Algorithm-based vehicle driving route positioning point offset correction method and system

Abstract

The invention relates to the technical field of navigation positioning, in particular to a vehicle driving route positioning point offset correction method and system based on an algorithm. The method comprises the following steps: acquiring an initial positioning point of a vehicle on a preset route, and performing offset correction to obtain a first corrected positioning point; then, first offset information of the vehicle leaving the first correction positioning point is obtained, and the first offset information comprises a first offset distance; when the vehicle runs to a second correction positioning point, a second offset distance is obtained; based on the two offset distances, an offset variable is calculated for correcting the positioning point of the vehicle on the next preset route to obtain a third corrected positioning point, predictive correction is introduced, and the offset variable is calculated by analyzing historical offset data, so that the technical span from passive correction to active prediction is realized, the correction of the positioning point is more prospective, and the accuracy of positioning point correction is improved. The navigation precision is obviously improved; and an efficient and reliable solution is provided for accurate positioning and offset correction of a vehicle driving route.

Description

Technical Field

[0001] The present invention relates to the field of navigation and positioning technology, and in particular to an algorithm-based method and system for correcting the offset of positioning points on a vehicle's route, which is used to improve the accuracy of positioning points during vehicle travel, thereby improving navigation accuracy and user experience. Background Art

[0002] With the rapid development of global positioning systems (GPS) and mobile internet technologies, navigation systems have become an indispensable part of modern vehicles. However, due to factors such as satellite signal obstruction, multipath effects, and ionospheric interference, GPS positioning data often exhibits a certain degree of deviation, causing the vehicle's displayed position on the map to deviate from its actual location. This deviation not only affects the user experience but also poses a potential safety hazard in high-precision scenarios such as autonomous driving.

[0003] In the prior art, common methods for correcting positioning point offsets include map matching, Kalman filtering, and correction methods based on machine learning. The map matching method matches the vehicle's GPS trajectory with the road network on a digital map, and adsorbs the offset GPS point to the nearest road. However, this method is prone to mismatching in complex road network environments and is highly dependent on map data. The Kalman filtering method establishes a vehicle kinematic model and combines historical data for state estimation, but its prediction accuracy is greatly affected by the accuracy of the model and performs poorly when the GPS signal changes suddenly. Machine learning-based methods require a large amount of labeled data for training, and their real-time performance and generalization capabilities need to be improved.

[0004] Most of the above methods adopt a reactive strategy, correcting for deviations after they are detected. This lacks the ability to predict deviation trends. Furthermore, existing methods often treat the entire driving path as a homogeneous entity, ignoring the unique characteristics of different road sections, resulting in poor correction results. Furthermore, existing methods often struggle to maintain stable correction results in complex environments, such as urban areas with dense skyscrapers or when GPS signals are lost.

[0005] Therefore, there is an urgent need for a method and system that can predictably correct positioning point offsets and adapt to complex environmental changes. Summary of the Invention

[0006] The purpose of the present invention is to provide an algorithm-based method and system for correcting the offset of vehicle route positioning points. By analyzing historical positioning point offset data, introducing the concept of offset variables, and establishing a prediction model, the future positioning point offset trend can be predicted and corrected, thereby improving the positioning accuracy of the vehicle route.

[0007] The present invention proposes an algorithm-based method for correcting the deviation of a vehicle's route positioning point, comprising:

[0008] Obtain the initial positioning point of the vehicle on the preset route;

[0009] Performing offset correction on the initial positioning point to obtain a first corrected positioning point;

[0010] Acquire first offset information of the vehicle away from the first correction positioning point, where the first offset information includes a first offset distance;

[0011] When it is detected that the vehicle travels to the second correction positioning point, a second offset distance is obtained;

[0012] Calculating an offset variable based on the second offset distance and the first offset distance by using Lie group transformation and geodesic curvature on a Riemannian manifold;

[0013] Correcting the positioning point of the vehicle on the next preset route according to the offset variable to obtain a third corrected positioning point; and obtaining first offset information of the vehicle from the first corrected positioning point further includes:

[0014] Based on the driving direction of the first corrected positioning location, the second corrected positioning location is obtained according to the offset variable and the first offset distance. The coordinate expression of the second corrected positioning location is:

[0015] ,

[0016] , ,

[0017] in, The horizontal coordinate of the second corrected positioning location; The vertical coordinate of the second corrected positioning location; is the horizontal coordinate of the first corrected positioning location; The vertical coordinate of the first corrected positioning location; is the horizontal coordinate of the initial positioning location; The vertical coordinate of the initial positioning location, used to calculate navigation parameters; is the offset variable computed by a manifold-preserving homeomorphism in topological space, is the position correction parameter, which is calculated from the offset variable and the distance difference; % is the remainder function in the number, which represents the periodic boundary condition on the compact Lie group; Indicates the first time, in seconds; To correct the offset direction of the positioning location and the speed of the vehicle according to the first The calculated vehicle offset distance of the first corrected positioning location; is the speed of the vehicle in kilometers per hour; is the offset distance of the initial positioning point; is the first offset distance;

[0018] Preferably, the method further includes an adaptive offset threshold adjustment step:

[0019] The offset determination threshold is dynamically adjusted according to an adaptive threshold calculation model, wherein the adaptive threshold calculation model is:

[0020] ,

[0021] in, for Adaptive threshold at the moment; is the basic threshold; To adjust the sensitivity coefficient, the range is is the Lyapunov index, which characterizes the trajectory sensitivity of the system near the chaotic attractor; is the number of historical samples; The first The weight of historical samples; For the The offset distance of each historical positioning point; is the weighted average of historical offset distances; is the standard deviation of the historical offset distance; For environmental factors.

[0022] Preferably, when the vehicle is detected to have traveled to the second correction positioning point, obtaining the second offset distance includes:

[0023] Obtain positioning points on the vehicle's route;

[0024] Calculate a first distance between the positioning point and a first corrected positioning point based on the positioning point and the initial positioning location;

[0025] calculating a vehicle speed of the vehicle leaving the first correction positioning point based on the first distance;

[0026] Calculating a first time when the vehicle leaves the first correction positioning point according to the vehicle speed;

[0027] Calculate the distance traveled by the vehicle to the positioning point based on the initial positioning location and the positioning point position;

[0028] Calculating a vehicle speed of the vehicle traveling to the positioning point based on the travel distance and the first time;

[0029] A second offset distance is calculated based on the first offset distance of the first corrected positioning location and the first time.

[0030] Preferably, the geographical coordinate expression of the positioning location is:

[0031] longitude:

[0032] latitude:

[0033] in, 、 is the latitude and longitude of the vehicle's initial position; 、 The latitude and longitude values ​​of the positioning point; For the first time, The value obtained by converting seconds, minutes and seconds into hours; is the speed of the vehicle in kilometers per hour; is the time zone offset value; is the longitude ratio corresponding to the time; Whether it is a DST offset, 1 for yes, 0 for no; Is it the east longitude? 1 means yes, 0 means no. Is it east latitude, 1 for yes, 0 for no; Is it a small local sutra, 1 for yes, 0 for no; is the error correction parameter.

[0034] Preferably, the environmental factors The calculation formula is:

[0035] ,

[0036] in, is the satellite signal strength factor, ranging from is the terrain complexity factor, ranging from ; is the building occlusion factor, ranging from [1.0, 2.0];

[0037] The historical sample weight The calculation uses the time decay function:

[0038] ,

[0039] in, is the time attenuation coefficient, the default value is 0.1; is the current timestamp; For the The timestamp of each historical sample.

[0040] As an advantage, the method further includes the following steps for handling GPS signal loss:

[0041] When GPS signal loss is detected, the vehicle position is estimated using dead reckoning based on speed and direction:

[0042] ,

[0043] in, for Estimated location at the moment; The last valid GPS positioning point; For the moment speed; For the moment Direction of travel; The default sampling time interval is 0.1 seconds.

[0044] As an advantage, the method further includes the following steps:

[0045] Calculating the drift rate ;

[0046] Determine whether the offset rate exceeds the threshold ;

[0047] when When , it is marked as an abnormal point;

[0048] When three abnormal points are detected continuously, Kalman filtering is used for smoothing, and its state vector , state transition matrix is a 4×4 matrix, where , and the rest of the elements are 0.

[0049] A vehicle route positioning point deviation correction system based on an algorithm, comprising:

[0050] The initial positioning point acquisition module is used to obtain the initial positioning location of the vehicle when it is traveling on a preset route;

[0051] A first correction positioning point acquisition module is used to perform offset correction on the initial positioning location to obtain a first correction positioning location;

[0052] A first offset information acquisition module is configured to acquire first offset information of the vehicle moving away from the first corrected positioning location, wherein the first offset information includes a first offset distance;

[0053] A second corrected positioning location acquisition module is configured to obtain a second offset distance when detecting that the vehicle has traveled to the second corrected positioning location;

[0054] an offset variable calculation module, configured to calculate an offset variable of the vehicle traveling between the first corrected positioning location and the second corrected positioning location based on the second offset distance and the first offset distance;

[0055] The third corrected positioning location acquisition module is used to correct the positioning location of the vehicle when it travels on the next preset route according to the offset variable to obtain a third corrected positioning location.

