Beidou navigation software development data updating method and system

Abstract

The application relates to the field of Beidou navigation software development and testing, in particular to a Beidou navigation software development data updating method and system. The method comprises the following steps: constructing a parallel positioning data processing path; obtaining original positioning data flow; correcting or replacing low-precision data points in the original positioning data flow to obtain benchmark positioning data flow; testing the original positioning data flow according to a preset test rule to obtain test positioning data flow; recording benchmark internal key state variables of a benchmark data path processing process and measured internal key state variables of a measured data path processing process; performing failure determination based on the difference between the benchmark internal key state variables and the measured internal key state variables to determine whether there is an unexpected interaction logic abnormality in the positioning data processing path processing flow. The method has the advantages that unexpected interaction logic abnormality in an upper position fusion module caused by low-precision data flow can be detected, and the efficiency and accuracy of software testing are significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of Beidou navigation software development and testing, in particular to a Beidou navigation software development data updating method and system. BACKGROUND

[0002] In the software development and testing process of highly dependent positioning information such as vehicle navigation system, it is crucial to ensure the correctness and stability of the function of the software after the update of the underlying hardware (such as Beidou positioning chip or module). The traditional regression testing method usually relies on a large historical radio frequency signal database to conduct comprehensive testing on a hardware-in-the-loop simulation platform. This method often takes a lot of time and is difficult to adapt to the current fast iteration development rhythm, resulting in low testing efficiency. When only the underlying hardware driver and the original observation data analysis interface of the navigation system are modified, and the upper application logic (such as the position fusion module) remains unchanged, the error mode introduced by the new hardware that is not detailed in the public specification may cause deep interaction problems. For example, when facing a specific signal, the new generation of positioning chip may no longer simply discard the data, but output pseudo-range values with specific markers but significantly reduced precision. This subtle change in data characteristics may cause the unmodified upper position fusion module to misinterpret the continuous low-precision data stream output by the new chip based on its original error model (for example, assuming that the data will be lost). This may cause the trust level, fusion weight or data consistency index of the internal logic of the fusion module to be adjusted unexpectedly, thereby causing internal logic abnormalities, even though the final positioning output result may only show temporary and hidden deviations.

[0003] The hidden nature of such internal logic abnormalities lies in the fact that they are not caused by defects in the new driver itself, but by the fact that the internal processing logic of the upper module fails to respond correctly when receiving abnormal data streams, such as incorrectly assigning weights or generating unexpected residual characteristics. The traditional black box testing method based on the final output result is prone to miss such occasional and temporary internal logic problems that only appear under specific vehicle dynamics due to its invisibility to internal states.

[0004] In view of the above problems, the prior art needs to be improved. SUMMARY

[0005] The purpose of the present application is to provide a Beidou navigation software development data updating method and system, which can effectively detect the unexpected interaction logic abnormalities of the upper position fusion module caused by low-precision data streams after the update of the underlying hardware of the Beidou navigation software, and significantly improve the efficiency and accuracy of software testing.

[0006] The present application provides a Beidou navigation software development data updating method, which comprises:

[0007] constructing a parallel positioning data processing path, the positioning data processing path comprising a reference data path and a measured data path;

[0008] obtaining an original positioning data stream, the original positioning data stream containing a plurality of positioning data points and corresponding state markers, the state markers indicating data accuracy states;

[0009] processing the original positioning data stream in the reference data path: correcting or replacing low-accuracy data points in the original positioning data stream to obtain a reference positioning data stream;

[0010] processing the original positioning data stream in the measured data path: testing the original positioning data stream according to a preset test rule to obtain a test positioning data stream;

[0011] recording reference internal key state variables of the reference data path processing process and measured internal key state variables of the measured data path processing process, the reference internal key state variables being used to represent trustworthiness, fusion weight, data consistency index and / or residual characteristics of different types of positioning information sources in the reference data path processing process, the measured internal key state variables being consistent with the reference internal key state variables in terms of parameter types and being used to represent dynamic adjustment responses in the processing logic under the interference of low-accuracy data;

[0012] performing failure determination based on differences between the reference internal key state variables and the measured internal key state variables to determine whether there is an unexpected interaction logic abnormality in the positioning data processing path processing flow.

[0013] Through the above scheme, unexpected interaction logic abnormalities in the upper position fusion module caused by low-accuracy data streams after the underlying hardware is updated can be effectively detected, and the efficiency and accuracy of software testing are significantly improved.

[0014] Optionally, the present application also proposes that the low-accuracy data points in the original positioning data stream are corrected or replaced, comprising:

[0015] obtaining vehicle motion information for indicating a vehicle operating state, the vehicle motion information including vehicle speed, acceleration and / or steering angle;

[0016] determining whether a target processing data point in the original positioning data stream is a low-accuracy data point according to a preset accuracy determination rule;

[0017] if the target processing data point is a low-accuracy data point, determining whether the vehicle is in a stationary state or a moving state according to the vehicle motion information;

[0018] if the vehicle is in the stationary state, replacing the current target processing data point with a previous valid positioning data point.

[0019] If the vehicle is in motion, based on the previous valid positioning data point and vehicle motion information, a replacement positioning data point corresponding to the vehicle motion trend is calculated and generated, and the current target processing data point is corrected with the replacement positioning data point.

[0020] Optionally, the application also proposes that the reference internal key state variables of the reference data path processing process and the measured internal key state variables of the measured data path processing process are recorded, including:

[0021] The reference internal processing data generated in the reference data path processing process is obtained, and the reference internal processing data includes timestamps, position solution values, weight distribution values, and fusion residual values of each type of positioning information source;

[0022] At least one reference internal key state variable is extracted from the reference internal processing data based on a preset state extraction rule;

[0023] The measured internal processing data generated in the measured data path processing process is obtained, and the measured internal processing data includes data fields corresponding to the reference internal processing data;

[0024] Based on the state extraction rule, the measured internal key state variable corresponding to the parameter type of the reference internal key state variable is extracted from the measured internal processing data.

[0025] Optionally, the application also proposes that the failure is determined based on the difference between the reference internal key state variable and the measured internal key state variable, and whether there is an unexpected interaction logic abnormality in the positioning data processing path processing flow is determined, including:

[0026] The paired values of the reference internal key state variable and the measured internal key state variable at the same timestamp are obtained;

[0027] The difference index is calculated for each pair of paired values, and the form of the difference index includes absolute difference, normalized offset, and / or residual growth rate;

[0028] The difference index is compared with the corresponding preset difference threshold value, and the difference threshold value is adaptively adjusted according to the running state of the current vehicle;

[0029] If the difference index exceeds the difference threshold value, it is determined that there is an unexpected interaction logic abnormality in the measured data path, and the abnormal time period and the corresponding variable change trend are recorded.

[0030] Optionally, the application also proposes that the difference threshold value is adaptively adjusted according to the running state of the current vehicle, including:

[0031] The dynamic information of the current vehicle is obtained;

[0032] When the dynamic information indicates that the vehicle is in a high-speed or sharp turning state, the difference threshold value is increased according to a preset adjustment rule;

[0033] When the dynamic information indicates that the vehicle is in a low-speed or stable cruising state, the difference threshold value is decreased according to a preset adjustment rule.

[0034] Optionally, the application also proposes that the failure judgment is based on the difference between the reference internal key state variable and the measured internal key state variable, and whether there is an unexpected interaction logic anomaly in the positioning data processing path processing flow is determined, and the method further comprises:

[0035] Based on the time sequence of the measured internal key state variable, an abnormal time period in which the difference value index continuously exceeds the difference threshold value is identified;

[0036] The variable change mode in the abnormal time period is matched with a preset abnormal mode library, and the abnormal mode library includes a fusion drift mode, a residual burst mode, a trust reversal mode, and / or an information source switching abnormal mode;

[0037] According to the matching result, the abnormal type of the unexpected interaction logic anomaly is determined, a corresponding abnormal classification identifier is generated, and is added to the recorded abnormal time period.

