution is directly output as the final positioning result. This strategy not only achieves effective utilization of wide-lane ambiguity integer solutions, but also provides users with a reliable suboptimal solution with accuracy between the narrow-lane fixing solution and the floating-point solution in complex environments.

[0009] In summary, this invention ensures the reliability of fixed solutions through a rigorous verification mechanism, and through an innovative degradation and reuse strategy, significantly expands the usability of RTK technology in complex observation environments while ensuring the absolute reliability of the positioning results, effectively improving the overall fixation rate of the system.

[0010] This invention aims to address the technical bottlenecks of existing RTK technology, such as low integer ambiguity fixation rate and susceptibility to erroneous solution contamination in complex dynamic environments. By introducing a dual reliability check mechanism, this invention can effectively identify and eliminate erroneously fixed wide-lane ambiguities, preventing errors from contaminating the filter state space at the source and ensuring the purity and reliability of narrow-lane ambiguity solution.

[0011] Meanwhile, this invention innovatively proposes a fixed solution degradation and reuse strategy. When fixing in narrow alleys fails, it fully utilizes high-precision fixed solutions in wide alleys as suboptimal outputs, avoiding the waste of high-precision observation resources. Ultimately, this invention significantly improves the availability and overall fixation rate of high-precision positioning for RTK terminals in complex scenarios such as urban canyons and vegetation obstruction, while ensuring the absolute reliability of the positioning results.

[0012] Principle of this invention:

[0013] The application scenario of this invention is to improve the fixation rate of RTK in occluded scenarios such as tree-lined areas, urban canyons, and under elevated roads by utilizing wide-lane fixed solutions. Specifically, the steps involve using the precision-checked wide-lane integer ambiguity to constrain the RTK floating-point ambiguity, improving the precision of the floating-point ambiguity and thus increasing the floating-point ambiguity fixation rate. Simultaneously, when floating-point ambiguity fixation fails, the wide-lane fixed solution is output as the RTK fixed solution, achieving fixed solution degradation and reuse, thus ensuring the fixation rate. In this process, to ensure the precision of the wide-lane fixed solution, a common method is to increase the threshold of the Ratio test, for example, setting the Ratio threshold to 5-10. However, since wide-lane combinations amplify observation errors by 5-6 times, the Ratio test alone cannot guarantee the precision of the wide-lane fixed solution. This invention introduces residual and ADOP tests on top of the Ratio test to ensure that high-precision wide-lane results are used for subsequent floating-point ambiguity fixation and fixed solution degradation and reuse.

[0014] Therefore, a "dual reliability check" and a "fixed solution degradation reuse" mechanism are introduced. Specifically, after fixing the wide-lane ambiguity, wide-lane observations with longer wavelengths (e.g., about 86 cm for GPS) and obvious integer characteristics are formed by combining the L1 and L2 carrier phases, making them easy to fix quickly. Simultaneously, a dual check of residuals and ADOP is performed. ADOP (Ambiguity Dilution of Precision) is a dimensionless index that measures the accuracy of the ambiguity floating-point solution, reflecting the geometric strength of the ambiguity search space; the smaller the ADOP value, the more accurate the ambiguity floating-point solution and the higher the fixation success rate. The high-confidence wide-lane fixed solution, verified through the check, is used as a constraint condition in subsequent floating-point solution filtering, thereby preventing erroneous wide-lane solutions from contaminating the state space at the source and ensuring the purity of the narrow-lane ambiguity solution environment. Narrow-lane ambiguity typically refers to ambiguity in L1 or ionospheric-free combinations, with short wavelengths (e.g., 19 cm), and is highly susceptible to noise and atmospheric errors, making it difficult to directly fix integer solutions. When narrow-lane ambiguity fixation fails, a fixed solution degradation and reuse strategy is initiated: real-time evaluation is performed to determine if the current epoch's wide-lane fixed solution meets a preset accuracy threshold; if so, the wide-lane fixed solution is directly output as the final positioning result. This invention not only achieves effective utilization of wide-lane ambiguity integer solutions but also provides users with a reliable suboptimal solution in complex environments, with accuracy between narrow-lane fixed solutions and floating-point solutions.

[0015] The technical solution of this invention for improving the integer ambiguity fixation rate of RTK is as follows:

[0016] A method for improving the integer ambiguity fixation rate in RTK, applicable to complex dynamic environments such as urban canyons and vegetation shading, includes the following steps:

[0017] S1: Based on real-time satellite signal quality and observation noise level, pre-screen effective single-difference ambiguities, remove satellites that are severely polluted by multipath effects or signal attenuation, and obtain double-difference wide-lane ambiguities that reach the preset upper limit, so as to provide a clean ambiguity candidate set for subsequent fixing.

[0018] S2: Construct the double-difference wide-lane ambiguity parameter matrix:

[0019] S2-1. After completing the single-difference ambiguity screening in S1, construct the double-difference wide-lane ambiguity parameter matrix and its corresponding covariance matrix;

[0020] S2-2. The construction process of the double-difference wide-lane ambiguity correlation matrix includes the following two matrix operations:

[0021] Step 1: Extract the double-difference ambiguity and its covariance matrix from the single-difference ambiguity in the state vector. Assume that the current filtered state vector contains... A single-difference ambiguity parameter, denoted as a vector. Its corresponding covariance matrix is Based on the selected double-difference ambiguity information, a dimension is constructed as follows: Double difference operator matrix Based on double difference operator matrix Calculate the double-difference ambiguity vector Double-difference ambiguity covariance matrix And the cross-covariance matrix of double-difference ambiguity and position parameters ;

[0022] Step 2: Construct the double-difference wide-lane ambiguity and its covariance matrix by linear combination of double-difference ambiguities. Assume that based on the selected double-difference wide-lane ambiguity combinations, such as L1 and L2, L1 and L5, L2 and L5, the dimension is... Wide lane combination operator matrix In particular, for situations where the number of visible satellites is small in obstructed scenarios, when three-frequency data is available, the number of double-difference wide-lane ambiguities is increased by combining multiple frequencies to ensure the dimensionality requirement of LAMBDA search, and the ultra-wide-lane wavelength characteristics formed by the combination of L2 and L5 are used to reduce the difficulty of ambiguity fixation.

