Vehicle-mounted Beidou positioning deviation self-calibration system fused with inertial navigation information

Abstract

The invention discloses a vehicle-mounted Beidou positioning deviation self-calibration system fused with inertial navigation information, and relates to the technical field of navigation and positioning. The method is used for solving the problems of positioning deviation accumulation and signal failure in a complex environment. The system realizes space-time alignment of inertial navigation and Beidou observation data through timestamp interpolation and coordinate system conversion, and constructs a multi-source synchronous data stream. A difference value between an inertial navigation calculation speed and a Beidou Doppler speed is analyzed based on a vehicle incomplete constraint model, and a weight factor is generated in combination with a confidence threshold to suppress abnormal observation. A geometric matching error is calculated through a high-precision lane map curvature and a vehicle steering angle, and candidate tracks are screened to generate transverse calibration parameters. The noise covariance is dynamically adjusted through a tight coupling filtering algorithm, the mode is switched to a lane constraint calibration mode when a Beidou signal is interrupted, inertial navigation errors are compensated through error boundary constraint, and stable output is achieved. Multi-source data deep fusion and adaptive robust calibration are realized, and system robustness and positioning continuity are improved.

Description

Technical Field

[0001] The present invention relates to the field of navigation and positioning technology, and in particular to a vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information. Background Art

[0002] With the rapid development of smart cars, vehicle-road collaboration and intelligent transportation systems, the high-precision positioning capability of vehicles during driving has become the basic support for realizing functions such as lane-level assisted driving, automatic parking, path planning and active safety control. Especially in complex traffic environments such as urban elevated roads, tunnels, underground garages, and high-density urban areas, traditional positioning methods are difficult to ensure continuity and accuracy, and are prone to problems such as signal loss, jumps or error accumulation, which seriously affect the operation safety and control effect of vehicles. Therefore, there is an urgent need for a technical solution that can integrate information from multiple sensors and achieve stable, reliable and high-precision positioning to meet the urgent needs of intelligent transportation systems for positioning continuity, accuracy and robustness.

[0003] At present, the commonly used vehicle positioning technology mainly relies on the combined application of satellite navigation system and inertial navigation system. Common combinations include loose combination and tight combination. In the loose combination structure, the position result provided by the satellite positioning system is used as an external input to correct the inertial navigation error. This method is highly dependent on satellite signals. Once the signal is blocked, the positioning error accumulates rapidly; the tight combination structure uses the original observation values such as pseudorange or carrier phase of the satellite and the inertial data for joint solution. Although the fusion accuracy is improved to a certain extent, there are still problems such as insufficient sensor error modeling and insufficient use of motion constraints in practical applications. In addition, the existing methods generally ignore the non-complete kinematic constraints of the vehicle during actual road driving, as well as the lane geometry information contained in the high-precision map, and fail to effectively introduce them into the error analysis and compensation process. As a result, in the scenario where the satellite signal is unavailable or severely degraded, the system lacks an effective positioning calibration mechanism, and there are technical bottlenecks such as poor stability and fast error drift. There is an urgent need for a highly robust vehicle positioning system that integrates vehicle kinematic constraints and road geometry characteristics and has adaptive error control capabilities. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information, which solves the problems of the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: A vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information, comprising the following modules: a multi-source data synchronization and alignment module, a kinematic constraint consistency analysis module, a lane geometry constraint analysis module, and an adaptive fusion and calibration module; The multi-source data synchronization and alignment module is used to perform time-series synchronization based on timestamp interpolation on the three-axis acceleration and three-axis angular velocity data output by the inertial navigation system, as well as the pseudorange observation value and carrier phase observation value output by the Beidou satellite navigation system, and achieve spatial reference alignment through coordinate system conversion, and output a multi-source data stream with a unified spatio-temporal reference; The kinematic constraint consistency analysis module is used to calculate the difference between the speed estimated by the inertial navigation system and the speed calculated by the Beidou system based on Doppler frequency shift according to the vehicle non-holonomic constraint model. When the difference exceeds the dynamic confidence threshold, a corresponding data credibility weight factor is generated; The lane geometry constraint analysis module is used to call the lane curvature information in the high-precision lane-level map, combine the vehicle's real-time steering angle data, calculate the matching error between the lane geometry constraint and the vehicle kinematic model, screen the candidate trajectories that meet the matching error threshold, and output the lateral position calibration parameter and error boundary; The adaptive fusion and calibration module is used to fuse multi-source data based on the tightly coupled filtering algorithm when the vehicle-mounted Beidou signal is available, dynamically adjust the noise covariance weight of the Beidou observation value through the weight factor, and output a robust fusion positioning result. When the vehicle-mounted Beidou signal is continuously lost for more than a preset threshold, it switches to the lane geometry constraint-driven calibration mode, suppresses the inertial navigation cumulative error based on the candidate trajectory and error boundary, and outputs a continuously stable position information.

[0006] Further, the time-series synchronization of the multi-source data synchronization and alignment module includes the following contents: Perform timestamp interpolation alignment on the high-frequency three-axis acceleration and angular velocity data output by the inertial navigation system, as well as the low-frequency pseudorange and carrier phase observation values output by the Beidou satellite navigation system, to eliminate the time-series mismatch caused by the sampling rate difference; Through the dynamic delay compensation mechanism, correct the fixed delay and random jitter during the data transmission process to ensure the time-series consistency of the multi-source data stream; Perform conflict detection on the interpolated time-series sequence, identify abnormal segments with reversed timestamps and excessive intervals, and perform local resampling repair through adjacent valid data segments.

[0007] Further, the specific process of calculating the difference between the inertial navigation system's estimated speed and the speed calculated by the Beidou system based on Doppler frequency shift according to the vehicle non-holonomic constraint model is as follows: According to the Ackermann steering geometry relationship, a theoretical constraint equation for the lateral speed is constructed, and the theoretical value of the lateral speed estimated by the inertial navigation system is derived; the speed calculated by the Beidou Doppler frequency shift is subjected to a rotation transformation from the vehicle body coordinate system to the navigation coordinate system, and the antenna installation angle error is compensated; the confidence interval of the vehicle acceleration is introduced, and the confidence threshold of the speed difference is dynamically softened: the threshold is expanded in the case of rapid acceleration or braking scenarios, and tightened in the uniform speed scenario.