[0038] Optionally, the application also proposes that according to the matching result, the abnormal type of the unexpected interaction logic anomaly is determined, a corresponding abnormal classification identifier is generated, and is added to the recorded abnormal time period, which comprises:

[0039] According to the matching result, the abnormal type of the unexpected interaction logic anomaly is determined;

[0040] If the abnormal type is a fusion drift mode or a residual burst mode, an abnormal segment elimination operation of the test positioning data stream is triggered, the data points in the corresponding abnormal time period are excluded, and a corresponding abnormal response execution result is generated;

[0041] If the abnormal type is a trust reversal mode or an information source switching abnormal mode, the current positioning information source is marked as a suspicious information source, the fusion weight of the suspicious information source is reduced or a backup information source is enabled in the subsequent data synchronization process, and a corresponding abnormal response execution result is generated;

[0042] The abnormal response execution result and the abnormal classification identifier corresponding to the abnormal type are attached to the abnormal time period to form an abnormal response chain.

[0043] Optionally, the application also proposes that the method further comprises:

[0044] The number and proportion of data segments excluded due to unexpected interaction logic anomalies in the test positioning data stream are counted, and an abnormal elimination index for evaluating positioning robustness is generated.

[0045] analyze the frequency and associated time period of each type of anomaly in the test positioning data stream, and construct an anomaly distribution spectrum at the version level;

[0046] Based on the anomaly elimination index and the anomaly distribution spectrum, update the test evaluation result of the software version corresponding to the measured data path.

[0047] Through the above scheme, a method for quantitative evaluation and version-level analysis of test results is provided, which helps to comprehensively evaluate the positioning robustness of the software version and guide subsequent optimization and improvement.

[0048] Optionally, the present application also proposes that the method further comprises:

[0049] During the processing of the measured data path, a source mapping relationship is established for each positioning data point of the original positioning data stream, and the original positioning information type, timestamp and correction path identifier of the positioning data point are recorded;

[0050] Based on the source mapping relationship, a path tree structure of the positioning data point is constructed, and the path nodes of the path tree structure record the processing stage, state marker, correction method and variable value;

[0051] When the failure determination result is an unexpected interaction logic anomaly, the formation process of the positioning data point is traced according to the path tree structure, and a key variable evolution chain related to the abnormal response chain is extracted;

[0052] The key variable evolution chain is used as a backtracking basis for processing the measured internal key state variable.

[0053] Optionally, the present application also proposes a Beidou navigation software development data updating system, the system comprising:

[0054] A path construction module is configured to construct a parallel positioning data processing path, the positioning data processing path comprising a reference data path and a measured data path;

[0055] A data acquisition module is configured to acquire an original positioning data stream, the original positioning data stream comprising a plurality of positioning data points and corresponding state markers, the state markers indicating data accuracy states;

[0056] A reference data processing module is configured to process the original positioning data stream through the reference data path: correcting or replacing low-precision data points in the original positioning data stream to obtain a reference positioning data stream;

[0057] A test data processing module is configured to process the original positioning data stream through the measured data path: testing the original positioning data stream according to a preset test rule to obtain a test positioning data stream;

[0058] The state variable recording module is configured to record a reference internal key state variable of the reference data path processing process and a measured internal key state variable of the measured data path processing process, the reference internal key state variable is used to represent a trust degree, a fusion weight, a data consistency index and / or a residual feature of different types of positioning information sources in the reference data path processing process, and the measured internal key state variable is consistent with the reference internal key state variable in terms of a parameter type and is used to represent a dynamic adjustment response in the processing logic under low-precision data interference.

[0059] The judging module is configured to perform failure determination based on a difference between the reference internal key state variable and the measured internal key state variable, and determine whether there is an unexpected interaction logic abnormality in the positioning data processing path processing flow.

[0060] Through the above scheme, a system architecture for implementing the above method is provided, so that the method can be actually deployed and applied, and a complete solution is provided for the development and testing of Beidou navigation software.

[0061] As can be seen from the above, the Beidou navigation software development data updating method and system provided by the present application can effectively detect unexpected interaction logic abnormalities in the upper position fusion module caused by low-precision data streams after the underlying hardware is updated, and significantly improve the efficiency and accuracy of software testing. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 A flowchart of a Beidou navigation software development data updating method provided by one embodiment of the present application.

[0063] Figure 2 A flowchart of a Beidou navigation software development data updating method provided by one embodiment of the present application.

[0064] Figure 3 A flowchart of a Beidou navigation software development data updating method provided by one embodiment of the present application.

[0065] Figure 4 A flowchart of a Beidou navigation software development data updating method provided by one embodiment of the present application.

[0066] Figure 5 A flowchart of a Beidou navigation software development data updating method provided by one embodiment of the present application.

[0067] Figure 6A flowchart of a Beidou navigation software development data updating method according to another embodiment of the present application.

[0068] Figure 7 A flowchart of a Beidou navigation software development data updating method according to another embodiment of the present application.

[0069] Figure 8 A flowchart of a Beidou navigation software development data updating method according to another embodiment of the present application.

[0070] Figure 9 A flowchart of a Beidou navigation software development data updating method according to another embodiment of the present application.

[0071] Figure 10 A program block diagram of a Beidou navigation software development data updating system according to another embodiment of the present application.

[0072] In the figure: 1, path construction module; 2, data acquisition module; 3, reference data processing module; 4, test data processing module; 5, state variable recording module; 6, judgment module. DETAILED DESCRIPTION

[0073] The technical solutions in the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the present application described and shown in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0074] REFERENCE Figure 1 To this end, the present application provides a Beidou navigation software development data updating method, which comprises:

[0075] S1000: Construct a parallel positioning data processing path, which comprises a reference data path and a measured data path;

[0076] S2000: Obtain an original positioning data stream, which contains a plurality of positioning data points and corresponding state marks, the state marks indicating data accuracy states;

[0077] S3000: Process the original positioning data stream through the reference data path: correct or replace low-precision data points of the original positioning data stream to obtain a reference positioning data stream;

[0078] S4000: Perform the measured data path processing on the original positioning data stream: perform the test on the original positioning data stream according to the preset test rule to obtain a test positioning data stream;

[0079] S5000: Record the reference internal key state variables of the reference data path processing process and the measured internal key state variables of the measured data path processing process, the reference internal key state variables are used to represent the trust degree, fusion weight, data consistency index and / or residual feature of the reference data path processing process on different types of positioning information sources, and the parameter types of the measured internal key state variables are consistent with those of the reference internal key state variables and are used to represent the dynamic adjustment response of the processing logic under the interference of low-precision data;

[0080] S6000: Perform failure determination based on the difference between the reference internal key state variables and the measured internal key state variables, and determine whether there is an unexpected interaction logic abnormality in the positioning data processing path processing flow.

[0081] In this embodiment, the parallel positioning data processing paths refer to two or more data processing flows that exist or are executed simultaneously, which can be realized by independent instantiation of software modules, multi-threaded processing or virtualization environment, etc., such as running two independent processes on the same server, or creating two logically isolated data stream pipelines in a simulation environment. The reference data path refers to a data stream channel that simulates ideal or known correct data processing logic, which can be realized by using pre-configured, verified algorithm modules or data correction rules, such as removing outliers by data cleaning algorithm or filling missing data by interpolation method. The measured data path refers to a data stream channel that simulates specific data patterns of preset test rules that may be encountered in actual applications, including software modules to be verified or new hardware, which can be realized by injecting specific error patterns, simulating real scene data or integrating software versions to be tested, such as injecting low-precision data with specific markers of new chips.

[0082] The original positioning data stream refers to the continuous positioning information sequence obtained directly from the positioning hardware or sensor without any advanced processing, which can be obtained by serial data interface, network transmission protocol or file reading, such as reading NMEA format data from the UART interface of a Beidou positioning chip. The state marker indicates the data precision state, which can be represented in the form of enumeration value, bit field or numerical range, such as 0 for high precision, 1 for medium precision and 2 for low precision.

[0083] The application provides a Beidou navigation software development data updating method, which realizes in-depth verification of internal processing logic of navigation software by constructing parallel positioning data processing paths, including a reference data path and a measured data path. The system simultaneously inputs original positioning data streams containing multiple positioning data points and state markers into the two paths. In the reference data path, the system generates high-quality reference positioning data streams by correcting or replacing low-precision data points to simulate processing behavior under ideal data input conditions; in the measured data path, the system generates test positioning data streams according to preset test rules to simulate actual complex input conditions.