[0023] S3: Fixed ambiguity of double-difference wide-lane:

[0024] S3-1. Use the LAMBDA method to fix ambiguity;

[0025] First, the ADOP value is calculated. ADOP measures the precision and geometric strength of the ambiguity floating-point solution; the smaller the value, the more accurate the floating-point solution, the flatter the search space, and the more reliable the fixation. In this invention, the empirical threshold for ADOP is set to 1.8.

[0026] At this point, the covariance matrix of the double-difference wide-lane ambiguity... Perform LDL decomposition:

[0027]

[0028] in, It is a unit lower triangular matrix with diagonal elements all being 1. It is a diagonal matrix, and the elements on its diagonal are represented as:

[0029]

[0030] This element is also a matrix. Approximation of eigenvalues. Based on the properties of matrix decomposition, determinant Represented as:

[0031]

[0032] because If it is a unit lower triangular matrix, then the determinant is... ,therefore That is, matrix Multiplication of elements along the main diagonal:

[0033]

[0034] The final result of ADOP is:

[0035]

[0036] in, This indicates the number of double-difference wide-lane ambiguities. Representation matrix The main diagonal element;

[0037] S3-2. Calculate the success rate of a fixed solution:

[0038] The success rate of a fixed solution reflects the current covariance matrix. Theoretically, the probability of successfully fixing the ambiguity is calculated using the following formula:

[0039]

[0040] Among them, the function The cumulative distribution function represents the standard normal distribution. After the LAMBDA method is executed, the RATIO value, ADOP value, and fixed success rate can be obtained simultaneously.

[0041] S3-3. Output the above parameters to determine whether the double-difference wide-lane ambiguity is fixed:

[0042] The RATIO value, ADOP value, and fixation success rate are used together as a joint judgment index for fixation success. This overcomes the defect that a single test index is prone to getting trapped in local optima when the downwave signal stability is poor and cycle slip and multipath effects are significant in complex occlusion scenarios (specifically, when the number of real ambiguities decreases, the RATIO value shows a false increasing trend, while the actual positioning accuracy is difficult to guarantee). Accordingly, the current double-difference wide-lane ambiguity is determined to be fixed successfully only when all three meet the preset threshold conditions. This constitutes the dual reliability check mechanism, which blocks the pollution of the subsequent filtering state space by the error introduced by incorrect fixation from the source.

[0043] S3-4. Calculate the difference between the double-difference real ambiguity and the integer ambiguity. :

[0044]

[0045] in, This represents the real solution for the double-difference wide-lane ambiguity. Represents the integer solution of the double-difference wide-lane ambiguity;

[0046] Using the error propagation law, the floating-point solution coordinates are transformed into double-difference wide-lane fixed solution coordinates. :

[0047]

[0048] in, Solve for floating-point coordinates. Let be the covariance matrix of the coordinates of the real and floating-point solutions to the double-difference wide-lane ambiguity;

[0049] Therefore, the covariance matrix of the fixed solution of the double-difference wide alley Updated to:

[0050]

[0051] in, Represents the floating-point solution covariance matrix;

[0052] S3-5: If the fixation of the wide alley ambiguity of the double difference is unsuccessful, proceed directly to the S5 "Narrow alley ambiguity fixation" stage;

[0053] S4: Double-difference wide-lane ambiguity integer solution constrained floating-point solution filtering;

[0054] S5: Enter the narrow alley ambiguity fixing stage. If the narrow alley ambiguity cannot be fixed in the current complex environment, the fixed solution degradation reuse strategy is started: the fixed high-precision double-difference wide alley ambiguity solution result is used as the suboptimal positioning output to avoid discarding all high-precision observation information due to narrow alley fixing failure. This significantly improves the available positioning accuracy and overall integer ambiguity fixing rate of the RTK terminal in scenarios such as urban canyons and vegetation cover.

[0055] According to the present invention:

[0056] S1: Select single-difference ambiguity;

[0057] First, the state vector of the current epoch is traversed according to the constellation system. For each satellite within each system, the single-difference ambiguity at each carrier frequency is pre-screened, eliminating invalid data with ambiguity values ​​of zero, satellite unavailable, or abnormal carrier lockout markers. Further, single-difference ambiguities that meet the criteria of continuous lock duration greater than a preset lock threshold, no cycle slip marker, and satellite elevation angle greater than a preset elevation angle threshold are used to construct a single-difference ambiguity information table based on the correspondence between satellites and frequencies. Subsequently, based on the single-difference ambiguity information table, inter-frequency combination matching is performed on different frequency signals of the same satellite. Satellite combinations that simultaneously possess two valid single-difference ambiguities are selected and classified into corresponding wide-lane combination sets (such as L1 and L2 combination, L1 and L5 combination, L2 and L5 combination). Finally, for each wide-lane combination set that meets the satellite quantity requirement, the first satellite is selected as the reference satellite and the remaining satellites are selected as non-reference satellites. The single-difference ambiguity index and satellite identifier of the reference satellite and non-reference satellites at the corresponding two frequencies are extracted respectively. The difference between each pair is used to construct multiple double-difference wide-lane ambiguity information until the preset upper limit of the number of double-difference wide-lane ambiguities is reached.

[0058] S1. The preset upper limit of the number of double-difference wide-lane ambiguities is 30. This upper limit is obtained through multiple tests and calibrations based on the hardware configuration of this invention. Under this upper limit, a fixed success rate of over 95% can be achieved, and the time consumed by a single RTK calculation is controlled within 90ms, thus taking into account both accuracy and real-time requirements.