[0056] The beneficial effects of the present invention include but are not limited to:

[0057] 1. Predictive correction: This invention achieves a technological leap from passive correction to active prediction by analyzing historical offset data and calculating offset variables, making positioning point correction more forward-looking and improving navigation accuracy.

[0058] 2. Strong adaptability: The present invention introduces an adaptive offset threshold adjustment mechanism, which dynamically adjusts parameters according to vehicle speed, road type, environmental factors, etc., to adapt to the needs of different scenarios and improve system robustness.

[0059] 3. Segment-by-segment fine processing: Differentiated processing strategies are adopted according to the characteristics of different road sections, which solves the limitation of traditional methods that treat the entire path as a homogeneous body and improves correction accuracy.

[0060] 4. Abnormal handling capability: A complete GPS signal loss handling mechanism and abnormal offset detection and processing process are designed to ensure the stable operation of the system in various complex environments.

[0061] 5. Strong practicality: The algorithm adopted by the present invention has the characteristics of high computational efficiency and low resource consumption. It is suitable for implementation on vehicle-mounted terminals and mobile devices and has good engineering practical value.

[0062] After testing, the present invention has achieved remarkable results in various road environments, with the average positioning accuracy increased by more than 70%. In particular, in high-offset environments such as complex urban intersections, the improvement is particularly obvious, greatly enhancing the navigation experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a flow chart of the algorithm-based vehicle route positioning point offset correction method of the present invention;

[0064] Figure 2 It is a workflow diagram of the offset variable calculation module of the present invention;

[0065] Figure 3 is a schematic diagram of the adaptive offset threshold adjustment mechanism of the present invention;

[0066] Figure 4 This is a flow chart of the GPS signal loss processing of the present invention;

[0067] Figure 5 is a schematic diagram of abnormal deviation detection and processing of the present invention;

[0068] Figure 6It is a system architecture diagram of the present invention;

[0069] Figure 7 This is a comparison chart of the experimental results of the present invention in a highway scenario;

[0070] Figure 8 This is a comparison chart of the experimental results of the present invention in a complex urban intersection scenario. DETAILED DESCRIPTION

[0071] Please refer to the attached Figure 1-8 , embodiments of the present invention will be described in detail below with reference to the accompanying drawings. The accompanying drawings illustrate embodiments of various aspects of the present invention. However, the present invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0072] The proposed method and system for correcting vehicle route positioning point offsets address the technical problem of GPS positioning point offsets in traditional navigation, which impacts navigation accuracy. Specifically, by introducing the concept of offset variables, the present invention achieves a technological leap from passive correction to active prediction, maintaining effective correction results in a variety of complex environments.

[0073] Before further explaining this invention, it's necessary to define the relevant concepts. In this invention, a positioning point refers to the geographic coordinates of a vehicle at a specific moment, as determined by GPS or other positioning methods. An offset is the deviation between a specified location and the vehicle's actual position. The offset variable, the core innovation of this invention, represents the trend of the offset change per unit time. And a corrected positioning point refers to the more accurate location coordinates obtained after processing using this invention's offset correction algorithm.

[0074] Reference Figure 1 As shown in the flowchart of the vehicle driving route positioning point offset correction method of the present invention, the overall method includes the key steps of initial positioning point acquisition, offset correction, offset information acquisition, offset variable calculation and application. At the system level, such as Figure 6 As shown, the system architecture of the present invention covers four main levels: data acquisition, processing, algorithm, and output distribution, forming a complete technical solution, specifically including:

[0075] Obtain the initial positioning point of the vehicle on the preset route;

[0076] Performing offset correction on the initial positioning point to obtain a first corrected positioning point;

[0077] Acquire first offset information of the vehicle away from the first correction positioning point, where the first offset information includes a first offset distance;

[0078] When it is detected that the vehicle travels to the second correction positioning point, a second offset distance is obtained;

[0079] Calculating an offset variable based on the second offset distance and the first offset distance by using Lie group transformation and geodesic curvature on a Riemannian manifold;

[0080] Correcting the positioning point of the vehicle on the next preset route according to the offset variable to obtain a third corrected positioning point; and obtaining first offset information of the vehicle from the first corrected positioning point further includes:

[0081] Based on the driving direction of the first corrected positioning location, the second corrected positioning location is obtained according to the offset variable and the first offset distance. The coordinate expression of the second corrected positioning location is:

[0082] ,

[0083] , ,

[0084] in, The horizontal coordinate of the second corrected positioning location; The vertical coordinate of the second corrected positioning location; is the horizontal coordinate of the first corrected positioning location; The vertical coordinate of the first corrected positioning location; is the horizontal coordinate of the initial positioning location; The vertical coordinate of the initial positioning location, used to calculate navigation parameters; is the offset variable computed by a manifold-preserving homeomorphism in topological space, is the position correction parameter, which is calculated from the offset variable and the distance difference; % is the remainder function in the number, which represents the periodic boundary condition on the compact Lie group; Indicates the first time, in seconds; To correct the offset direction of the positioning location and the speed of the vehicle according to the first The calculated vehicle offset distance of the first corrected positioning location; is the speed of the vehicle in kilometers per hour; is the offset distance of the initial positioning point; is the first offset distance;

[0085] Also includes an adaptive offset threshold adjustment step:

[0086] The offset determination threshold is dynamically adjusted according to an adaptive threshold calculation model, wherein the adaptive threshold calculation model is:

[0087] ,

[0088] in, for Adaptive threshold at the moment; is the basic threshold; To adjust the sensitivity coefficient, the range is is the Lyapunov index, which characterizes the trajectory sensitivity of the system near the chaotic attractor; is the number of historical samples; The first The weight of historical samples; For the The offset distance of each historical positioning point; is the weighted average of historical offset distances; is the standard deviation of the historical offset distance; For environmental factors.

[0089] When the vehicle continues to travel and leaves the first correction positioning point, the system begins to obtain the first offset information, wherein the key parameter is the first offset distance, that is, the vertical distance between the actual driving trajectory of the vehicle and the preset route.

[0090] This invention innovatively incorporates Lie group transformations and Riemannian manifold theory to calculate the offset variable. The vehicle's motion path is considered a curve embedded on a Riemannian manifold, and the optimal offset variable is estimated by calculating the geodesic curvature. This mathematical model enables the system to more accurately capture the inherent geometric characteristics of vehicle motion, improving the accuracy of offset prediction.

[0091] The remainder operation here This is of special significance because it is not just a simple mathematical operation, but rather represents a periodic boundary condition on a compact Lie group, ensuring that the calculation results are within a reasonable range and avoiding divergence caused by cumulative errors. In practical applications, this operation effectively handles looping roads and recurring navigation environments.

[0092] Example: Assume that a vehicle is driving in an urban environment. The initial positioning point coordinates are (116.403884, 39.914889). After map matching, the first corrected positioning point is (116.403902, 39.914875). The vehicle is driving at a speed of 30 km / h. After 5 seconds, the offset distance is 3.2 meters. Using the above formula, the offset variable The speed is about 0.64m / s. Combining Lie group transformation and geodesic curvature calculation, the coordinates of the second corrected positioning point are (116.403947, 39.914821). This point is closer to the actual position of the vehicle than the result calculated by traditional methods, and the deviation is reduced by about 68%.

[0093] In a preferred embodiment of the present invention, an initial positioning point is first acquired for the vehicle along a preset route. This process is typically accomplished using an onboard GPS receiver or other positioning device. In addition to basic latitude and longitude information, this initial positioning point also includes auxiliary data such as a timestamp and positioning accuracy. It is worth noting that, in practical applications, it is also crucial to simultaneously acquire information about the vehicle's speed and direction of travel; this information plays a key role in subsequent correction calculations.

[0094] After obtaining the initial positioning point, the offset correction link is immediately entered. The present invention uses an improved map matching technology to intelligently match the initial positioning point with the road network of the electronic map to obtain the first corrected positioning point. Unlike traditional methods, the matching process of the present invention not only takes into account the distance factor, but also makes full use of information such as road direction and vehicle historical trajectory, significantly reducing the probability of mismatching. Preferably, in open environments such as highways, the matching radius can be set to 15 meters; in complex environments such as urban blocks, the radius can be dynamically adjusted to 8-12 meters to meet the needs of different scenarios. Actual tests show that this preliminary correction process can reduce the original offset by about 50%, laying the foundation for subsequent fine correction.