[0084] During the processing, the system records internal key state variables of the two paths, including but not limited to trustworthiness of different types of positioning information sources, fusion weight, data consistency index and residual characteristics. Among them, the reference internal key state variable represents the system behavior under ideal processing logic; the measured internal key state variable has the same type and is used to reflect the dynamic adjustment response of the system under interference input. The system compares the differences between the two types of state variables to perform failure judgment and identify whether there is an unexpected interaction logic abnormality in the positioning data processing path, such as module coordination disorder, fusion drift, source switching error, etc. This method breaks through the visibility limitation of traditional black box testing, realizes fine verification of the internal robustness and processing logic integrity of the Beidou navigation software, and has significant technical progress and creativity.

[0085] In some preferred embodiments, the application is implemented as follows: a parallel positioning data processing path is built using a software test framework, including two positioning data processing engines configured as a reference mode and a test mode. The system obtains an original positioning data stream from the simulation interface of the vehicle-mounted Beidou positioning module, where each positioning data point is attached with a state marker byte representing the data accuracy state. For positioning data points with low accuracy markers, the system performs correction or replacement processing in the reference data path according to a preset strategy: if the vehicle is in a stationary state, the system replaces the point with the previous valid positioning data point; if the vehicle is in a motion state, the system generates a replacement positioning data point consistent with the vehicle motion trend through Kalman filtering based on the previous valid point and inertial measurement unit data, and obtains a reference positioning data stream. At the same time, the original positioning data stream is processed in the measured data path according to a preset test rule, such as simulating a continuous low-precision data stream input, to observe the response behavior of the fusion module. The system records the reference internal key state variables and the measured internal key state variables respectively during processing, including the fusion weights, trustworthiness, residual values and consistency indicators of different information sources (such as Beidou, IMU, and wheel speed meter), and organizes them into a time series structure. Subsequently, the system calculates the difference value of the paired values of the key variables at the same timestamp, and if the absolute difference value of the Beidou signal weight exceeds the difference threshold value, it is determined that there is an unexpected interaction logic abnormality, and the abnormal time period and variable change trend are recorded.

[0086] Reference Figure 2 In another embodiment of the application, it is further proposed that the step of correcting or replacing low-precision data points in the original positioning data stream includes:

[0087] S3100: Obtain vehicle motion information for indicating the vehicle running state, the vehicle motion information including vehicle speed, acceleration and / or steering angle;

[0088] S3200: Determine whether a target processing data point in the original positioning data stream is a low-precision data point according to a preset accuracy determination rule;

[0089] S3300: If the target processing data point is a low-precision point, determine whether the vehicle is in a stationary state or a motion state according to the vehicle motion information;

[0090] S3400: If the vehicle is in a stationary state, replace the current target processing data point with the previous valid positioning data point;

[0091] S3500: If the vehicle is in a motion state, calculate and generate a replacement positioning data point corresponding to the vehicle motion trend based on the previous valid positioning data point and the vehicle motion information, and correct the current target processing data point with the replacement positioning data point.

[0092] wherein, in the embodiment, the vehicle motion information refers to a parameter set for indicating the vehicle running state, and can specifically include the speed, acceleration or steering angle of the vehicle, etc.; the preset accuracy judgment rule refers to a standard for evaluating the quality of the positioning data point, and can specifically be that whether the data point meets the expected accuracy requirement is judged by setting one or more threshold values, such as based on the horizontal dilution of precision (HDOP), vertical dilution of precision (VDOP), signal-to-noise ratio (SNR) or number of satellites of the positioning data point, etc., and the purpose is to identify the low-accuracy data point that needs to be corrected or replaced; the previous valid positioning data point refers to the positioning data that has been confirmed and accepted by the system as high-accuracy or corrected positioning data before the current target processing data point, and can specifically be the latest data point meeting the accuracy requirement selected from the historical positioning data stream. The substitute positioning data point refers to the estimated position calculated according to the vehicle motion trend, which is used to replace or correct the low-accuracy data point, and can specifically be the estimated position at the current time calculated from the previous valid positioning data point by a motion model (such as a uniform straight line model, a uniform acceleration model or a turning model) combined with the vehicle motion information, and the purpose is to provide a correction value more consistent with the actual trajectory than simple replacement in the vehicle motion state.

[0093] The scheme of the present application constructs a dynamic correction mechanism for low-accuracy data points based on the state flag attached in the original positioning data stream combined with the vehicle motion information (including the vehicle speed, acceleration and / or steering angle). First, the system identifies whether the target processing data point in the original positioning data stream is a low-accuracy data point through the accuracy judgment rule. Then, according to the current state of the vehicle, the system distinguishes between the static state and the motion state, and accordingly performs differentiated correction operations. If the vehicle is in a static state, the system directly replaces the current target processing data point with the previous valid positioning data point to eliminate short-term errors and maintain position stability. If the vehicle is in a motion state, the system generates a substitute positioning data point matching the motion trend based on the last valid positioning point and the vehicle motion information by using Kalman filtering or an equivalent motion model, which is used to correct the current low-accuracy point to ensure the smoothness and continuity of the trajectory. The method realizes the dynamic replacement of low-accuracy positioning data points in the reference data path processing, and can effectively improve the overall accuracy and reliability of the reference positioning data stream. In this way, the high-quality reference data not only provides a positioning performance reference in the ideal input state, but also enhances the sensitivity in the subsequent comparison process with the test positioning data stream, making the failure judgment based on the difference in internal key state variables more accurate, effectively revealing the abnormality of the unexpected interaction logic, and improving the controllability and adaptability of the overall navigation software test method.

[0094] Reference Figure 3 In another embodiment of the present application, it is further proposed that step S5000 comprises:

[0095] S5100: Obtain reference internal processing data generated in a reference data path processing process, the reference internal processing data including timestamps, position solution values, weight distribution values, and fusion residual values of each type of positioning information source;

[0096] S5200: Extract at least one reference internal key state variable from the reference internal processing data based on a preset state extraction rule;

[0097] S5300: Obtain measured internal processing data generated in a measured data path processing process, the measured internal processing data including data fields corresponding to the reference internal processing data;

[0098] S5400: Extract a measured internal key state variable corresponding to a parameter type of the reference internal key state variable from the measured internal processing data based on the state extraction rule.

[0099] In this embodiment, the reference internal processing data refers to original or semi-finished data sets generated internally by the system when positioning data passes through a processing module in the reference data path, used to describe the intermediate state in the processing process. The timestamps of each type of positioning information source refer to specific time markers when different positioning information sources (such as Beidou satellites, inertial measurement units IMU, and wheel speed meters) are generated or received by the system, aiming to ensure time synchronization between different data sources. The position solution values are position results calculated independently by the positioning algorithm at different stages or using different information sources, aiming to reflect real-time estimation of position information by the positioning system in the processing process, and preliminary representation of the contribution of different information sources to the final position. The weight distribution values refer to the trustworthiness or importance coefficients assigned by the system to different positioning information sources in the multi-source positioning information fusion process.

[0100] The fusion residual value refers to the difference or inconsistency between the fusion result and each original positioning information source data after the positioning information fusion, aiming to represent the fitting degree of the fusion model and the consistency of the data. The preset state extraction rule refers to a set of pre-defined logic or algorithms used to filter, calculate, or aggregate key indicators that are representative and can reflect the core behavior or potential abnormalities of the system from the massive reference internal processing data. The measured internal processing data refers to original or semi-finished data sets generated internally by the system when positioning data passes through a processing module in the measured data path, matching the structure and type of the reference internal processing data, aiming to provide comparable internal processing information for the reference data path and lay the foundation for subsequent difference analysis. The measured internal key state variable corresponding to the parameter type of the reference internal key state variable refers to an index that is consistent with the reference internal key state variable in terms of data type, meaning, and calculation method, obtained from the measured internal processing data by applying the same state extraction rule as the reference data path.