[0008] Further, when the difference exceeds the dynamic confidence threshold, the specific process of generating the corresponding data credibility weight factor is as follows: Design a weight decay curve based on the S-shaped function. When the difference exceeds the threshold, the weight factor decreases non-linearly with the overrun ratio, avoiding the weight jump caused by hard decision; introduce a time decay factor to exponentially decay and weight historical abnormal events, ensuring that short-term anomalies do not permanently reduce the credibility of the data source; adaptively adjust the weight recovery rate according to the road type, accelerating the recovery of the Beidou data weight in the curve scene and preferentially trusting the inertial data in the straight road scene.

[0009] Further, the specific process of calling the lane curvature information in the high-precision lane-level map and combining the vehicle's real-time steering angle data to calculate the matching error between the lane geometric constraint and the vehicle kinematic model is as follows: Extract the discrete curvature points of the lane centerline from the high-precision map, construct a continuous curvature function through multiple spline interpolations, and calculate its change rate; according to the vehicle's real-time steering angle and longitudinal vehicle speed, inversely deduce the theoretical lane curvature through the kinematic model, and introduce a tire side slip stiffness compensation term to improve the model accuracy; take the matching error between the actual curvature and the theoretical curvature as the main term of the error function, and map the vehicle's lateral offset to the curvature error gain coefficient to weight and correct the error function; increase the penalty weight for the trajectory near the lane edge to enhance the model's response ability to boundary constraints.

[0010] Further, the specific process of screening candidate trajectories that meet the matching error threshold and outputting the lateral position calibration parameter and error boundary is as follows: Segment the track generated by inertial navigation estimation through a sliding window mechanism to construct a candidate trajectory cluster; calculate the weight matching error for each candidate trajectory and sort based on probability scoring, and screen the non-dominated solution trajectory cluster through the Pareto front; apply the Gaussian mixture model to the selected preferred trajectory cluster for clustering analysis, and extract the lateral position mean and covariance matrix of the clustering center trajectory as the calibration parameter and confidence interval.

[0011] Further, when the on-vehicle Beidou signal is available, the specific process of fusing multi-source data based on the tightly coupled filtering algorithm, dynamically adjusting the noise covariance weight of the Beidou observation value through the weight factor, and outputting the robust fusion positioning result is as follows: Construct a joint state vector including the biases of inertial devices, scale factor errors, and the clock error of the Beidou receiver. Explicitly model the temperature drift term of the inertial device in the state vector and establish a covariance correlation matrix with the Beidou clock error term; Map the weight factor to the independent noise covariance components of the Beidou pseudorange and carrier phase observation values, downweight the low-confidence Beidou observation values, and achieve heterogeneous downweighting of multiple observation sources; Construct a cost function through a robust M-estimator, perform reweighted iteration on the filtering residuals, and suppress the influence of multipath abnormal observations.

[0012] Further, when the on-vehicle Beidou signal is continuously lost for more than a preset threshold, switch to the calibration mode driven by lane geometric constraints. The specific process of suppressing the cumulative error of inertial navigation based on the candidate trajectory and error boundary is as follows: In the initial stage of Beidou signal failure, generate a lane-bound corridor based on the historical positioning result and high-precision map to constrain the lateral divergence of inertial dead reckoning; Through the improved residual resampling particle filter, inject lane curvature constraints into the proposal distribution to force the particles to be distributed along the reasonable lane centerline; Inject the calibrated bias estimate value back into the tightly coupled filter state vector to achieve a smooth mode switch when the signal is restored and avoid positioning jumps.

[0013] The present invention has the following beneficial effects: (1) A vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information can achieve precise alignment of the output data of the inertial navigation system and the Beidou system in terms of time sequence and spatial reference by setting a multi-source data synchronization and alignment module and a kinematic constraint consistency analysis module, improving the basic consistency before multi-source information fusion; At the same time, dynamically analyze the speed calculation difference based on the vehicle nonholonomic constraint model, generate a weight factor in combination with the confidence judgment mechanism, effectively suppress the interference of abnormal data on the fusion result, and enhance the system's discrimination ability for the reliability of navigation data in abnormal states.

[0014] (2) A vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information can establish a spatial matching constraint relationship by setting a lane geometric constraint analysis module and an adaptive fusion and calibration module, using a high-precision lane-level map and real-time steering angle data, to achieve accurate estimation and constraint correction of the vehicle's lateral deviation; When the Beidou signal is available, the system fuses multi-source observation information based on the tightly coupled filtering algorithm and improves the robust performance through dynamic weight regulation; In the case of continuous loss of the Beidou signal, it automatically switches to the calibration mode driven by geometric constraints, which can effectively limit the cumulative error of inertial navigation and ensure the continuity and stability of the positioning result in complex environments.

[0015] Of course, it is not necessary for any product implementing the present invention to simultaneously achieve all the above-mentioned advantages. Brief Description of the Drawings

[0016] Figure 1 It is a flowchart of a vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information according to the present invention.

[0017] Figure 2 It is a flowchart of the adaptive fusion and calibration module according to the present invention. Detailed Embodiments

[0018] In the embodiments of the present application, a vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information is adopted to solve the problems of the existing vehicle-mounted Beidou positioning system, including the decline in positioning accuracy in complex environments, the inability to effectively suppress the error accumulation of the inertial navigation system, and the lack of a dynamic credibility evaluation mechanism. The spatio-temporal reference consistency is achieved through the multi-source data synchronization and alignment module, enhancing the basis for multi-source observation data fusion; the dynamic detection and confidence quantification of speed deviation are realized through the kinematic constraint consistency analysis module, improving the ability to identify abnormal data; the lateral position correction is carried out by combining the lane geometry constraint analysis module with high-precision maps and vehicle state data, enhancing the vehicle path constraint accuracy; finally, the dynamic robust fusion of multi-source information and the failure switching strategy are realized through the adaptive fusion and calibration module, significantly enhancing the continuity and reliability of the vehicle-mounted positioning system in scenarios where satellite signals are weakened or lost.

[0019] The overall idea of the solution in the embodiments of the present application is as follows: First, through the timing synchronization module, the interpolation synchronization based on timestamps is performed on the three-axis acceleration and three-axis angular velocity data output by the inertial navigation system (INS), and the pseudorange observation values and carrier phase observation values output by the Beidou satellite navigation system (BDS).

[0020] Then, using the coordinate transformation method, the inertial navigation data and Beidou satellite navigation data are unified to the same spatial reference, thereby outputting a multi-source data stream with a unified spatio-temporal reference, preparing for subsequent processing.

[0021] According to the non-holonomic constraint model of the vehicle, the difference between the speed estimated by the inertial navigation system and the speed calculated by the Beidou system based on Doppler frequency shift is calculated.