[0059] It should be noted that this upper limit is not the only constraint; if higher-cost, higher-performance hardware is used, the upper limit can be appropriately increased or even eliminated depending on the actual computing power, in order to further improve the fixation rate and solution time.

[0060] S2: Construct the double-difference wide-lane ambiguity parameter matrix;

[0061] After completing the single-difference ambiguity filtering in step one, it is necessary to construct the double-difference wide-lane ambiguity parameter matrix and its corresponding covariance matrix, and then fix them using the LAMBDA method. The construction process of the double-difference wide-lane ambiguity correlation matrix includes the following two matrix operations:

[0062] Step 1: Extract the double-difference ambiguity and its covariance matrix from the single-difference ambiguity in the state vector. Assume the current filtered state vector contains... One single-difference ambiguity parameter (denoted as a vector) Its corresponding covariance matrix is Based on the selected double-difference ambiguity information, a dimension is constructed as follows: Double difference operator matrix For matrices The first in line (corresponding to the first) (a double-difference ambiguity), if the single-difference ambiguity corresponding to its reference star is in The index in is The index corresponding to the non-reference star is Then the line is in The element at column position is 1. The element in column 1 is -1, and the rest are 0.

[0063] Based on double difference operator matrix Calculate the double-difference ambiguity vector Double-difference ambiguity covariance matrix The cross-covariance matrix of double-difference ambiguity and position parameters The formula is as follows:

[0064]

[0065]

[0066]

[0067] in, Position parameters in the state vector The initial cross-covariance submatrix of the single-difference ambiguity parameter, For matrix The transpose of .

[0068] Step 2: Construct the double-difference wide-lane ambiguity and its covariance matrix from a linear combination of double-difference ambiguities. Assume that based on the selected double-difference wide-lane ambiguity pairs (such as L1 and L2 frequency points), the dimension is constructed as follows: Wide lane combination operator matrix For matrices The first in line (corresponding to the first) (a double-difference wide-lane ambiguity), if the wide-lane is composed of a double-difference ambiguity at frequency 1. Double difference ambiguity at frequency point 2 Subtraction constitutes (i.e.) ), then in the matrix In the middle, corresponding In vector The column element in the middle position is set to 1, corresponding to In vector The column element at the middle position is set to -1, and the rest are 0.

[0069] Based on the wide lane combination operator matrix A linear transformation is performed on the double-difference ambiguity parameters obtained in the first step to calculate the double-difference wide-lane ambiguity vector. Double-difference wide-lane ambiguity covariance matrix And the ambiguity of the double-difference wide-lane and the location parameters cross-covariance matrix The formula is as follows:

[0070]

[0071]

[0072]

[0073] in, For the wide lane combination operator matrix The transpose matrix is ​​obtained. Through the above two matrix operations, the double-difference wide-lane floating-point solution vector for subsequent fixed-solution calculations is finally extracted. and its corresponding precision information matrix and .

[0074] S3: Fixed ambiguity of double-difference wide alleyway;

[0075] When using the LAMBDA method for ambiguity fixation, it is necessary to calculate the ADOP value and the fixation success rate. ADOP measures the precision and geometric strength of the floating-point ambiguity solution; the smaller the value, the more accurate the floating-point solution, the flatter the search space, and the more reliable the fixation. At this point, the covariance matrix of the double-difference wide-lane ambiguity needs to be considered. Perform LDL decomposition:

[0076]

[0077] in, It is a unit lower triangular matrix (with diagonal elements of 1). It is a diagonal matrix, and its diagonal elements are... It is a matrix Approximation of eigenvalues. Based on the properties of matrix decomposition, determinant It can be represented as:

[0078]

[0079] because If it is a unit lower triangular matrix, then the determinant is... ,therefore That is, matrix Multiplication of elements along the main diagonal:

[0080]

[0081] The final result of ADOP is:

[0082]

[0083] The success rate of a fixed solution reflects the current covariance matrix. Theoretically, the probability of successfully fixing the ambiguity is calculated using the following formula:

[0084]

[0085] Where the function This represents the cumulative distribution function of the standard normal distribution. After the LAMBDA method is executed, the RATIO value, ADOP value, and fixation success rate can be obtained. These parameters are output to determine whether the double-difference wide-lane ambiguity is fixed.

[0086] According to the present invention, when the RATIO value, ADOP value and fixation success rate all meet the requirements, the double-difference wide-lane ambiguity is considered to be successfully fixed.

[0087] According to the present invention, the requirements for the specific three data are as follows:

[0088] The empirical threshold for RATIO is 3.0, the empirical threshold for ADOP is 1.8, and the threshold for a fixed success rate must be greater than 0.99.

[0089] At this point, calculate the difference between the real and integer solutions of the double-difference wide-lane ambiguity:

[0090]

[0091] Using the law of error propagation, the floating-point solution coordinates are transformed into double-difference wide-lane fixed solution coordinates:

[0092]

[0093] Meanwhile, the covariance matrix of the fixed solution for the double-difference wide alley is updated as follows:

[0094]

[0095] S4: Integer double-difference wide-lane fuzzy constraint floating-point solution;

[0096] After obtaining the integer double-difference wide-lane ambiguity, a strong constraint correction is applied to the state vector of the Kalman filter through a virtual observation update mechanism. The virtual observation equation is the integer value of the double-difference wide-lane ambiguity:

[0097]

[0098] in Indicates a non-reference star. Indicates the reference star. Indicates two frequencies, This represents the inter-station single-difference ambiguity in the state matrix. The linear equation for the state variables is then expressed as:

[0099]

[0100] Therefore, the observation residual (Kalman filter innovation) can be obtained as follows:

[0101]

[0102] Since the Kalman filter state vector stores single-difference ambiguity parameters, a mapping relationship between double-difference ambiguity and single-difference ambiguity states needs to be established. For a double-difference combination consisting of four single-difference ambiguity state indices, the coefficient matrix... row vectors It can be represented as:

[0103]

[0104] Finally, the virtual observation equations described above are substituted into the extended Kalman filter for updating. This process forces the floating-point solution of the single-difference ambiguity to be constrained to an integer solution, and also uses the correlation between states to correct other state parameters such as coordinates, thereby achieving accurate convergence of the filter state and improving the success rate of subsequent narrow-lane ambiguity fixing and the accuracy of the positioning solution.