[0101] The scheme of the present application systematically obtains internal detailed data of the reference data path and the measured data path in the processing process, including timestamps, position solution values, weight distribution values, and fusion residual values of various types of positioning information sources. Based on the pre-set, unified state extraction rules, representative and comparable reference internal key state variables and measured internal key state variables can be accurately extracted from these detailed data. This ensures that the extracted variables can truly reflect the internal state of the data processing path, and due to the consistency of the reference and measured variables in parameter type and extraction rules, the comparability between them is greatly enhanced. Thus, a reliable, accurate, and highly comparable data basis is provided for subsequent failure judgment based on variable differences, thereby effectively detecting unexpected interaction logic abnormalities in the positioning data processing flow caused by changes in underlying data characteristics, even if these abnormalities are not obvious at the final output level. This detailed recording and extraction method provides a reliable, accurate, and highly comparable data basis for subsequent failure judgment, thereby effectively detecting unexpected interaction logic abnormalities in the positioning data processing flow caused by changes in underlying data characteristics, even if these abnormalities are not obvious at the final output level. This method provides high-quality input for the overall failure judgment process, enabling even subtle internal logic deviations to be effectively identified, solving the challenge of traditional methods being difficult to find hidden problems.

[0102] Reference Figure 4 In another embodiment of the present application, it is further proposed that step S6000 includes:

[0103] S6100: Obtain paired values of the reference internal key state variables and the measured internal key state variables at the same timestamp;

[0104] S6200: Calculate a difference value index for each pair of paired values, the form of the difference value index including absolute difference, normalized offset, and / or residual growth rate;

[0105] S6300: Compare the difference value index with the corresponding pre-set difference threshold value, and the difference threshold value is adaptively adjusted according to the running state of the current vehicle;

[0106] S6400: If the difference value index exceeds the difference threshold value, it is determined that there is an unexpected interaction logic abnormality in the measured data path, and the abnormal time period and the corresponding variable change trend are recorded.

[0107] In the present embodiment, the pair value is a one-to-one variable pair composed of the reference internal key state variable and the measured internal key state variable, which are two key state variable values with the same parameter type extracted from the reference data path and the measured data path respectively at the same time point (timestamp) in the parallel positioning data processing path. The difference value index is a value used to quantify the degree of difference between the reference internal key state variable and the measured internal key state variable, which can be implemented in the form of absolute difference, normalized offset or residual growth rate. The normalized offset is an index that eliminates the influence of dimension and reflects the relative difference by dividing the difference value index of the two variables by the reference value or range of one of the variables, which can be calculated in the form of (measured value-reference value) / reference value or (measured value-reference value) / variable range, etc. The purpose is to provide a difference measure independent of the absolute size of the variable, so that the difference of variables with different dimensions or value ranges can be effectively compared. The residual growth rate is the rate or trend of the residual (i.e. the difference between the measured value and the reference value) over time in a continuous time series, which can be calculated in the form of the ratio of the current residual to the residual at the previous time, or the derivative of the residual sequence. The difference threshold value is a preset limit used to judge whether the difference value index indicates an abnormality, which can be dynamically generated according to empirical values, statistical analysis results or machine learning models, and the purpose is to provide a basis for judging abnormality and to be able to adaptively adjust according to the vehicle operating state to improve the accuracy and robustness of the judgment.

[0108] The unexpected interaction logic anomaly includes internal state changes or incorrect decisions that do not conform to the expected behavior of the upper layer unmodified fusion or processing logic due to changes in the underlying data characteristics or specific input patterns in the positioning data processing path, such as weight allocation errors, data trust reversal, etc. The purpose is to clearly define the type of anomaly to be detected, which is different from simple driver errors or data transmission errors.

[0109] In this embodiment, the system records the reference internal key state variables and the measured internal key state variables respectively based on the constructed parallel localization data processing paths, and extracts paired values, i.e., variable pairs with the same parameter type, for corresponding data points at the same timestamp of the two paths. This pairing mechanism ensures the time synchronization and parameter consistency of comparative analysis. Subsequently, the system calculates difference indicators for each pair of paired values, including but not limited to: absolute difference: reflecting the direct deviation of variable values; normalized offset: eliminating the dimensional influence by dividing by the reference value or fluctuation scale; residual growth rate: measuring the amplification trend of errors between variables over time. Compare the above difference indicators with their respective difference threshold values to determine whether there is an anomaly. Among them, the difference threshold value is dynamically adjusted according to the current vehicle operating state: when the vehicle is in high speed or severe steering state, the system automatically increases the threshold tolerance fluctuation; when the vehicle is in low speed or stable cruising state, the threshold is lowered to enhance the sensitivity of anomaly detection. Once any difference indicator exceeds the adaptive threshold, the system determines that there is an unexpected interaction logic anomaly in the current processing path, and records the corresponding abnormal time period and the key state variable change trend in real time, providing a complete data chain for subsequent analysis and tracking. This scheme realizes high-sensitivity detection of abnormal logic within the navigation software by combining paired value calculation and operating state adaptive mechanism, significantly improves robustness and error tolerance, and is suitable for automated software verification tasks in multiple scenarios.

[0110] In some preferred embodiments, the application is implemented as follows: the system runs the reference data path and the measured data path in parallel, and obtains the reference internal key state variables and the measured internal key state variables at the same timestamp from the two paths respectively. For example, the system records the following paired values at a certain time: the trust degree of the Beidou signal in the reference path is 0.9, the fusion weight is 0.7, and the residual error is 0.5 meters; the corresponding trust degree in the measured path is 0.5, the fusion weight is 0.3, and the residual error is 1.2 meters. Subsequently, the system calculates the difference indicators based on these paired values, including: the absolute difference of the trust degree: |0.5-0.9|=0.4; the normalized offset of the fusion weight: (0.3-0.7) / 0.7≈-0.57; the growth rate of the residual error: if the residual error at the last time is 0.8 meters, then the growth rate is (1.2-0.8) / 0.8=0.5. The system compares the above difference indicators with the preset difference threshold values respectively. The threshold values are adaptively adjusted according to the current vehicle operating state: if the vehicle is in a high-speed or severe turning state, the threshold value is appropriately increased (for example, the residual error growth rate threshold is increased from 0.3 to 0.5); otherwise, in the stable cruising state, the threshold is lowered to improve the detection sensitivity. If any difference indicator exceeds the adaptive threshold value, the system determines that there is an unexpected interaction logic anomaly in the positioning data processing path processing flow, and records the abnormal time period and the change trend of the key variables (such as the fusion weight and the residual error). This mechanism realizes sensitive identification of the internal logic deviation of the measured path, and provides real-time and fine abnormal monitoring capability for software testing.

[0111] Referring to Figure 5 In another embodiment of the application, it is further proposed that the step of adaptively adjusting the difference threshold value according to the current vehicle operating state comprises:

[0112] S6310: obtaining dynamic information of the current vehicle;

[0113] S6320: when the dynamic information indicates that the vehicle is in a high-speed or severe turning state, increasing the difference threshold value according to a preset adjustment rule;

[0114] S6330: when the dynamic information indicates that the vehicle is in a low-speed or stable cruising state, decreasing the difference threshold value according to a preset adjustment rule.

[0115] The dynamic information refers to data reflecting the current motion characteristics of the vehicle, and can be specifically a speed, an acceleration, a steering angle, an angular velocity, or a vehicle attitude of the vehicle, etc. The high-speed or sharp turning state refers to a state in which the vehicle travels at a high speed or a state in which a significant change in direction occurs in a short time, and can be specifically determined by monitoring whether the speed of the vehicle exceeds a preset threshold, or whether the rate of change of the steering angle or the lateral acceleration of the vehicle exceeds a preset threshold. The preset adjustment rule refers to a predetermined algorithm or lookup table for modifying the difference threshold value under a specific vehicle operating state, and can be specifically a function defined to map the vehicle dynamic information to an adjustment factor of the threshold value, or a list of threshold values under different states preset in advance. The weight fluctuation refers to an unexpected change in the trust degree or fusion weight of different positioning information sources (such as Beidou and IMU) caused by data quality or external interference in the process of positioning data fusion, and aims to describe the instability of the internal state of data processing in a dynamic environment. The low-speed or stable cruising state refers to a state in which the vehicle travels at a low speed or a state in which the vehicle maintains a relatively stable straight line or gentle curve motion, and can be specifically determined by monitoring whether the speed of the vehicle is lower than a preset threshold, or whether the acceleration and the rate of change of the steering angle of the vehicle are both lower than a preset threshold. The abnormal weight deviation refers to a deviation of the trust degree or fusion weight of a certain positioning information source from its normal or expected range in the process of positioning data fusion, and the deviation is not caused by normal system logic adjustment.