[0022] When the speed difference between the two exceeds the dynamic confidence threshold, a corresponding data credibility weight factor is generated. This weight factor will dynamically adjust the weights of the observation values in the subsequent data fusion process to ensure that data with higher credibility has a greater impact on the system output.

[0023] Call the lane curvature information in the high-precision lane-level map, and combine the real-time steering angle data of the vehicle to calculate the matching error between the lane geometric constraint and the vehicle kinematic model.

[0024] According to the calculated matching error, filter out the candidate trajectories that meet the matching error threshold, and output the corresponding lateral position calibration parameters and error boundaries to provide support for subsequent calibration.

[0025] When the on-vehicle Beidou signal is available, fuse multi-source data based on the tightly coupled filtering algorithm. At this time, by dynamically adjusting the weight factor, the noise covariance weight of the Beidou observation value is dynamically adjusted based on the calculated credibility weight factor, so as to output a robust fused positioning result.

[0026] When the loss of the on-vehicle Beidou signal exceeds the preset threshold, the system will automatically switch to the calibration mode driven by lane geometric constraints. At this time, by using the candidate trajectory and error boundary to suppress the cumulative error of the inertial navigation system, the stability and continuity of the positioning result are ensured.

[0027] Please refer to Figure 1 , an embodiment of the present invention provides a technical solution: a vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information, including the following modules: a multi-source data synchronization and alignment module, a kinematic constraint consistency analysis module, a lane geometric constraint analysis module, and an adaptive fusion and calibration module; the multi-source data synchronization and alignment module is used to perform time-series synchronization based on timestamp interpolation on the three-axis acceleration and three-axis angular velocity data output by the inertial navigation system, and the pseudorange observation value and carrier phase observation value output by the Beidou satellite navigation system, and realize spatial reference alignment through coordinate system transformation, and output a multi-source data stream with a unified spatio-temporal reference; the kinematic constraint consistency analysis module is used to calculate the difference between the speed estimated by the inertial navigation system and the speed calculated by the Beidou system based on Doppler frequency shift according to the vehicle non-holonomic constraint model, and generate the corresponding data credibility weight factor when the difference exceeds the dynamic confidence threshold; the lane geometric constraint analysis module is used to call the lane curvature information in the high-precision lane-level map, combine the real-time steering angle data of the vehicle, calculate the matching error between the lane geometric constraint and the vehicle kinematic model, filter out the candidate trajectories that meet the matching error threshold, and output the lateral position calibration parameters and error boundaries; the adaptive fusion and calibration module is used to fuse multi-source data based on the tightly coupled filtering algorithm when the on-vehicle Beidou signal is available, dynamically adjust the noise covariance weight of the Beidou observation value through the weight factor, and output a robust fused positioning result. When the on-vehicle Beidou signal is continuously lost for more than the preset threshold, it switches to the calibration mode driven by lane geometric constraints, suppresses the inertial navigation cumulative error based on the candidate trajectory and error boundary, and outputs continuous and stable position information.

[0028] In this implementation plan, the multi-source data synchronization and alignment module: The core task of this module is to uniformly process the observation data from different sensor systems (inertial navigation system and Beidou satellite navigation system) in terms of time and space. Since the sampling frequencies of each sensor are different and there are clock biases, it is necessary to achieve time series synchronization through an interpolation method based on timestamps to align each data under the same time reference. At the same time, since each system operates in a different coordinate system (for example, the inertial navigation system often uses the vehicle body coordinate system, while the Beidou system uses the geographic coordinate system or the geocentric coordinate system), this module realizes the unification of the spatial reference through coordinate transformation. Three-axis acceleration and three-axis angular velocity: They are respectively the accelerometer and gyroscope data obtained by the inertial measurement unit (IMU) in the inertial navigation system, which are the basic inputs for calculating the vehicle attitude and speed; Pseudorange observation value: The propagation time between the receiver and the satellite measured by the Beidou receiver multiplied by the speed of light, which contains various error components; Carrier phase observation value: Compared with the pseudorange, it has higher measurement accuracy and is used for precise positioning. Kinematic constraint consistency analysis module: This module establishes a vehicle kinematic inference model according to the physical characteristics of the vehicle, that is, the "nonholonomic constraints model". By comparing the difference between the speed estimated by the inertial navigation system and the speed measured by the Beidou system through Doppler frequency shift, the consistency of the data of the two systems is evaluated. Once the difference exceeds the preset dynamic confidence threshold, it indicates that there may be anomalies or errors, and this module will generate a data credibility weight factor for dynamic weighting in the subsequent fusion process. Nonholonomic constraints model: It is often used to describe the actual motion constraints of a vehicle that can only move along the forward direction and is restricted in lateral motion; Doppler frequency shift resolved speed: Based on the satellite signal frequency shift, it measures the radial speed of the receiver relative to the satellite, reflecting the actual driving speed of the vehicle; Weight factor: It is used to adjust the influence degree of data on the fusion result according to the data reliability degree during the multi-source information fusion process. Lane geometric constraint analysis module: This module uses the detailed road geometric information (such as curvature, width, etc.) in the high-precision lane-level map (HD Map), and combines the vehicle's real-time steering angle data to perform geometric constraint matching on the vehicle's driving trajectory. By constructing a geometric error model, it calculates the deviation between the actual motion trajectory of the vehicle and the lane model, screens out the candidate trajectories with errors within a reasonable range, and then outputs the lateral position calibration parameters and error boundaries to achieve the lateral compensation for the position drift of the inertial navigation system. High-precision lane-level map: It has a higher resolution than the traditional navigation map and can be accurate to the lane line level; Steering angle data: It comes from the steering wheel angle sensor or the front wheel steering angle sensor; Error boundary: It represents the confidence interval within which the trajectory error is acceptable and is used for subsequent fusion judgment. Adaptive fusion and calibration module: This module is the fusion core of the entire system.Under normal Beidou signal conditions, the tight coupling filter algorithm is used to fuse and estimate inertial navigation and Beidou data. The noise covariance matrix of Beidou observations is dynamically adjusted according to the aforementioned weight factors to achieve robust estimation and output high-precision positioning results. When the Beidou signal is continuously lost for more than a threshold (such as in an occlusion scenario), the system automatically switches to an error suppression mode based on lane geometry constraints to suppress the cumulative error of the inertial navigation system with the candidate trajectory and error boundary, ensuring the continuous and stable output of position information. Tight coupling filter algorithm: Perform joint state space estimation between inertial navigation and GNSS observation information, without relying on external preliminary positioning; Noise covariance weight adjustment: Adjust the contribution degree of different observation sources to the estimated state according to their reliability; Robust fusion: Even if some observations are abnormal, the overall estimation can remain stable and not be dominated by outliers; Cumulative error: The integration drift problem caused by the long-term operation of the inertial navigation system, which usually increases with time accumulation.