[0105] S5: Residual inspection and output after fixed solution of wide lane;

[0106] After fixing the narrow-lane ambiguity, if the narrow-lane ambiguity cannot be fixed at this point, a floating-point solution should be output. This invention, however, performs a post-hoc residual test on the wide-lane fixed solution. The core of this test is to construct a chi-square test statistic by calculating the quadratic form of the double-difference residuals (i.e., the sum of squared residuals) to evaluate the overall internal consistency quality of the wide-lane fixed solution. First, the residual vector of all double-difference observations is calculated through dot product operations. Sum of squares:

[0107]

[0108] in, This represents the total number of double-difference observations actually involved in the solution. Under the assumption that the observation errors follow a normal distribution, the sum of squared residuals... Obeying the degree of freedom The centered chi-square distribution. Its degrees of freedom (number of redundant observations) are calculated using the following formula:

[0109]

[0110] In the formula, This represents the total number of parameters to be estimated. The calculated values ​​will be used to determine the total number of parameters to be estimated. Compare with the critical value of the chi-square distribution (at a 95% confidence level). The judgment criterion is:

[0111]

[0112] If the above inequality holds, it indicates that the sum of squared residuals of the current wide-lane fixed solution is significantly large, and the model as a whole has a large error. The test fails, and a floating-point solution is still output. Conversely, if it does not exceed the threshold, it indicates that the post-hoc residuals of the wide-lane fixed solution are statistically within the normal fluctuation range, and the overall accuracy meets the standard. The test passes, and RTK returns the wide-lane fixed solution. This mechanism effectively avoids the output of erroneous solution results containing gross errors.

[0113] According to the present invention, the degrees of freedom are The "centralized" aspect of a centered chi-square distribution emphasizes that the mean of these standard normal variables is 0; if the mean is non-zero, it is called a non-centered chi-square distribution.

[0114] According to the present invention, the model as a whole has large errors, such as uncorrected cycle slips, gross error interference, or ambiguity fixing errors.

[0115] According to the present invention, the failure to fix the ambiguity of narrow lanes usually refers to the ambiguity parameter of a single frequency point with a short wavelength (e.g., 19 cm), which is greatly affected by noise and atmospheric errors, and the integer solution is difficult to fix directly.

[0116] The present invention also provides an apparatus and terminal for improving the integer ambiguity fixation rate of RTK.

[0117] An apparatus for improving the integer ambiguity fixation rate of RTK according to the present invention includes: a preprocessing module, a filtering module, and an ambiguity fixation module; wherein:

[0118] The preprocessing module performs parsing, transcoding, and quality assessment on the input satellite observation signals and differential data to obtain usable raw data.

[0119] The filtering module works by using the original data to perform state prediction and measurement updates on the localization solution, thereby obtaining new RTK localization results.

[0120] The ambiguity fixing module works by filtering the floating-point solution data obtained from the filtering module to obtain high-precision ambiguity parameters, and then using the LAMBDA method to fix the ambiguity.

[0121] A terminal for improving the integer ambiguity fixation rate of RTK according to the present invention includes:

[0122] GNSS antennas are used to receive radio frequency signals from GNSS satellites in the sky.

[0123] The radio frequency front-end circuit, connected to the GNSS antenna, is used to perform low-noise amplification, filtering, down-conversion, and analog-to-digital conversion on the radio frequency signal, and output a digital intermediate frequency signal.

[0124] The baseband chip, connected to the radio frequency front-end circuit, is used to perform correlation processing on the digital intermediate frequency signal to capture and track satellite signals, and output pseudorange, carrier phase observations and navigation messages;

[0125] A processor, connected to the baseband chip, is used to perform RTK calculations based on the observations and navigation messages, in order to improve the RTK integer ambiguity fixation rate.

[0126] A memory, connected to the processor, is used to store intermediate data and integer ambiguity fixing results during the RTK solution process;

[0127] A communication interface, connected to the processor, is used to receive differential correction data sent by the base station and output the final positioning result.

[0128] Beneficial effects of the present invention

[0129] The technical solution of this invention, based on the existing RTK step-by-step fixed strategy, makes targeted improvements to address the deficiencies in the wide-lane ambiguity resolution and utilization process. Its beneficial effects are mainly reflected in the following two aspects:

[0130] I. Introducing Multiple Indicator Checks and Virtual Observation Constraints to Reduce the Risk of State Space Contamination by Erroneous Wide-Lane Fixed Solutions. In conventional RTK solutions, wide-lane ambiguities are typically used directly as constraints in subsequent filtering. If the wide-lane solution itself contains errors, these errors propagate through the state covariance of the Kalman filter, causing distortion of the narrow-lane ambiguity floating-point solution and its covariance matrix. In step S3, this invention employs a combined check using RATIO, ADOP values, and the theoretical fixation success rate based on the normal distribution cumulative function. In step four, a strict mapping relationship is established between integer wide-lane ambiguities and single-difference ambiguities in the state vector by constructing a double-difference operator matrix and a wide-lane combination operator matrix, using these as virtual observations introduced into the extended Kalman filter for state updates. This mechanism restricts erroneous wide-lane solutions from entering the filtering process at the mathematical model level, ensuring the accuracy of the covariance information required for subsequent narrow-lane ambiguity solutions.