[0116] The scheme of the present application effectively solves the problem of insufficient accuracy of failure determination of fixed threshold value in dynamic vehicle operating environment by introducing an adaptive adjustment mechanism of difference threshold value. Specifically, when determining failure based on the difference between the benchmark internal key state variable and the measured internal key state variable, first, the dynamic information of the current vehicle is obtained. These dynamic information, such as the speed, acceleration or steering angle of the vehicle, can reflect the running state of the vehicle in real time. Based on these dynamic information, the system can determine whether the vehicle is currently in a high-speed or sharp turning state, or in a low-speed or stable cruising state. When the dynamic information indicates that the vehicle is in a high-speed or sharp turning state, the motion posture of the vehicle changes sharply, and the positioning data is easily affected by dynamic disturbance, causing the weight of the internal key state variable to fluctuate. In this case, if the fixed difference threshold value is followed, the normal dynamic disturbance may be misjudged as an unexpected interaction logic anomaly, resulting in false positives. Therefore, according to the preset adjustment rule, the difference threshold value is increased, thereby providing greater tolerance for weight fluctuations caused by dynamic disturbance, avoiding unnecessary false positives. On the contrary, when the dynamic information indicates that the vehicle is in a low-speed or stable cruising state, the motion posture of the vehicle is relatively stable, and the positioning data is also more stable. At this time, if the difference threshold value is set too high, it may not be able to detect the subtle abnormal weight deviation in time, resulting in missed judgment. Therefore, according to the preset adjustment rule, the difference threshold value is reduced to enhance the detection sensitivity of abnormal weight deviation, ensuring that even a small abnormality can be discovered in time. Through this way of adaptively adjusting the difference threshold value according to the vehicle operating state, the scheme makes the comparison result of the difference index and the difference threshold value more accurately reflect whether there is an unexpected interaction logic anomaly in the positioning data processing path processing flow.

[0117] In some preferred embodiments, the adaptive adjustment of the difference threshold value according to the current operating state of the vehicle can be implemented as follows. First, the dynamic information of the current vehicle is obtained, which can be achieved by the on-board sensor network, for example, the speed information of the vehicle can be obtained from the global positioning system (GPS) receiver, the acceleration and angular velocity information of the vehicle can be obtained from the inertial measurement unit (IMU), and the steering angle information can be obtained from the vehicle bus (CANbus). These information can be sampled and updated at a frequency of several times per second. Specifically, the system can set a series of threshold values to determine the operating state of the vehicle. For example, when the vehicle speed exceeds 80 kilometers per hour, or the absolute value of the lateral acceleration exceeds 2 meters per second squared, or the absolute value of the steering angular velocity exceeds 30 degrees per second, it can be judged that the vehicle is in a high-speed or sharp turning state. In this state, the difference threshold value can be increased according to the preset adjustment rule. For example, the basic difference threshold value can be multiplied by an adjustment coefficient greater than 1, such as 1.5 or 2.0, or a fixed offset can be added to the basic threshold value to tolerate the normal fluctuations of the internal state variables of the positioning data caused by the dynamic changes of the vehicle. On the contrary, when the vehicle speed is lower than 20 kilometers per hour, and the absolute value of the lateral acceleration is lower than 0.5 meters per second squared, and the absolute value of the steering angular velocity is lower than 5 degrees per second, it can be judged that the vehicle is in a low-speed or stable cruising state. In this state, the difference threshold value can be reduced according to the preset adjustment rule. For example, the basic difference threshold value can be multiplied by an adjustment coefficient less than 1, such as 0.8 or 0.5, or a fixed offset can be subtracted from the basic threshold value to enhance the detection sensitivity of the slight abnormal weight offset. These adjustment coefficients or offsets can be calibrated and optimized according to a large amount of actual test data and experience, and stored in the configuration parameters of the system. In this way, the system can dynamically adjust the strictness of the abnormality determination according to the real-time operating conditions of the vehicle, thereby improving the accuracy and reliability of the overall failure determination.

[0118] Referring to Figure 6 In another embodiment of the present application, it is further proposed that step S6000 further comprises:

[0119] S6500: Based on the time series of the measured internal key state variables, identify the abnormal time period in which the difference value indicator continuously exceeds the difference threshold value;

[0120] S6600: Match the variable change pattern in the abnormal time period with a preset abnormal pattern library, the abnormal pattern library including a fusion drift pattern, a residual burst pattern, a trust reversal pattern and / or an information source switching abnormal pattern;

[0121] S6700: Determine the abnormal type of the unexpected interaction logic abnormality according to the matching result, generate the corresponding abnormal classification identifier and add it to the recorded abnormal time period.

[0122] wherein the variable variation pattern refers to the numerical change trend, fluctuation characteristics or statistical distribution characteristics of the measured internal key state variable over time within a certain time period, which can be identified by time series analysis, feature extraction algorithm or machine learning model, and the purpose is to capture the inherent behavior of the variable when the anomaly occurs; wherein the anomaly pattern library refers to a set of pre-defined or learned typical anomaly behavior patterns, which can include a variety of known anomaly types with specific characteristics, such as fusion drift pattern, residual burst pattern, trust reversal pattern and information source switching anomaly pattern, and the purpose is to provide a reference standard for identifying and classifying unexpected interaction logic anomalies; the fusion drift pattern is that when the positioning fusion algorithm processes abnormal data, due to improper weight allocation or data source trust error, the final positioning result deviates from the true trajectory or benchmark result in a persistent and cumulative manner, which can be manifested as a slow but continuous deviation of the positioning point in a certain direction, and the purpose is to represent the long-term stability problem of the fusion logic; the residual burst pattern refers to the rapid increase or violent fluctuation of the residual value of the key state variable within a short time during the positioning calculation or data fusion process, which far exceeds the normal range, which can be manifested as sudden deterioration or jump of the positioning accuracy index, and the purpose is to represent the instantaneous failure of data consistency or model fitting; the trust reversal pattern is that the trust degree or weight allocation logic of the positioning system for different information sources appears unexpected reversal or abnormal adjustment, such as reducing the weight of high-quality data input while increasing the weight of low-quality data input, which can be manifested as abnormal fluctuation or reverse change of the weight allocation curve, and the purpose is to represent the logic error of the information source management strategy; wherein the information source switching anomaly pattern is that when the positioning system switches between different positioning information sources, due to switching logic defects or data connection problems, positioning interruption, jump or non-smooth transition occurs during switching, which can be manifested as obvious breakage or jitter of the positioning trajectory near the switching point, and the purpose is to represent the robustness problem of the multi-source fusion system in dynamic switching scenarios; the anomaly classification identifier refers to a symbol or code used to uniquely identify and mark a specific anomaly type, which can be a string, enumeration value or binary bit, and the purpose is to classify and distinguish the detected unexpected interaction logic anomalies for subsequent analysis, processing and statistics.

[0123] The solution of the present application classifies the unexpected interaction logic anomalies detected in the positioning data processing path in a refined manner. Specifically, based on the preliminary failure determination, the solution further analyzes the time series of the measured internal key state variables, identifies the abnormal time period where the difference index continuously exceeds the difference threshold value, and accurately defines the time range of the anomaly occurrence. Subsequently, the change pattern exhibited by the variables in these abnormal time periods is extracted and matched with a pre-set abnormal pattern library. The abnormal pattern library contains various typical abnormal types, such as fusion drift pattern, residual burst pattern, trust reversal pattern, and information source switching abnormal pattern. Through this pattern matching mechanism, the specific type of unexpected interaction logic anomaly can be determined, and the corresponding abnormal classification identifier is generated and attached to the recorded abnormal time period. This classification mechanism enables the understanding of anomalies to be upgraded from a simple "existence or not" to a "specific type", thereby enabling more targeted diagnostic and processing measures to be taken for different types of anomalies. When combined with the previous steps of constructing parallel positioning data processing paths, recording reference and measured internal key state variables, and failure determination based on differences, the present solution forms a comprehensive testing and diagnosis framework. The parallel path ensures that the behavior of the new software version under disturbed data can be directly compared with the known reference; the recording of internal key state variables provides the required data granularity for in-depth analysis; the preliminary failure determination identifies the existence and time of the anomaly; and the classification mechanism of the present solution further accurately identifies the nature of the anomaly. This layered and cooperative method can detect subtle internal logic errors that may not be apparent in the final output, thereby significantly improving the reliability and stability of the positioning system in complex data interaction scenarios.