[0029] Specifically, the time synchronization of the multi-source data synchronization and alignment module includes the following: For the high-frequency three-axis acceleration and angular velocity data output by the inertial navigation system, and the low-frequency pseudorange and carrier phase observations output by the Beidou satellite navigation system, perform timestamp interpolation alignment to eliminate the timing mismatch caused by sampling rate differences; Through the dynamic delay compensation mechanism, correct the fixed delay and random jitter during data transmission to ensure the timing consistency of multi-source data streams; Perform conflict detection on the interpolated timing sequence, identify abnormal segments with reversed timestamps and excessive intervals, and repair them through local resampling of adjacent valid data segments.

[0030] In this implementation scheme, timestamp interpolation alignment: The sampling frequencies of the three-axis acceleration and three-axis angular velocity data output by the Inertial Navigation System (INS) are relatively high (usually hundreds of Hz or higher), while the sampling frequencies of the pseudorange and carrier phase observations output by the BeiDou Satellite Navigation System (BDS) are relatively low (generally several Hz or lower). Due to the sampling rate difference, when directly comparing these two types of data, a timing mismatch will occur, resulting in incorrect fusion. Operation: Through the timestamp interpolation algorithm, the high-frequency inertial navigation data and the low-frequency BeiDou satellite data are unified to the same time reference. This is achieved by filling the gaps between the lower-frequency data through interpolation methods (such as linear interpolation or spline interpolation) to ensure that there are corresponding values at the same time point for both, eliminating the timing deviation caused by the sampling frequency difference. Dynamic delay compensation: When data is transmitted from sensors (such as inertial navigation systems, satellite navigation systems, etc.) to the data fusion system, certain transmission delays will occur. These delays can be fixed or may be affected by random jitter. Operation: By introducing a dynamic delay compensation mechanism, the transmission delays of data from different sources are corrected. Fixed delay refers to the conventional delay caused by the system architecture or communication link, while random jitter is the delay variation caused by the uncertainty of the communication network or the sensor itself. This mechanism can estimate and compensate for these delays based on real-time measurements and historical data, thus ensuring the timing consistency of multi-source data streams. Conflict detection and repair: Due to various reasons (such as clock deviation or sensor failure), timestamps may be in reverse order or have abnormal time intervals. For example, the sensor may experience data loss or confusion at certain moments, resulting in reverse timestamps or overly long intervals. Operation: Conflict detection is performed on the interpolated timing data to identify possible abnormal situations, such as reverse timestamps (i.e., timestamps not arranged in ascending order) or interval overrun (i.e., abnormally large data time intervals). Once such conflicts are detected, the system will handle them through the method of local resampling repair, that is, using the values of adjacent valid data segments to compensate or resample the abnormal data segments to restore them to a reasonable time series state. Through this mechanism, it can be ensured that the finally output data stream is continuous and free of abnormalities in terms of timing.

[0031] Specifically, the specific process of calculating the difference between the inertial navigation system's estimated speed and the speed calculated by the BeiDou system based on Doppler frequency shift according to the vehicle non-holonomic constraint model is as follows: According to the Ackermann steering geometry relationship, a theoretical constraint equation for the lateral speed is constructed, and the theoretical value of the lateral speed estimated by the inertial navigation system is derived; the speed calculated by the BeiDou Doppler frequency shift is subjected to a rotation transformation from the vehicle body coordinate system to the navigation coordinate system, and the antenna installation angle error is compensated; the confidence interval of the vehicle acceleration is introduced, and the confidence threshold of the speed difference is dynamically softened: the threshold is expanded in the case of rapid acceleration or braking scenarios, and tightened in the uniform speed scenario.

[0032] In this implementation, a theoretical constraint equation for the lateral velocity is constructed: Based on the Ackermann steering geometry relationship, the lateral velocity can be derived from the longitudinal velocity and the steering angle . For a vehicle with a steering system, the lateral velocity is determined by the curve radius of the vehicle's travel and the vehicle's steering angle. The formula is as follows: Assume that the steering angle and the longitudinal velocity of the vehicle are known. The lateral velocity can be calculated through the following more complex formula: ; where: : The lateral velocity of the vehicle. : The longitudinal velocity of the vehicle. : The wheelbase of the vehicle (the distance between the wheels). : The turning radius of the vehicle. : The steering angle of the vehicle. Coordinate system conversion: The Beidou satellite navigation system and the inertial navigation system use different coordinate systems. It is necessary to convert the velocity solved by Beidou from the satellite coordinate system to the vehicle's navigation coordinate system. Generally, the velocity solved by Beidou is based on the inertial coordinate system of the earth, while the inertial navigation system is based on the vehicle's coordinate system. Therefore, a rotation matrix is needed to convert the coordinate system. The formula is as follows: The velocity of the Beidou system is converted from the satellite coordinate system to the vehicle coordinate system through the rotation matrix : ; where: : The velocity in the vehicle navigation coordinate system. : The velocity solved by Beidou Doppler. : The coordinate system rotation matrix, is the rotation angle between the coordinate systems. Antenna mounting angle error compensation: The antenna mounting angle error will affect the velocity solved by the Beidou system. Therefore, it is necessary to compensate for the antenna mounting error according to the known antenna mounting error angle . This compensation process corrects the velocity data solved by Beidou through a rotation matrix. The formula is as follows: If the antenna mounting error angle is considered, the compensated velocity can be calculated through the following formula: ; where: : The compensated velocity. : The velocity in the vehicle navigation coordinate system. : The antenna mounting angle error. : The velocity components of the vehicle on the x and y axes. Dynamic softening process: According to the vehicle's acceleration , dynamically adjust the confidence threshold of the speed difference. When the vehicle is in a state of rapid acceleration or rapid braking, a larger speed difference is allowed, so the threshold will be expanded; while in a uniform speed state, the threshold will be tightened to enhance accuracy. This dynamic adjustment helps to adapt to different driving scenarios. The formula is as follows: According to the acceleration of the vehicle and the acceleration threshold , the confidence threshold can be dynamically adjusted through the following formula : ; where: : The speed difference threshold after dynamic adjustment. : The basic speed difference threshold. : The threshold adjustment coefficient, which is used to determine the amplitude of threshold expansion. : The acceleration of the vehicle. : The acceleration threshold, which is used to determine whether the vehicle is in a state of rapid acceleration or rapid braking.