[0131] II. A wide-lane fixed solution degradation and reuse mechanism based on chi-square test is proposed to improve the continuity of positioning accuracy in complex environments. Addressing the problem of accuracy degradation caused by existing technologies directly outputting floating-point solutions when narrow-lane ambiguity fixation fails, this invention proposes a post-hoc residual test method for wide-lane fixed solutions in step five. This method calculates the sum of squares of the residuals of all double-difference observations and performs a hypothesis test against the chi-square distribution critical value with a 95% confidence level. If the statistic exceeds the threshold, it is determined that the model has gross errors or fixation errors, and a floating-point solution is still output to ensure the reliability of the results; if the statistic does not exceed the threshold, it indicates that the overall internal conformity quality of the current wide-lane fixed solution meets the standard, and it is output as the final positioning result. This strategy fully utilizes the superior accuracy of wide-lane fixed solutions compared to floating-point solutions, effectively reducing missing epochs in high-precision positioning results in complex observation environments, while ensuring that the positioning result does not output erroneous high-precision solutions due to undetected gross errors. Attached Figure Description

[0132] Figure 1 This is a flowchart summarizing the process of this invention.

[0133] Figure 2 This is a detailed flowchart of the present invention.

[0134] Figure 3a The route is a driving trajectory that does not use the strategy of this invention (green trajectory indicates a fixed solution).

[0135] Figure 3b The route is a driving trajectory using the strategy of this invention (green trajectory indicates a fixed solution).

[0136] Figure 4a The route trajectory for Route 2 without using the strategy of this invention is shown (green trajectory indicates a fixed solution).

[0137] Figure 4b The route trajectory for Route 2 using the strategy of this invention (green trajectory indicates a fixed solution).

[0138] Figure 5a The route trajectory for Route 3 without using the strategy of this invention is shown (green trajectory indicates a fixed solution).

[0139] Figure 5b The route trajectory for Route 3 using the strategy of this invention (green trajectory indicates a fixed solution).

[0140] Figure 6 This is a structural diagram of the receiver terminal implemented in this invention. Detailed Implementation

[0141] Example:

[0142] To verify the effectiveness of the algorithm proposed in this invention, a real-vehicle dynamic test platform was constructed. Specifically, the firmware of the positioning algorithm to be verified was embedded in a GNSS receiver terminal, which communicates with the onboard computer via a serial port. The test terminal uses the network RTK differential service provided by China Mobile as the base station data source. This data stream is routed to the onboard computer via the wireless network and then forwarded to the receiver via the serial port, thereby realizing differential positioning calculation.

[0143] In terms of hardware configuration, the satellite signal receiving antenna is mounted on the top of the test vehicle and physically connected to the test terminal. To effectively suppress multipath effects, this experiment uses a "mushroom head" measurement antenna with a choke coil structure, and rigidly fixes it to the roof using magnetic attraction and binding to avoid antenna phase center shift caused by vehicle bumps or wind resistance.

[0144] The test scenarios covered typical complex environments such as open roads, tree-lined areas, under viaducts, and urban high-rise blocks, aiming to comprehensively evaluate the algorithm's performance under different signal conditions. Three test routes were used: Route 1 was approximately 20 kilometers long, Route 2 was approximately 50 kilometers long, and Route 3 was approximately 100 kilometers long. After the test vehicle started, the GNSS receiver board acquired data from the base station and performed RTK calculations. First, the velocity value calculated using Doppler and the RTK result from the previous moment was used to recursively deduce the velocity, obtaining the Kalman filter's predicted state. Subsequently, a double-difference observation equation for pseudorange and carrier phase observations was established for observation updates.

[0145]

[0146] in, This represents the double-difference carrier phase observation value. Represents the distance between sets with two differences. Indicates wavelength. Indicates an unknown number for the entire week. This represents the double-difference random error.

[0147] After the observations are updated, the RTK floating-point solution is obtained. At this point, it is necessary to select the inter-station single-difference ambiguities that meet the requirements from the state matrix X, where the state matrix X can be represented as follows:

[0148]

[0149] in, Represents coordinate values. This indicates the ambiguity of a single difference.

[0150] First, the single-difference ambiguities in the state vector are traversed by constellation system. After eliminating systems that do not participate in the fixed configuration, the data of each satellite is screened one by one by frequency dimension. If the ambiguity value is zero or the satellite is unavailable, it is skipped; if the satellite meets the requirements of continuous lock duration greater than 20, no cycle slip marker, and elevation angle greater than 30 degrees, the satellite identifier and its single-difference ambiguity index in the state vector are stored in the single-difference ambiguity information table.

[0151] Subsequently, based on the above information table, frequency combination matching is performed. Each satellite in the table is traversed to detect whether it has two effective single-difference ambiguities at the same time. If so, it is classified into the L1L2, L1L5 or L2L5 wide-lane combination list according to the frequency type.

[0152] Next, double-difference wide-lane ambiguities are generated based on the list of wide-lane combinations. For lists with more than one satellite, the first satellite is selected as the reference satellite, and the rest are non-reference satellites. The single-difference ambiguity indices, satellite identifiers, and frequency types of the reference and non-reference satellites at corresponding two frequencies are extracted, and the pairwise differences are used to construct the double-difference wide-lane ambiguity information structure.

[0153] Finally, during the construction process, the total number of generated double-difference wide-lane ambiguities is judged in real time. If the preset upper limit is reached, the selection process is terminated immediately and the current set is returned to prevent the dimension of subsequent matrix operations from exceeding the limit. If the upper limit is not reached, the traversal continues until all systems have been processed.

[0154] After selection, it is determined whether the number of double-difference wide-lane ambiguities is less than 4. If it is less than 4, the ambiguity fixing operation is stopped and the process exits directly. If it is greater than 4, matrix operations are used to construct the double-difference ambiguities, their corresponding covariances, and the covariance between the double-difference ambiguities and the position parameters. After obtaining the above covariances, a matrix related to the double-difference wide-lane ambiguities is constructed. This matrix is ​​then substituted into the LAMBDA algorithm for fixing, and its Ratio, ADOP, and fixing success rate are calculated.