[0124] In some preferred embodiments, the application is implemented as follows: the system identifies a time period during which the difference indicator continuously exceeds the difference threshold value as an abnormal time period based on the time series of the measured internal key state variables, wherein a minimum continuous duration threshold value (such as more than 5 consecutive timestamps) is set. The start and end timestamps of the abnormal time period are accurately recorded. Subsequently, the system extracts the change pattern features of the key state variables within the abnormal time period, including the average value, standard deviation, maximum / minimum value, change rate, trend slope, main frequency component (after Fourier transform), and multi-scale wavelet features, etc. The above feature vectors are used to match with the preset abnormal pattern library, which contains labeled feature representations of typical anomalies such as drift pattern, residual burst pattern, trust reversal pattern, and information source switching anomaly pattern. The system determines the abnormal type through a trained classification model (such as support vector machine, random forest, or neural network). The model is trained based on a large amount of historical abnormal data and can automatically identify the abnormal type corresponding to the current abnormal time period. Once a match is successful, the system generates the corresponding abnormal classification identifier (such as "FUSION_DRIFT_001") and attaches it to the abnormal time period, combining feature extraction and machine learning to realize the structuring and interpretable classification of abnormal detection results, providing a reliable data-driven basis for the testing and optimization of Beidou navigation software.

[0125] Reference Figure 7 In another embodiment of the application, it is further proposed that step S6700 comprises:

[0126] S6710: determining the abnormal type of the unexpected interaction logic anomaly according to the matching result;

[0127] S6720: if the abnormal type is a drift pattern or a residual burst pattern, triggering an abnormal segment rejection operation of the test positioning data stream, excluding the data points within the corresponding abnormal time period, and generating the corresponding abnormal response execution result;

[0128] S6730: if the abnormal type is a trust reversal pattern or an information source switching anomaly pattern, marking the current positioning information source as a suspicious information source, and reducing the fusion weight of the suspicious information source or enabling a backup information source in the subsequent data synchronization process, and generating the corresponding abnormal response execution result;

[0129] S6740: attaching the abnormal response execution result and the abnormal classification identifier corresponding to the abnormal type to the abnormal time period to form an abnormal response chain.

[0130] The abnormal segment rejection operation refers to removing data points in the time period identified as abnormal in the test positioning data stream from the data stream to avoid affecting subsequent processing, which can be implemented in the form of data filtering or data truncation. The suspicious information source refers to a positioning information source whose output data or internal state exhibits abnormal characteristics during positioning data processing, which may affect the positioning accuracy or robustness, and it can be marked as a low-reliability source by the system. Reducing the fusion weight of the suspicious information source refers to reducing or limiting the contribution of the information source marked as suspicious to the final positioning result in the multi-source positioning data fusion algorithm, which can be implemented by adjusting the weight coefficient.

[0131] Enabling a backup information source refers to switching to a pre-configured other positioning information source for positioning when the main positioning information source is marked as suspicious or fails, which can be implemented using an information source priority switching mechanism.

[0132] The abnormal response execution result refers to the output record of the processing measures taken for a specific abnormal type, which can include processing actions, processing time, affected data range, etc.

[0133] The scheme identifies the unexpected interaction logic exception in the positioning data processing path, executes differentiated response strategies based on the abnormal type, and records the response results associated with the abnormal information to build a complete abnormal response chain. Specifically, after identifying the abnormal time period and determining the abnormal type (such as fusion drift, residual burst, trust reversal, information source switching exception), the system takes corresponding response measures according to the nature of the abnormality.

[0134] If the abnormal type is fusion drift or residual burst, it indicates that the positioning data has been significantly distorted in the abnormal time period, and the system performs an abnormal segment rejection operation on the test positioning data stream to shield the data points in that time period, preventing abnormal data from interfering with subsequent processing, and generates a response execution result detailing the rejection range and execution basis.

[0135] If the abnormal type is trust reversal or information source switching exception, it indicates that there is a problem with the weight logic of individual positioning information sources during fusion, and the system will mark the information source as a suspicious information source and reduce its fusion weight or enable a backup information source to replace it in subsequent processing. The corresponding source adjustment operation and strategy selection are recorded as the response execution result.

[0136] Finally, all response execution results and abnormal classification identifiers are recorded together in the abnormal time period to form a traceable response record chain. This mechanism not only improves the intelligent intervention ability of the system to abnormality, but also enhances the closed-loop manageability of the processing process, significantly improving the robustness and reliability of the Beidou navigation software in complex data interaction scenarios.

[0137] In some embodiments, the application is implemented as follows: after identifying the type of unexpected interaction logic anomaly, the scheme executes a difference response strategy based on the anomaly type, and associates the response result with the anomaly information for record, forming a traceable anomaly response chain. Specifically, when the anomaly type is fusion drift mode or residual burst mode, the system considers that the test positioning data stream has significant deviation or residual sharp increase in the abnormal time period, so the data processing module receives the start and end timestamps of the abnormal time period, logically screens or physically removes all data points (such as position coordinates, speed information, timestamps) in the corresponding time period of the test positioning data stream, or marks them as invalid, to avoid their interference with subsequent positioning calculation. Subsequently, an anomaly response execution result is generated, recording the type, time range, data point quantity and execution module information of the removal operation.

[0138] If the anomaly type is trust reversal mode or information source switching anomaly mode, the system judges that the fusion strategy of a certain positioning information source is abnormal, such as a low-quality signal being given too high a weight, or a jump in positioning output when the information source is switched. At this time, the information source management module marks the information source as "suspicious", and dynamically adjusts its fusion weight (such as from 0.8 to 0.2 or to 0) or enables a backup information source (such as an inertial navigation system) in the subsequent fusion process. The corresponding adjustment action forms another anomaly response execution result, recording the information source name, adjustment content and timestamp. Finally, all anomaly response execution results are recorded together with the anomaly classification identifier in the metadata of the abnormal time period, forming a complete response chain, providing accurate and structured data support for subsequent system diagnosis and optimization.

[0139] Reference Figure 8 In another embodiment of the application, it is further proposed that the method further comprises:

[0140] S6750: Count the number and proportion of data segments removed due to unexpected interaction logic anomalies in the test positioning data stream, and generate an anomaly removal index for evaluating positioning robustness;

[0141] S6760: Analyze the frequency and associated time period of each type of anomaly in the test positioning data stream, and construct a version-level anomaly distribution spectrum;

[0142] S6770: Update the test evaluation result of the software version corresponding to the measured data path based on the anomaly removal index and the anomaly distribution spectrum.

[0143] The abnormality rejection index refers to quantitative data for evaluating positioning robustness, and can be the total duration of the rejected data segment, the proportion of the rejected data points in the total data points, or the average number of rejected data segments per unit time, and the purpose is to intuitively reflect the influence degree of unexpected interaction logic abnormality on the integrity and availability of the test positioning data stream. The version-level abnormality distribution spectrum refers to the comprehensive statistical and analysis results of the occurrence of each type of abnormality in the test positioning data stream of a specific software version, which can be a report or chart containing the occurrence frequency, duration, and relevance or clustering effect of different abnormality types (such as fusion drift mode, residual burst mode, etc.) on the time axis, and the purpose is to fully reveal the performance characteristics and potential defect modes of the software version under different abnormality scenarios. The test evaluation result refers to the conclusion of the comprehensive performance and stability evaluation of the software version corresponding to the measured data path, which can be a report containing robustness score, abnormality type distribution overview, key performance indicator (KPI) compliance, and potential risk prompt, and the purpose is to provide data support for the release decision or further optimization of the software version.