[0033] Specifically, when the difference exceeds the dynamic confidence threshold, the specific process of generating the corresponding data credibility weight factor is as follows: Design a weight decay curve based on the S-shaped function. When the difference exceeds the threshold, the weight factor decreases non-linearly with the excess ratio, avoiding the weight jump caused by hard decision; Introduce a time decay factor to exponentially decay and weight historical abnormal events to ensure that short-term anomalies do not permanently reduce the credibility of the data source; Adaptively adjust the weight recovery rate according to the road type, accelerate the recovery of the Beidou data weight in the curve scene, and give priority to trusting the inertial data in the straight road scene.

[0034] In this implementation plan, the S-shaped function is usually used to describe a smooth non-linear change process, which is widely used in control systems, especially suitable for scenarios that require smooth transition such as weight factors. Its output value range is usually between 0 and 1, which makes it suitable for adjusting the weight factor to avoid sudden changes caused by hard decision. Design a weight decay curve based on the S-shaped function: When the speed difference exceeds the dynamic confidence threshold, the weight factor will decay non-linearly according to the S-shaped function. Through the gradual change property of the S-shaped function, the sudden change caused by hard decision is avoided, ensuring the smooth change of data credibility. The formula is as follows: The weight factor decays using the S-shaped function with the change of the excess ratio of the difference . The specific formula is: ; where: : The current time 's weight factor. : The ratio of the current speed difference exceeding the threshold. : The dynamic confidence threshold, and the exceeded difference will trigger decay. : A parameter that controls the attenuation rate. The larger the value, the faster the attenuation. Introduction of the time decay factor: To prevent historical abnormal events from permanently affecting the credibility of the data source, the time decay factor is designed. The weights of historical abnormal events will decay exponentially over time, making the more recent abnormal events have a greater impact on data credibility, while the impact of long - ago abnormal events on credibility decreases. Formula representation: The weighting of historical anomalies can be corrected by the following formula: ; where: : The current time is the historical weight factor of : The historical weight factor of the previous time point. : The exponential decay rate, which controls the speed at which historical abnormal events decay. : The time difference from the occurrence of the historical abnormal event to the current moment. Adaptive adjustment of the weight recovery rate according to road type: Different road scenarios have different requirements for weight recovery. In the curve scenario, the inertial navigation of the vehicle may generate large errors, so it is necessary to accelerate the recovery of the weight of Beidou data; while in the straight - road scenario, the inertial navigation data is more reliable, and inertial data should be trusted first. Therefore, the recovery rate is adjusted according to the road type. Formula representation: According to different road types, the weight recovery rate will be adjusted. Assume that the road type is represented by , where represents a curve, represents a straight road. The adjustment formula for the weight recovery rate is: ; where: : The weight recovery rate at the current moment of : The basic recovery rate, which is used for the straight - road scenario. : The adjustment coefficient, which determines the acceleration degree of the recovery rate in the curve scenario. : The road type indicator (curve or straight road). Generating the final weight factor by combining the above factors: The final weight factor is the result of the combined action of the above several factors. First, calculate the current weight decay factor based on the S - shaped function, then adjust the decay according to historical anomalies, and finally adjust the recovery rate according to the road type to finally obtain the comprehensive weight factor. Formula representation: The final weight factor can be calculated by the following formula: ; where: : The finally adjusted weight factor. : The decay factor based on the S - shaped function. : The decay factor based on historical anomalies. : The recovery rate based on the road type.

[0035] Specifically, the specific process of calculating the matching error between the lane geometric constraint and the vehicle kinematic model by invoking the lane curvature information in the high-precision lane-level map and combining with the vehicle's real-time steering angle data is as follows: Extract the discrete curvature points of the lane centerline from the high-precision map, construct a continuous curvature function through multiple spline interpolations, and calculate its rate of change; According to the vehicle's real-time steering angle and longitudinal vehicle speed, inversely deduce the theoretical lane curvature through the kinematic model, and introduce a tire side slip stiffness compensation term to improve the model accuracy; Take the matching error between the actual curvature and the theoretical curvature as the main term of the error function, and map the vehicle's lateral offset to a curvature error gain coefficient to weight and correct the error function; Increase the penalty weight for the trajectory near the lane edge to enhance the model's response ability to boundary constraints.

[0036] In this implementation plan, extract the discrete curvature points of the lane centerline from the high-precision map. The high-precision map contains detailed geometric information of the lane, including the discrete curvature points of the lane centerline. We extract the curvature information from these curvature points and construct a continuous curvature function through the spline interpolation method. Construction of the curvature function: The curvature of the lane centerline is a measure of the degree of curve bending, where represents the arc length parameter on the curve. Through multiple spline interpolations, a smooth and continuous curvature function can be constructed: ; where: is the weight coefficient of the spline interpolation; is the number of nodes (number of discrete points), representing the total number of discrete curvature points used for interpolation, that is, the refinement degree of the entire lane centerline; is the spline basis function; is the arc length coordinate of the vehicle traveling along the lane. Inversely deduce the theoretical lane curvature based on the vehicle's real-time steering angle and longitudinal vehicle speed. According to the vehicle's steering angle data and longitudinal vehicle speed , we deduce the theoretical lane curvature of the vehicle through the vehicle's kinematic model . The basic relationship in the vehicle kinematic model is: ; where: is the steering angle of the vehicle; is the wheelbase of the vehicle. This theoretical lane curvature assumes an ideal vehicle steering process, where describes the curvature expected by the vehicle. Introduce a tire side slip stiffness compensation term. To improve the accuracy of the vehicle kinematic model, it is necessary to consider the lateral stiffness effect of the tire, that is, the tire will experience side slip during steering, resulting in a deviation between the theoretical curvature and the actual situation. The side slip stiffness compensation term can be expressed by the following formula: ; where: is the tire side slip stiffness compensation coefficient; is the lateral deviation angle of the vehicle, representing the angle by which the vehicle currently deviates from the ideal trajectory. This compensation enables the curvature calculation to better conform to the actual vehicle motion. Calculate the matching error between the actual curvature and the theoretical curvature. By comparing the actual curvature with the theoretical curvature we can calculate the matching error between the lane geometric constraint and the vehicle kinematic model: ; This error represents the degree of matching between the actual driving trajectory of the vehicle and the center line of the lane. Map the vehicle's lateral offset to the curvature error gain coefficient. The vehicle's lateral offset also affects the curvature error. The relationship between the lateral offset and the curvature error can be adjusted by the gain coefficient . The calculation formula for the gain coefficient is: ; where: is the adjustment coefficient; is the lateral offset of the vehicle on the lane. When the vehicle deviates from the center of the lane, the gain coefficient will decrease, thereby increasing the influence of the lateral offset on the curvature error. Weighted correction of the error function. The error function will be weighted and corrected according to the lateral offset , thereby dynamically adjusting the response to the error: ; Penalty weight and boundary constraint. To improve the response ability of the lane geometric constraint to the boundary, we introduce a penalty weight into the error function. Specifically, when the vehicle approaches the lane edge, the weight is increased to make the boundary constraint more stringent. Define the boundary constraint weight as: ; When the vehicle approaches the lane edge, the weight increases, thereby increasing the response to the boundary. The final error function. The final error function is: ; This error function combines the influences of lane geometric constraints, vehicle kinematic models, tire cornering stiffness, lateral offset, and boundary constraints, and is used to measure the matching error between the vehicle and the lane.