[0155]

[0156]

[0157]

[0158] in, This indicates the number of double-difference wide-lane ambiguities. The main diagonal elements of the D matrix after LD decomposition of the double-difference wide-lane ambiguity covariance matrix are represented by the function. The cumulative distribution function represents the standard normal distribution.

[0159] If the above indicators pass, then using the error propagation law, the floating-point solution coordinates are transformed into double-difference wide-lane fixed solution coordinates:

[0160]

[0161] After obtaining the integer double-difference wide-lane ambiguity, a strong constraint correction is applied to the state vector of the Kalman filter using a virtual observation update mechanism. Finally, after fixing the narrow-lane ambiguity, if the narrow-lane ambiguity cannot be fixed at this point, a chi-square test is performed on the apostolic residuals of the wide-lane fixed solution.

[0162]

[0163] If the sum of squares of the post-test residuals is less than the chi-square critical value, then the wide-lane fixed solution is output as the RTK fixed solution.

[0164] The following are the dynamic test trajectories with and without wide-lane ambiguity fixing enabled. It can be seen that enabling wide-lane ambiguity fixing significantly improves the fixing rate (the number of green trajectories):

[0165] The overall fixed rate for Route 1 increased from 69.7% to 94.3%, the overall fixed rate for Route 2 increased from 74.4% to 90.7%, and the overall fixed rate for Route 3 increased from 64.0% to 84.1%.

[0166] Furthermore, Table 1 compares the total positioning errors of the three lines with and without the strategy of this invention. The data shows that the positioning errors in the east, north, and zenith directions using the strategy of this invention are all smaller than those without the strategy. This demonstrates that, under complex traffic conditions, the accuracy of the output wide-lane fixed solution is still higher than that of the floating-point solution, verifying the effectiveness of the fixed solution degradation and reuse strategy.

[0167] Table 1. Positioning errors with and without the present invention's strategy enabled:

[0168] .

Claims

1. A method for improving the integer ambiguity fixation rate in RTK, applicable to complex dynamic environments such as urban canyons and vegetation shading, characterized in that, Includes the following steps: S1: Based on real-time satellite signal quality and observation noise level, pre-screen effective single-difference ambiguities, remove satellites that are severely polluted by multipath effects or signal attenuation, and obtain double-difference wide-lane ambiguities that reach the preset upper limit, so as to provide a clean ambiguity candidate set for subsequent fixing. S2: Construct the double-difference wide-lane ambiguity parameter matrix: S2-1. After completing the single-difference ambiguity screening in S1, construct the double-difference wide-lane ambiguity parameter matrix and its corresponding covariance matrix; S2-2. The construction process of the double-difference wide-lane ambiguity parameter matrix includes the following two matrix operations: Step 1: Extract the double-difference ambiguity and its covariance matrix from the single-difference ambiguity in the state vector. Assume that the current filtered state vector contains... A single-difference ambiguity parameter, denoted as a vector. Its corresponding covariance matrix is Based on the selected double-difference ambiguity information, a dimension is constructed as follows: Double difference operator matrix Based on double difference operator matrix Calculate the double-difference ambiguity vector Double-difference ambiguity covariance matrix And the cross-covariance matrix of double-difference ambiguity and position parameters ; Step 2: Construct the double-difference wide-lane ambiguity and its covariance matrix by linear combination of double-difference ambiguities. Assume that based on the selected double-difference wide-lane ambiguity combinations, such as L1 and L2, L1 and L5, L2 and L5, the dimension is... Wide lane combination operator matrix In particular, for situations where the number of visible satellites is small in obstructed scenarios, when three-frequency data is available, the number of double-difference wide-lane ambiguities is increased by combining multiple frequencies to ensure the dimensionality requirement of LAMBDA search, and the ultra-wide-lane wavelength characteristics formed by the combination of L2 and L5 are used to reduce the difficulty of ambiguity fixation. S3: Fixed ambiguity of double-difference wide-lane: S3-1. Use the LAMBDA method to fix ambiguity: First, the ADOP value is calculated. ADOP measures the precision and geometric strength of the ambiguity floating-point solution. The smaller the value, the more accurate the floating-point solution, the flatter the search space, and the more reliable the fixation. In this invention, the empirical threshold for ADOP is set to 1.

8. At this point, the covariance matrix of the double-difference wide-lane ambiguity... Perform LDL decomposition: in, It is a unit lower triangular matrix with diagonal elements all being 1. It is a diagonal matrix, and the elements on its diagonal are represented as: This element is also a matrix. Eigenvalue approximation, based on the properties of matrix decomposition. determinant Represented as: because If it is a unit lower triangular matrix, then the determinant is... ,therefore That is, matrix Multiplication of elements along the main diagonal: The final result of ADOP is: in, This indicates the number of double-difference wide-lane ambiguities. Representation matrix The main diagonal element, S3-2. Calculate the success rate of a fixed solution: The success rate of a fixed solution reflects the current covariance matrix. Theoretically, the probability of successfully fixing the ambiguity is calculated using the following formula: Among them, the function The cumulative distribution function represents the standard normal distribution. After the LAMBDA method is executed, the RATIO value, ADOP value, and fixed success rate can be obtained simultaneously. S3-3. Output the above parameters to determine whether the double-difference wide-lane ambiguity is fixed: The RATIO value, ADOP value, and fixation success rate are used together as a joint judgment index for fixation success. This is to overcome the defect that a single test index is prone to getting trapped in local optima when the downwave signal is unstable, cycle slip and multipath effect are significant in complex occlusion scenarios. Accordingly, the current double difference wide lane ambiguity is determined to be fixed successfully only when all three meet the preset threshold conditions. This constitutes a dual reliability check mechanism to prevent the pollution of the subsequent filtering state space by the error introduced by incorrect fixation from the source. S3-4. Calculate the difference between the double-difference real ambiguity and the integer ambiguity. : in, This represents the real solution for the double-difference wide-lane ambiguity. Denotes the integer solution of the double-difference wide-lane ambiguity. Using the error propagation law, the floating-point solution coordinates are transformed into double-difference wide-lane fixed solution coordinates. : in, Solve for floating-point coordinates. Let be the covariance matrix of the coordinates of the real and floating-point solutions to the double-difference wide-lane ambiguity. Therefore, the covariance matrix of the fixed solution of the double-difference wide alley Updated to: in, Represents the floating-point solution covariance matrix; S3-5: If the fixation of the wide-lane ambiguity due to the double difference fails, proceed directly to the S5 "Narrow-lane ambiguity fixation" stage; S4: Double-difference wide-lane ambiguity integer solution constrained floating-point solution filtering; S5: Enter the narrow alley ambiguity fixing stage. If the narrow alley ambiguity cannot be fixed in the current complex environment, the fixed solution degradation reuse strategy is started: the fixed high-precision double-difference wide alley ambiguity solution result is used as the suboptimal positioning output to avoid discarding all high-precision observation information due to narrow alley fixing failure. This significantly improves the available positioning accuracy and overall integer ambiguity fixing rate of the RTK terminal in scenarios such as urban canyons and vegetation cover.