[0144] After the identification, classification and response of unexpected interaction logic abnormality are completed, the present scheme further quantitatively analyzes the abnormality in the test positioning data stream, so as to realize the comprehensive update of the test evaluation result of the software version. Specifically, the system first counts the number of data segments rejected due to abnormal response operations in the test positioning data stream and its proportion in the overall data stream, and generates an abnormality rejection index for evaluating positioning robustness, which reflects the data retention capability and processing robustness of the software in the abnormality handling process. Subsequently, the system classifies and counts the identified abnormality types, analyzes the occurrence frequency, distribution law and time overlap relationship between each type of abnormality in the data stream, and constructs a version-level abnormality distribution spectrum for revealing the dominant mode and potential causal relationship of the abnormality types.

[0145] In some preferred embodiments, the application is implemented as follows: after the identification and classification of unexpected interaction logic anomalies are completed, the system starts the data post-processing module to statistically analyze the excluded data segments in the test positioning data stream and generate quantitative indicators for software version evaluation. First, traverse the entire test positioning data stream, identify all data segments marked as "excluded", and record the number of excluded data segments and the total number of data points they contain. Calculate the anomaly exclusion index (such as the data exclusion rate), which is the proportion of the number of excluded data points to the total number of data points, to measure the data retention capability and robustness of the software under abnormal interference. At the same time, maintain an anomaly type statistics table to count the frequency of all identified anomaly types (such as fusion drift pattern, residual burst pattern, trust reversal pattern, information source switching anomaly pattern), and identify the temporal association of different anomaly types based on the anomaly timestamp sequence. For example, when the trust reversal pattern frequently appears after the residual burst pattern, the system can mark it as a high correlation anomaly chain. Based on frequency statistics and temporal correlation analysis, the system generates a version-level anomaly distribution spectrum to show the occurrence proportion, distribution pattern and potential causal chain of each type of anomaly, which can be visualized in the form of bar charts, line charts or time axis charts. Finally, combined with the anomaly exclusion index and the anomaly distribution spectrum, the test evaluation results of the current software version are updated according to the preset version evaluation model. For example, if the data exclusion rate exceeds the set threshold or high-risk anomalies occur frequently, the system can trigger automatic degradation and generate a risk prompt to realize automated quantitative feedback and continuous optimization support for software quality.

[0146] Reference Figure 9 In another embodiment of the application, it is further proposed that the method further comprises:

[0147] S6780: During the processing of the measured data path, a source mapping relationship is established for each positioning data point of the original positioning data stream, and the original positioning information type, timestamp and correction path identifier of the positioning data point are recorded;

[0148] S6790: Based on the source mapping relationship, a path tree structure of the positioning data point is constructed, and the path nodes of the path tree structure record the processing stage, state marker, correction method and variable value;

[0149] S67100: When the failure determination result is that there is an unexpected interaction logic anomaly, the formation process of the positioning data point is traced and located based on the path tree structure, and the key variable evolution chain related to the anomaly response chain is extracted;

[0150] S67110: The key variable evolution chain is used as the backtracking basis for processing the measured internal key state variables.

[0151] The source mapping relationship refers to the association record established for each original positioning data point, which can trace the original information source and processing history, which can be realized by using the metadata field, association table or log entry in the data structure, and the purpose is to provide basic information for subsequent data tracing. The correction path identifier refers to the unique identifier of each correction operation or processing link experienced by the data point in the positioning data point processing process, which can be realized by using string encoding, enumeration value or bit mask, and the purpose is to clarify which specific correction is applied to the data point in the processing flow. The path tree structure refers to a hierarchical data organization form used to represent the evolution path of the positioning data point from the original state to the final processing result, which can be realized by using a multi-way tree, a binary tree or a graph structure. The path node refers to an independent unit in the path tree structure, representing the state snapshot of the positioning data point at a specific processing stage, which can be realized by using a data object or structure containing multiple attributes, and the purpose is to record the key information of the data point at each processing link in detail. The key variable evolution chain refers to the value sequence of the core variable directly related to the abnormality of the unexpected interaction logic at different processing stages extracted from the path tree structure, which can be realized by using an ordered list, time series data or state machine transition record, and the purpose is to reveal the dynamic change rule of the key variable in the abnormality occurrence process. The backtracking basis refers to the reference information used to guide the analysis, diagnosis and optimization of the processing logic of the measured internal key state variable, which can be realized by using a log analysis report, a visualization chart or an automatic diagnosis rule set, and the purpose is to provide data support for system debugging and performance improvement.

[0152] In the measured data path processing, the source mapping relationship of each positioning data point in the original positioning data stream is established, the original positioning information type, timestamp and correction path identifier are recorded, and a traceable identity file is constructed. Based on this, the path tree structure is further generated, the processing state, correction method and key variable value of each processing stage are recorded in the node, and a complete "growth trajectory diagram" is formed. When the failure determination identifies the unexpected interaction logic abnormality, the system can rely on the path tree structure to trace the abnormal data point, and extract the key variable evolution chain as the backtracking basis of the measured internal key state variable. For example, if the "trust degree reversal mode" is identified, the specific information source and the correction behavior of the processing stage can be traced, and the key path node causing the trust degree change is identified. This mechanism can deeply associate the abnormal response chain with the processing flow of the positioning data point, realize the microscopic level of abnormal cause analysis, and improve the accuracy and depth of software abnormal diagnosis.

[0153] In some embodiments, the present scheme establishes a source mapping relationship for each original positioning data point during the processing of the measured data path, and the mapping content includes the original positioning information type (such as "GPS" or "BDS"), the collection timestamp (such as the Unix timestamp), and the correction path identifier (such as the bit mask "0x01" indicating filtering and "0x04" indicating outlier rejection). Based on the mapping relationship, the system constructs a path tree structure for each data point, and records its processing stage, state marker, correction method, and key variable value at each stage. For example, the data point is marked as "filtered" in the "filtering" stage, the correction method is "Kalman filtering", and the pseudorange value after processing is recorded; if it is subsequently determined to be "low precision", the residual error and other variable values that cause the anomaly are further recorded. If the failure determination result shows that there is an unexpected interaction logic anomaly, the system performs source tracing on the abnormal data point based on the path tree structure, extracts its key variable evolution chain, and uses it as the basis for backtracking the processing of the measured internal key state variable. For example, in the "trust reversal mode", the system can backtrack and analyze the trust degree, fusion weight, and position solution value at each stage, find that the trust degree of Beidou information decreases at a certain time and the IMU weight abnormally increases, and then locate the logic defect of the fusion module. This mechanism realizes a closed-loop support from anomaly determination to root cause tracing, improves the granularity and traceability of the positioning data processing path anomaly analysis, and enhances the tunability and robustness of software development.

[0154] Reference Figure 10 The present application further proposes a Beidou navigation software development data updating system, which comprises:

[0155] A path construction module 1 is configured to construct a parallel positioning data processing path, and the positioning data processing path comprises a reference data path and a measured data path.

[0156] A data acquisition module 2 is configured to acquire an original positioning data stream, and the original positioning data stream comprises a plurality of positioning data points and corresponding state markers, and the state markers indicate the data precision state.

[0157] A reference data processing module 3 is configured to process the original positioning data stream through the reference data path: correct or replace the low-precision data points in the original positioning data stream to obtain a reference positioning data stream.

[0158] A test data processing module 4 is configured to process the original positioning data stream through the measured data path: test the original positioning data stream according to a preset test rule to obtain a test positioning data stream.

[0159] The state variable recording module 5 is configured to record the reference internal key state variables of the reference data path processing process and the measured internal key state variables of the measured data path processing process, the reference internal key state variables are used to represent the trust degree, fusion weight, data consistency index and / or residual feature of different types of positioning information sources in the reference data processing process, and the measured internal key state variables are consistent with the reference internal key state variables in terms of parameter types and are used to represent the dynamic adjustment response in the processing logic under the interference of low-precision data.

[0160] The judgment module 6 is configured to perform failure judgment based on the difference between the reference internal key state variables and the measured internal key state variables, and determine whether there is an unexpected interaction logic abnormality in the positioning data processing path processing flow.

[0161] The system implementation of the present application converts the abstract method steps into specific and executable modules, so that the data updating and testing methods which are only at the theoretical level can be efficiently run and have reliability. Through parallel processing and fine comparison of internal state variables, the system can deeply reach the internal logic level of the software, find hidden problems that are difficult to be detected by traditional black box testing, for example, the error weight allocation or trust degree reversal of the upper fusion module under specific low-precision data input, so as to improve the coverage of the testing in the development process of the Beidou navigation software and the accuracy of problem positioning.