[0037] Specifically, the process of screening candidate trajectories that meet the matching error threshold and outputting the lateral position calibration parameters and error boundaries is as follows: Segment the track generated by inertial navigation calculation through a sliding window mechanism to construct a candidate trajectory cluster; Calculate the weight matching error for each candidate trajectory and sort them based on probability scoring, and screen the non-dominated solution trajectory cluster through the Pareto front; Apply the Gaussian mixture model to the selected preferred trajectory cluster for clustering analysis, and extract the lateral position mean and covariance matrix of the clustering center trajectory as calibration parameters and confidence intervals.

[0038] In this implementation, the track is segmented through a sliding window mechanism. First, the track generated by the inertial navigation system is segmented, and the sliding window mechanism is used to divide the track. Set a window size and a sliding step . Slide on the track according to the window size and step, so as to obtain a number of candidate track segments. Each candidate track segment can be expressed as: ; where: are discrete points on the track, representing the lateral and longitudinal positions at time ; is the number of points within the window, which determines the length of each track segment. Calculate the matching error and probability score for each candidate track segment , and calculate the matching error according to the lane geometry constraint and the kinematic model of the vehicle. The matching error can be calculated based on the degree of deviation of the vehicle from the center line of the lane: ; where: is the lane curvature in the lane geometry information; is the theoretical lane curvature calculated based on the kinematic model of the vehicle. Then, according to the matching error, calculate the probability score of each candidate track. The probability score can be obtained by normalizing the matching error: ; where: is a positive constant used to adjust the influence of the matching error on the score; is the matching error of the track segment . Screen the non-dominated solution track clusters based on the Pareto front. Sort all candidate tracks according to the probability score , and select those track clusters that satisfy the Pareto front. The Pareto front refers to the set of solutions that are not dominated by other solutions under multiple objective functions. In this step, the selection of track clusters is based on two criteria: matching error and probability score. Through sorting, select those track clusters that perform best in terms of matching error and score to generate non-dominated solution track clusters : ; Apply the Gaussian mixture model for clustering analysis. Apply the Gaussian mixture model (GMM) to the screened non-dominated solution track clusters for clustering analysis. The Gaussian mixture model is a probability model based on multiple Gaussian distributions, used to identify multiple potential clusters from data. Assume that the distribution of each candidate track cluster is a Gaussian distribution, and the EM algorithm (expectation maximization algorithm) is used to estimate the mean and covariance of each cluster. Let be the set of screened non-dominated solution track clusters, where each represents the state vector of a track. The probability density function of the Gaussian mixture model can be expressed as: ; where: is the weight of the th Gaussian distribution; is the th Gaussian distribution with a mean of and a covariance of ; is the set of parameters of the model; is the number of Gaussian distributions. Through the clustering analysis of the Gaussian mixture model, the clustering center (i.e., the mean) and the covariance matrix of each candidate trajectory cluster can be obtained, so as to extract the mean and covariance of the lateral position as calibration parameters and confidence intervals. Output the lateral position calibration parameters and error bounds. The mean of the lateral position of the clustering center trajectory and the covariance matrix can be used as calibration parameters for the lateral position, and the covariance matrix can be used to represent the confidence interval of the lateral position: ; ; Through these parameters, a calibration value for the lateral position can be provided for each trajectory, and a confidence interval can be provided for position calibration in subsequent steps.

[0039] Please refer to Figure 2 , specifically, when the on-vehicle Beidou signal is available, multi-source data is fused based on the tightly coupled filtering algorithm, and the noise covariance weight of the Beidou observation value is dynamically adjusted through the weight factor. The specific process of outputting the robust fusion positioning result is as follows: construct a joint state vector including the zero bias of the inertial device, scale factor error, and Beidou receiver clock error, explicitly model the temperature drift term of the inertial device in the state vector, and establish a covariance correlation matrix with the Beidou clock error term; map the weight factor to the independent noise covariance components of the Beidou pseudorange and carrier phase observation values, downweight the low-confidence Beidou observation values, and achieve heterogeneous downweighting of multiple observation sources; construct a cost function through the robust M estimator, perform reweighted iteration on the filtering residuals, and suppress the influence of multipath abnormal observations.

[0040] In this implementation scheme, constructing the joint state vector and explicitly modeling the temperature drift term First, construct the joint state vector, which includes the zero bias (bias) of the inertial sensor, scale factor error, and the clock error of the Beidou receiver. The joint state vector x can be expressed as: ; where: represents the zero bias of the inertial sensor (gyroscope zero bias and accelerometer zero bias ); represents the clock error of the Beidou receiver; represents the temperature drift term. The covariance correlation matrix describes the error relationship between each state quantity: ; Mapping of Weight Factors and Adjustment of Noise Covariance Next, map the weight factors to the noise covariance matrix of Beidou pseudorange and carrier phase observations. The weight factors are dynamically adjusted according to the credibility of Beidou observations, and the observations with low credibility will have their weights reduced. Assume the noise covariance of Beidou pseudorange is , and the noise covariance of carrier phase is . Then the covariance after adjustment based on the weight factor is: ; ; where: and represent the original noise covariances of pseudorange and carrier phase respectively; is the dynamically adjusted weight factor, which changes with the credibility of Beidou signals. The dynamic adjustment of the weight factor can be carried out according to the actual Beidou signal quality. For example, when the signal is weak or the multipath effect is strong, the weight factor will decrease, thereby reducing its impact on the filtering result. Downweighting of Low-Credibility Observations Low-credibility Beidou observations will be dynamically downweighted. Specifically, low-credibility observations will be downweighted through the weight factor to reduce their impact on the final fusion result. At this time, the weight factor will be adjusted according to the following rules: ; where: is the adjustment factor used to adjust the relationship between signal quality and weight; is the error of Beidou observations (the deviation estimated through the multipath effect). Reweighted Iteration Using Robust M-Estimator To further suppress the impact of multipath effects and abnormal observations, a robust M-estimator is used to perform reweighted iteration on the filtering residuals. The M-estimator defines a cost function to weight the residuals , thereby reducing the impact of abnormal observations. The cost function can take forms such as the Huber function or the Tukey function: ; where: is the filtering residual; is a constant that determines the saturation threshold of the cost function. In each iteration, the impact of large residual values is suppressed by reweighting the filtering residuals. The final filtering result is optimized through weighted least squares.