2. The method for improving the fixed rate of integer ambiguity in RTK as described in claim 1, characterized in that, S1. Includes: S1-1: First, traverse the state vector of the current epoch according to the constellation system, pre-screen the single-difference ambiguity of each satellite in each system at each carrier frequency to obtain the effective single-difference ambiguity, and construct a single-difference ambiguity information table based on the correspondence between satellite and frequency. S1-2: Subsequently, based on the single-difference ambiguity information table, inter-frequency combination matching, i.e. signal fusion processing, is performed on different frequency signals of the same satellite to filter out satellite combinations that simultaneously possess two effective single-difference ambiguities and classify them into the corresponding wide-lane combination set; S1-3: Select and extract the single-difference ambiguity index and satellite identifier of the reference satellite and non-reference satellite at the corresponding two frequencies, respectively, and construct multiple double-difference wide-lane ambiguity information by subtracting them pairwise until the preset upper limit of the number of double-difference wide-lane ambiguities is reached.

3. The method for improving the fixed rate of integer ambiguity in RTK as described in claim 2, characterized in that, In S1-1: Remove invalid data with a ambiguity value of zero, satellite unavailable, or abnormal carrier lockout markers.

4. The method for improving the RTK integer ambiguity fixation rate as described in claim 2, characterized in that, In S1-2: The effective single-difference ambiguity satisfies the following conditions: continuous locking duration is greater than the preset locking threshold, there is no cycle slip, and the satellite elevation angle is greater than the preset elevation angle threshold. The preset locking threshold is 20, and the preset elevation angle threshold is 30 to 35 degrees.

5. A method for improving RTK integer ambiguity fixation rate as described in claim 1 or 2, characterized in that, S1. The preset upper limit of the number of double-difference wide-lane ambiguities is 30. This upper limit is based on the hardware configuration of the present invention and is obtained after multiple tests and calibrations. Under this upper limit, a fixed success rate of over 95% can be achieved, and the time consumed by a single RTK calculation can be controlled within 90ms, thus taking into account both accuracy and real-time requirements.

6. The method for improving the fixed rate of integer ambiguity in RTK as described in claim 1, characterized in that, For double difference operator matrices The first in The line, that is, the line corresponding to the first If a double-difference ambiguity exists, and the single-difference ambiguity corresponding to its reference star is in The index in is The index corresponding to the non-reference star is Then the double difference operator matrix The first in Walking The element at column position is 1. The element in column 1 is -1, and the rest are 0.

7. A method for improving RTK integer ambiguity fixation rate as described in claim 1 or 6, characterized in that, Based on double difference operator matrix Calculate the double-difference ambiguity vector Double-difference ambiguity covariance matrix And the cross-covariance matrix of double-difference ambiguity and position parameters The formula is as follows: in, Position parameters in the state vector The initial cross-covariance submatrix of the single-difference ambiguity parameter, For matrix The transpose of .

8. The method for improving the fixed rate of integer ambiguity in RTK as described in claim 1, characterized in that, After S2-2, S2-3. Based on the wide-lane combination operator matrix A linear transformation is performed on the double-difference ambiguity parameters obtained in the first step to calculate the double-difference wide-lane ambiguity vector. Double-difference wide-lane ambiguity covariance matrix And the ambiguity of the double-difference wide-lane and the location parameters cross-covariance matrix The formula is as follows: in, For the wide lane combination operator matrix The transpose of the matrix is ​​used to extract the double-difference wide-lane floating-point solution vector for subsequent fixed-solution calculations through the two matrix operations described above. and its corresponding precision information matrix and .

9. The method for improving the fixed rate of integer ambiguity in RTK as described in claim 1, characterized in that, In S3-3, when the RATIO value, ADOP value, and fixation success rate all meet the specific threshold requirements, the double-difference wide-lane ambiguity is considered to be successfully fixed. The specific threshold requirements include: the RATIO value is greater than the empirical threshold of 3.0, the ADOP value is less than the empirical threshold of 1.8, and the fixation success rate is greater than 0.

99.

10. The method for improving the RTK integer ambiguity fixation rate as described in claim 9, characterized in that, In S3-3, to avoid a single indicator getting stuck in a local optimum, it is also necessary to determine whether the ratio of RATIO value to ADOP value is less than a preset threshold. The preset threshold is dynamically set according to the number of double-difference wide-lane ambiguities: when the number of double-difference wide-lane ambiguities is less than 9, the preset threshold is set to 3.5; otherwise, the preset threshold is set to 2.

0.