[0162] The above only describes the embodiments of the present application and is not used to limit the protection scope of the present application. For those skilled in the art, the present application can have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for updating data in BeiDou navigation software development, characterized in that, The method includes: Construct parallel positioning data processing paths, which include a reference data path and a measured data path. Acquire the raw positioning data stream, which contains multiple positioning data points and corresponding status markers, the status markers indicating the data accuracy status; The original positioning data stream is processed using the reference data path: low-precision data points in the original positioning data stream are corrected or replaced to obtain the reference positioning data stream. The original positioning data stream is processed according to the test data path: the original positioning data stream is tested according to preset test rules to obtain the test positioning data stream; The reference internal key state variables and the measured internal key state variables of the measured data path processing process are recorded. The reference internal key state variables are used to characterize the trust in different types of positioning information sources, fusion weight, data consistency index and / or residual characteristics during the reference data path processing process. The parameter types of the measured internal key state variables are consistent with those of the reference internal key state variables and are used to characterize the dynamic adjustment response in the processing logic under low-precision data interference. Failure determination is performed based on the difference between the benchmark internal key state variables and the measured internal key state variables to determine whether there are any unexpected interaction logic anomalies in the location data processing path.

2. The BeiDou navigation software development data update method according to claim 1, characterized in that, Correcting or replacing low-precision data points in the original positioning data stream includes: Acquire vehicle motion information to indicate the vehicle's operating status, the vehicle motion information including vehicle speed, acceleration, and / or steering angle; Based on the preset accuracy judgment rules, determine whether the target processing data points in the original positioning data stream are low-precision data points; If the target data point is a low-precision point, the vehicle is determined to be in a stationary or moving state based on the vehicle motion information. If the vehicle is stationary, the current target processing data point will be replaced with the previous valid positioning data point. If the vehicle is in motion, then based on the previous valid positioning data point and the vehicle motion information, an alternative positioning data point corresponding to the vehicle motion trend is calculated and generated, and the current target processing data point is corrected using the alternative positioning data point.

3. The BeiDou navigation software development data update method according to claim 1, characterized in that, The baseline internal key state variables and the measured internal key state variables of the measured data path processing process are recorded, including: Acquire the internal processing data of the benchmark generated during the benchmark data path processing process. The internal processing data of the benchmark includes the timestamps, location calculation values, weight allocation values, and fusion residual values ​​of each type of positioning information source. Based on preset state extraction rules, at least one key state variable within the benchmark is extracted from the internal processed data of the benchmark. Obtain the internal processing data generated during the processing of the data path under test, wherein the internal processing data under test includes data fields corresponding to the benchmark internal processing data; Based on the state extraction rules, the key internal state variables corresponding to the parameter types of the benchmark key internal state variables are extracted from the tested internal processing data.

4. The BeiDou navigation software development data update method according to claim 3, characterized in that, Failure determination is performed based on the differences between the benchmark internal key state variables and the measured internal key state variables, determining whether there are any unexpected interaction logic anomalies in the positioning data processing path, including: Obtain the paired values ​​of the benchmark internal key state variable and the measured internal key state variable at the same timestamp; For each pair of said paired values, a difference index is calculated, which may include, in the form of, absolute difference, normalized offset, and / or residual growth rate. The difference index is compared with the corresponding preset difference threshold value, which is adaptively adjusted according to the current operating status of the vehicle. If the difference index exceeds the difference threshold, it is determined that there is an unexpected interaction logic anomaly in the tested data path, and the abnormal time period and the corresponding variable change trend are recorded.

5. The BeiDou navigation software development data update method according to claim 4, characterized in that, The difference threshold value is adaptively adjusted according to the current operating status of the vehicle, including: Obtain the current dynamic information of the vehicle; When the dynamic information indicates that the vehicle is in a high-speed or violent change of direction state, the difference threshold value is increased according to the preset adjustment rule. When the dynamic information indicates that the vehicle is in a low-speed or stable cruising state, the difference threshold value is reduced according to a preset adjustment rule.

6. The BeiDou navigation software development data update method according to claim 4, characterized in that, Failure determination is performed based on the difference between the benchmark internal key state variables and the measured internal key state variables. This includes determining whether there are any unexpected interaction logic anomalies in the positioning data processing path. Based on the time series of the measured internal key state variables, identify abnormal time periods in which the difference index continuously exceeds the difference threshold. The variable change pattern within the abnormal time period is matched with a preset abnormal pattern library, which includes fusion drift pattern, residual burst pattern, trust reversal pattern and / or information source switching abnormal pattern. Based on the matching results, determine the exception type of the unexpected interaction logic exception, generate the corresponding exception classification identifier, and add it to the recorded exception time period.

7. The BeiDou navigation software development data update method according to claim 6, characterized in that, Based on the matching results, the exception type of the unexpected interaction logic is determined, a corresponding exception classification identifier is generated and added to the recorded exception time period, including: The exception type of the unexpected interaction logic exception is determined based on the matching results; If the anomaly type is a fusion drift mode or a residual burst mode, the anomaly segment removal operation of the test positioning data stream is triggered to exclude the data points within the corresponding anomaly time period and generate the corresponding anomaly response execution result. If the anomaly type is a trust reversal mode or an information source switching anomaly mode, then the current location information source is marked as a suspicious information source, and the fusion weight of the suspicious information source is reduced or a backup information source is enabled during subsequent data synchronization, and a corresponding anomaly response execution result is generated. The execution result of the abnormal response and the abnormal classification identifier corresponding to the abnormal type are appended and recorded in the abnormal time period to form an abnormal response chain.

8. The BeiDou navigation software development data update method according to claim 7, characterized in that, The method further includes: The number and percentage of data segments that were removed from the test positioning data stream due to unexpected interaction logic anomalies were statistically analyzed, and an anomaly removal index was generated to evaluate the robustness of positioning. Analyze the frequency and associated time periods of various anomaly types in the test location data stream to construct a version-level anomaly distribution spectrum; Based on the anomaly removal index and the anomaly distribution spectrum, update the test evaluation results of the software version corresponding to the tested data path.

9. The BeiDou navigation software development data update method according to claim 7, characterized in that, The method further includes: During the processing of the tested data path, a source mapping relationship is established for each location data point in the original location data stream, and the original location information type, timestamp, and corrected path identifier of the location data point are recorded. Based on the source mapping relationship, a path tree structure for the location data points is constructed. The path nodes of the path tree structure record the processing stage, status marker, correction method, and variable value. When the failure determination result indicates that there is an unexpected interaction logic anomaly, the formation process of the location data point is traced according to the path tree structure, and the evolution chain of key variables related to the anomaly response chain is extracted. The evolution chain of the key variables is used as the backtracking basis for processing the internal key state variables under test.

10. A Beidou navigation software development data update system, characterized in that, The system includes: A path construction module is used to construct parallel positioning data processing paths, which include a reference data path and a measured data path. The data acquisition module is used to acquire the raw positioning data stream, which contains multiple positioning data points and corresponding status markers, and the status markers indicate the data accuracy status. The reference data processing module is used to process the original positioning data stream using the reference data path: correcting or replacing low-precision data points in the original positioning data stream to obtain the reference positioning data stream; The test data processing module is used to process the raw positioning data stream according to the test data path: the raw positioning data stream is tested according to preset test rules to obtain the test positioning data stream; The state variable recording module is used to record the benchmark internal key state variables of the benchmark data path processing process and the measured internal key state variables of the measured data path processing process. The benchmark internal key state variables are used to characterize the trust level, fusion weight, data consistency index and / or residual characteristics of different types of positioning information sources during the benchmark data path processing process. The parameter types of the measured internal key state variables are consistent with those of the benchmark internal key state variables and are used to characterize the dynamic adjustment response in the processing logic under low-precision data interference. The judgment module is used to determine failure based on the difference between the internal key state variables of the benchmark and the internal key state variables of the tested entity, and to determine whether there is any unexpected interaction logic abnormality in the processing flow of the positioning data processing path.

Images