[0041] Specifically, when the on-vehicle Beidou signal is continuously lost for more than a preset threshold, it switches to the calibration mode driven by lane geometry constraints. The specific process of suppressing the cumulative error of inertial navigation based on the candidate trajectory and the error boundary is as follows: In the initial stage of the Beidou signal failure, a lane-bound corridor is generated based on the historical positioning results and the high-precision map to constrain the lateral divergence of inertial dead reckoning; through an improved residual resampling particle filter, lane curvature constraints are injected into the proposal distribution to force the particles to be distributed along the reasonable lane centerline; the calibrated zero-bias estimated value is injected back into the tightly coupled filter state vector to achieve a smooth mode switch when the signal is restored, avoiding positioning jumps.

[0042] In this implementation plan, the lane-bound corridor is generated in the initial stage of the Beidou signal failure through the historical positioning results and the high-precision map data. The lane-bound corridor is a constraint area that restricts lateral divergence and is usually defined by the lane centerline and the lane width. Assume that the discrete points of the lane centerline are , where each point represents the coordinates of the lane center, and the lane width is , then the lane-bound corridor can be expressed as: ; where: is the z-th discrete point of the lane centerline, representing the coordinate point of the lane center on the two-dimensional plane. is the width of the lane. represents the constructed lane-bound corridor area, which is a set of lateral constraints for the vehicle's inertial dead reckoning position. represents the offset of a certain point on the lane centerline in the lateral direction, and its range is limited to [−w / 2, w / 2] to ensure that the offset point is still within the lane range. represents a set of ordered pairs composed of two real numbers, that is, the Euclidean two-dimensional plane, indicating that is a two-dimensional position point. At this time, the lateral divergence of the inertial navigation system will be constrained by the lane corridor to ensure that the vehicle travels along the lane as much as possible during the reckoning process, reducing the lateral deviation. Residual resampling through an improved particle filter To further suppress the error accumulation of the inertial navigation system, an improved residual resampling particle filter is adopted. This filtering method injects lane curvature constraints into the proposal distribution to force the particles to be distributed along the reasonable lane centerline. The core idea of the particle filter algorithm is to estimate the state of the system based on the position and weight of each particle. Here, each particle represents a possible position of the vehicle, and the update process can be described through the following steps: Particle state: The particle represents the state vector of the -th particle at the current moment k, including position and speed information. Prediction step: Predict the particle state according to the kinematic model of the vehicle: ; where: is the kinematic model of the vehicle; is the control input (such as steering angle, vehicle speed, etc.); is the process noise. Injecting lane curvature constraint: Inject lane curvature information into the predicted particle state to ensure that the particles are distributed along the center line of the lane. Assume the lane curvature is , and the lane geometry constraint needs to be satisfied during particle update: ; Resampling step: Select more likely particles through weighted resampling and eliminate particles that do not meet the constraints. Finally, after multiple resampling and update steps, the particles will gather near reasonable lane positions, suppressing the cumulative error of the inertial navigation system. Reverse injection of the calibrated zero bias estimate value During the calibration process, the zero bias estimate value of the inertial navigation system is reversely injected into the state vector of the tightly coupled filter to provide a corrected state estimate. Assume the state vector of the tightly coupled filter is , and the zero bias estimate value is , then the calibrated state vector is: ; During this process, the zero bias estimate value represents the bias in the inertial navigation system caused by sensor errors or other factors. Smooth mode switching When the Beidou signal is restored, the system needs to smoothly switch from the lane geometry constraint mode back to the inertial navigation system mode. At this time, through the smooth switching mechanism of the state vector, the system is prevented from experiencing a positioning jump during the switch. The transition of the smooth mode can be represented by the following formula: ; where: represents the state vector after sliding, which finally fuses the estimates of the inertial navigation system and the lane geometry constraint. α is the smoothing coefficient, which controls how the system smoothly transitions; is the calibrated inertial navigation state vector. When the Beidou signal is restored and the positioning accuracy is high, the value of α will gradually increase, transitioning to the original filtering result of the inertial navigation system; when the quality of the Beidou signal is poor, a low value of α is maintained, and the positioning continues to rely on the lane geometry constraint.

[0043] In summary, the present application has at least the following effects: A vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information effectively integrates inertial navigation and Beidou observation data through spatio-temporal reference alignment and tightly coupled filtering algorithms, improving positioning accuracy and continuity in complex scenarios. Based on kinematic constraints and dynamic adjustment of weight factors, it suppresses the interference of abnormal observation values and enhances the robustness of the system under signal occlusion or multipath effects. Through lane geometry constraints and error boundary analysis, it corrects inertial navigation drift using a high-precision map when the Beidou signal is lost, significantly reducing long-term positioning deviation. Combining real-time signal quality and vehicle motion state, it autonomously switches fusion strategies and calibration modes to ensure stable output in extreme environments such as urban canyons and tunnels. Through an incomplete constraint model and candidate trajectory screening mechanism, it reduces redundant computational workload and achieves efficient real-time processing on a vehicle-mounted embedded platform.

[0044] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can be in the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can be in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0045] The present invention is described with reference to the flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in Figure 1 one or more of the processes or multiple processes and / or blocks Figure 1 one or more of the blocks or multiple blocks.

[0046] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured product including an instruction device, and the instruction device implements the functions specified in Figure 1 one or more of the processes or multiple processes and / or blocks Figure 1 one or more of the blocks or multiple blocks.

[0047] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions for implementing the process Figure 1 one process or a plurality of processes and / or blocks Figure 1 steps for implementing the functions specified in one block or a plurality of blocks.