11. The method for improving the fixed rate of integer ambiguity in RTK as described in claim 1, characterized in that, In S4: Double-difference wide-lane ambiguity integer solution constrained floating-point solution filtering: S4-1. After obtaining the integer solution of the double-difference wide-lane ambiguity, the state vector of the Kalman filter is strongly constrained and corrected through a virtual observation update mechanism. The virtual observation equation is the integer value of the double-difference wide-lane ambiguity. : in Indicates a non-reference star. Indicates the reference star. Indicates two frequencies, This represents the inter-station single-difference ambiguity in the state matrix. The linear equation for the state variables is then expressed as: From this, the observation residual can be obtained. (Kalman filter information) is: 。 12. The method for improving the fixed rate of integer ambiguity in RTK as described in claim 11, characterized in that, Since the Kalman filter state vector stores single-difference ambiguity parameters, a mapping relationship between double-difference wide-lane ambiguity and single-difference ambiguity states needs to be established. Specifically, for a certain double-difference wide-lane combination composed of four single-difference ambiguity state indices, its corresponding coefficient matrix... row vector It can be represented as: In this row vector, the single-difference ambiguity state position corresponding to the non-reference star is set to 1, the single-difference ambiguity state position corresponding to the reference star is set to -1, and the other state positions not involved in this combination are set to 0. By multiplying this row vector with the single-difference ambiguity state vector, the corresponding double-difference wide-lane ambiguity observation value can be obtained.

13. The method for improving the fixed rate of integer ambiguity in RTK as described in claim 11, characterized in that, S4-2. Introduce the above virtual observation equation into the state update step of the Extended Kalman Filter (EKF). Use the difference between the integer and floating-point solutions of the double-difference wide-lane ambiguity as the virtual observation vector. Construct a gain matrix by combining the cross-covariance matrix between the double-difference wide-lane ambiguity and the position parameters. Correct the filter's state vector through the filter state update equation. The state vector includes the position parameters, and the calculation formula is as follows: in, This is the corrected state vector. This is the state vector before correction. The filter gain matrix is ​​determined using the cross-covariance matrix. For the integer solution of the double-difference wide-lane ambiguity, The floating-point solution is the double-difference wide-lane ambiguity. Through the above correction, the position parameters are constrained to a high-precision space consistent with the integer solution of the wide-lane ambiguity, thereby improving the success rate of subsequent narrow-lane ambiguity fixation and the accuracy of the positioning solution.

14. The method for improving the fixed rate of integer ambiguity in RTK as described in claim 1, characterized in that, S5-1: After a narrow alleyway fix fails, a fixed solution degradation and reuse strategy is initiated. A post-test residual check is performed on the wide alleyway fix solution. If the post-test residual passes the chi-square test, the wide alleyway fix solution is output as the final RTK positioning result. The core of this test is to construct the chi-square test statistic by calculating the quadratic form of the double-difference residuals, i.e., the sum of squared residuals, to assess the overall internal consistency quality of the wide-lane fixed solution. First, the residual vector of all double-difference observations is calculated using the dot product operation. Sum of squares: in, The sum of squared residuals represents the total number of double-difference observations actually used in the solution, assuming the observation errors follow a normal distribution. Obeying the degree of freedom The centralized chi-square distribution, Its degrees of freedom, i.e., the number of redundant observations, are calculated using the following formula: In the formula, The parameter to be estimated is the total number of necessary observations.

15. The method for improving the fixed rate of integer ambiguity in RTK as described in claim 14, characterized in that, After the aforementioned dual reliability check mechanism and wide-lane fixed solution constraint processing, the residuals of the double-difference observations approximately satisfy the characteristics of zero mean, mutual independence, and following the same Gaussian distribution, and their sum of squared residuals thus follows a chi-square distribution; based on this, the current epoch's wide-lane fixed solution is evaluated in real time to determine whether it meets the threshold of the chi-square test, and the calculated... Compared with the critical value of the chi-square distribution, the confidence level of the critical value of the chi-square distribution is set at 95%, and the judgment criterion is: The critical value of the chi-square distribution is obtained by looking up a table. The inequality holds true, indicating that the sum of squared residuals of the current wide-lane fixed solution is significantly larger, and the model as a whole has a large error. Therefore, the test fails, and a floating-point solution is still output. Conversely, if the critical value is not exceeded, it indicates that the post-test residual of the wide-lane fixed solution is statistically within the normal fluctuation range, the overall accuracy meets the standard, the test is passed, and RTK returns the wide-lane fixed solution.

16. An apparatus for improving RTK integer ambiguity fixation rate, used in the method for improving RTK integer ambiguity fixation rate as described in claim 1, characterized in that, include: The module consists of a preprocessing module, a filtering module, and a fuzziness fixing module; among which, The preprocessing module performs parsing, transcoding, and quality assessment on the input satellite observation signals and differential data to obtain usable raw data. The filtering module works by using the original data to perform state prediction and measurement updates on the localization solution, resulting in a new RTK localization result. The ambiguity fixing module works by filtering the floating-point solution data obtained from the filtering module to obtain high-precision ambiguity parameters, and then using the LAMBDA method to fix the ambiguity.

17. A terminal for improving RTK integer ambiguity fixation rate, used in the method for improving RTK integer ambiguity fixation rate as described in claim 1, characterized in that, include: GNSS antennas are used to receive radio frequency signals from GNSS satellites in the sky. The radio frequency front-end circuit, connected to the GNSS antenna, is used to perform low-noise amplification, filtering, down-conversion, and analog-to-digital conversion on the radio frequency signal, and output a digital intermediate frequency signal. The baseband chip, connected to the radio frequency front-end circuit, is used to perform correlation processing on the digital intermediate frequency signal to capture and track satellite signals, and output pseudorange, carrier phase observations and navigation messages; A processor, connected to the baseband chip, is used to perform RTK calculations based on the observations and navigation messages to implement the method for improving the RTK integer ambiguity fixation rate as described in claim 1; A memory, connected to the processor, is used to store intermediate data and integer ambiguity fixing results during the RTK solution process; A communication interface, connected to the processor, is used to receive differential correction data sent by the base station and output the final positioning result.

Images