[0048] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn of the basic inventive concept. Therefore, the appended claims are intended to be construed to cover the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0049] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information, characterized in that, It includes the following modules: multi-source data synchronization and alignment module, kinematic constraint consistency analysis module, lane geometric constraint analysis module, and adaptive fusion and calibration module; The multi-source data synchronization and alignment module is used to perform time-series synchronization based on timestamp interpolation on the three-axis acceleration and three-axis angular velocity data output by the inertial navigation system, as well as the pseudorange observation value and carrier phase observation value output by the Beidou satellite navigation system, and achieve spatial reference alignment through coordinate system transformation, and output a multi-source data stream with a unified spatio-temporal reference; The kinematic constraint consistency analysis module is used to calculate the difference between the estimated speed of the inertial navigation system and the speed calculated by the Beidou system based on Doppler frequency shift according to the vehicle non-holonomic constraint model. When the difference exceeds the dynamic confidence threshold, a corresponding data credibility weight factor is generated; The lane geometric constraint analysis module is used to call the lane curvature information in the high-precision lane-level map, combine the vehicle's real-time steering angle data, calculate the matching error between the lane geometric constraint and the vehicle kinematic model, screen the candidate trajectories that meet the matching error threshold, and output the lateral position calibration parameter and error boundary; The adaptive fusion and calibration module is used to fuse multi-source data based on the tightly coupled filtering algorithm when the on-vehicle Beidou signal is available, dynamically adjust the noise covariance weight of the Beidou observation value through the weight factor, and output the robust fusion positioning result. When the on-vehicle Beidou signal is continuously lost for more than the preset threshold, it switches to the lane geometric constraint-driven calibration mode, suppresses the inertial navigation cumulative error based on the candidate trajectory and error boundary, and outputs a continuous and stable position information.

2. The vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information according to claim 1, characterized in that: The time-series synchronization of the multi-source data synchronization and alignment module includes the following contents: Perform timestamp interpolation alignment on the high-frequency three-axis acceleration and angular velocity data output by the inertial navigation system, as well as the low-frequency pseudorange and carrier phase observation values output by the Beidou satellite navigation system, and eliminate the time-series mismatch caused by the sampling rate difference; Through the dynamic delay compensation mechanism, correct the fixed delay and random jitter during the data transmission process to ensure the time-series consistency of the multi-source data stream; Perform conflict detection on the interpolated time-series sequence, identify abnormal segments with reversed timestamps and excessive intervals, and perform local resampling repair through adjacent valid data segments.

3. The vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information according to claim 2, characterized in that: The specific process of calculating the difference between the estimated speed of the inertial navigation system and the speed calculated by the Beidou system based on Doppler frequency shift according to the vehicle non-holonomic constraint model is as follows: According to the Ackermann steering geometry relationship, construct the theoretical constraint equation of the lateral speed, and deduce the theoretical value of the lateral speed estimated by the inertial navigation system; Perform a rotation transformation of the speed calculated by the Beidou Doppler frequency shift from the carrier coordinate system to the navigation coordinate system, and compensate for the antenna installation angle error; Introduce the vehicle acceleration confidence interval to perform dynamic softening processing on the confidence threshold of the speed difference: Expand the threshold in the case of rapid acceleration or braking scenarios, and tighten the threshold in the case of uniform speed scenarios.

4. The vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information according to claim 3, characterized in that: The specific process of generating a corresponding data credibility weight factor when the difference exceeds the dynamic confidence threshold is as follows: Design a weight decay curve based on the S-shaped function. When the difference exceeds the threshold, the weight factor decreases non-linearly with the exceeding ratio, avoiding the weight jump caused by hard decision; Introduce a time decay factor to exponentially decay-weight historical abnormal events, ensuring that short-term anomalies do not permanently reduce the credibility of the data source; Adaptively adjust the weight recovery rate according to the road type, accelerating the recovery of the Beidou data weight in the curve scenario and preferentially trusting the inertial data in the straight road scenario.

5. The vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information according to claim 4, wherein: The specific process of calling the lane curvature information in the high-precision lane-level map and combining the vehicle's real-time steering angle data to calculate the matching error between the lane geometric constraint and the vehicle kinematic model is as follows: Extract the discrete curvature points of the lane centerline from the high-precision map, construct a continuous curvature function through multiple spline interpolations, and calculate its change rate; According to the vehicle's real-time steering angle and longitudinal vehicle speed, inversely deduce the theoretical lane curvature through the kinematic model, and introduce a tire side slip stiffness compensation term to improve the model accuracy; Take the matching error between the actual curvature and the theoretical curvature as the main term of the error function, and map the vehicle's lateral offset to a curvature error gain coefficient to weight and correct the error function; Increase the penalty weight for the trajectory close to the lane edge to enhance the model's response ability to boundary constraints.

6. The vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information according to claim 5, characterized in that: The specific process of screening candidate trajectories that meet the matching error threshold and outputting the lateral position calibration parameters and error boundaries is as follows: Segment the track generated according to inertial navigation by the sliding window mechanism to construct a candidate trajectory cluster; Calculate the weight matching error for each candidate trajectory, sort based on probability scoring, and screen out the non-dominated solution trajectory cluster through the Pareto front; Apply the Gaussian mixture model to the screened preferred trajectory cluster for clustering analysis, and extract the lateral position mean and covariance matrix of the clustering center trajectory as the calibration parameters and confidence interval.

7. An in-vehicle Beidou positioning deviation self-calibration system integrating inertial navigation information, characterized in that: When the on-vehicle Beidou signal is available, fuse multi-source data based on the tightly coupled filtering algorithm, dynamically adjust the noise covariance weight of the Beidou observation value through the weight factor, and the specific process of outputting the robust fusion positioning result is as follows: Construct a joint state vector including the biases of inertial devices, scale factor errors, and Beidou receiver clock errors, explicitly model the temperature drift term of inertial devices in the state vector, and establish a covariance correlation matrix with the Beidou clock error term; Map the weight factor to the independent noise covariance components of the Beidou pseudorange and carrier phase observations, downweight the low-credibility Beidou observations, and achieve heterogeneous downweighting of multiple observation sources; Construct a cost function through a robust M-estimator, perform reweighted iteration on the filtering residuals, and suppress the influence of multipath abnormal observations.

8. The vehicle-mounted Beidou positioning deviation self-calibration system integrating inertial navigation information according to claim 7, characterized in that: When the on-vehicle Beidou signal is continuously lost for more than the preset threshold, switch to the calibration mode driven by lane geometric constraints, and the specific process of suppressing the cumulative error of inertial navigation based on candidate trajectories and error boundaries is as follows: In the initial stage of Beidou signal failure, generate a lane-bound corridor based on the historical positioning result and the high-precision map to constrain the lateral divergence of inertial dead reckoning; Through the improved residual resampling particle filter, inject lane curvature constraints into the proposal distribution to force the particles to be distributed along the reasonable lane centerline; Inject the calibrated bias estimate value back into the tightly coupled filter state vector to achieve a smooth mode switch when the signal is restored and avoid positioning jumps.

Images