[0095] As the vehicle continues to travel and leaves the first corrected positioning point, the system begins acquiring first offset information. The most critical parameter is the first offset distance. This distance is defined as the vertical distance between the vehicle's actual trajectory and the pre-set route. It is calculated by comparing real-time GPS data with the centerline of the road on the map. In one embodiment of the present invention, this distance can be calculated using the closest point projection method. This method projects the vehicle's current position onto the nearest road segment and calculates the distance between the projected point and the original point.

[0096] The present invention further introduces an intelligent offset determination mechanism. Specifically, the system compares the first offset distance with a preset threshold to determine the subsequent processing strategy. Based on extensive experimental data analysis, this threshold is typically set at 5 meters, a value consistent with the average error range of common commercial GPS receivers. When the offset distance is greater than or equal to this threshold, it indicates a significant offset, and the system directly enters the offset variable calculation process. Conversely, if the offset distance is within the threshold, it is considered a normal error within an acceptable range, and the system initiates an alternative processing process to avoid unnecessary corrective operations, thereby saving computing resources and maintaining navigation stability.

[0097] In the alternative processing flow, the system first obtains a preset offset distance range, typically 0-10 meters, covering the common GPS offset range. It then obtains the vehicle's second offset information, calculates the second offset distance, and compares it with the first offset distance to obtain the offset difference. If the difference exceeds a second preset value (empirically 2 meters), the offset difference is retained for subsequent calculations. Otherwise, the system deems the offset change insignificant and proceeds directly to the normal processing flow. This multi-level judgment mechanism significantly improves the system's adaptability and resource efficiency.

[0098] When the vehicle reaches the second corrected position, the system needs to calculate the second offset distance. This calculation process is relatively complex and involves multiple detailed steps. First, the real-time position point on the vehicle's route is obtained. Then, the first distance between the two points is calculated based on this point and the initial position point. The spherical distance calculation formula is used here, which takes into account the influence of the earth's curvature and is more accurate than the straight-line distance calculation:

[0099] ,

[0100] in, is the average radius of the Earth, about 6371 kilometers; and Respectively represent the latitude of two points (radians); and This method of calculation is more accurate than plane distance calculation at larger distances, especially when moving long distances in the north-south direction.

[0101] After obtaining the distance, the system further calculates the speed at which the vehicle moves away from the first correction positioning point. Here, a simple physical formula is used: speed equals distance divided by time, that is, .in is the vehicle speed (km / h), is the first distance calculated (km), is the time difference (hours). It is worth noting that in practical applications, in addition to obtaining the vehicle speed through calculation, the speed information in the vehicle CAN bus can also be directly read, which often produces more accurate results.

[0102] Next, the system uses the vehicle speed information to calculate the time it takes the vehicle to leave the first corrected positioning point (i.e., the first time). In most practical applications, this time can be directly obtained from the system timestamp. However, in cases where the data is incomplete or the timestamp is unreliable, it can be inferred from the distance and speed.

[0103] The system also needs to calculate the vehicle's actual distance traveled from its initial point to its current location. Unlike simple straight-line distance calculations, this method takes the actual road shape into account, employing a segment-by-segment integration approach to achieve a more accurate distance. Specifically, the road can be divided into multiple small segments, and the total distance is calculated by summing the lengths of each segment. While this method requires a slightly higher computational effort, it significantly improves accuracy, especially on roads with numerous curves.

[0104] Combining the distance traveled and the initial time, the system calculates the vehicle's actual speed at the current location. Finally, based on the offset distance and time interval of the first corrected location, the system calculates the second offset distance. This continuous chain of calculations ensures the accuracy and consistency of offset distance calculations.

[0105] After obtaining the two key parameters of the first and second offset distances, the system enters the core link of the present invention: the calculation of the offset variable. The offset variable represents the trend of the offset change per unit time and is the basis for predictive correction. Its mathematical expression is:

[0106] ,

[0107] In this formula, Indicates the offset variable in meters per second; and The offset distances of the first and second correction positioning points are in meters; is the time interval between two correction positioning points, in seconds; The road condition adjustment coefficient varies in different road environments, for example, it can be 1.2 for urban roads, 0.8 for highways, and 1.0 for rural roads; The offset direction influence factor usually ranges from 0.1 to 0.9. Based on experience, 0.5 is sufficient for most scenarios. It represents the angle between the vehicle's direction of travel and the road direction, in radians.

[0108] The ingenuity of this formula lies in its simultaneous consideration of both the rate of change and the direction of the offset. The former reflects the temporal trend of the offset, while the latter accounts for the impact of vehicle steering on the offset. The contribution of the directional term becomes particularly important when driving on curves, significantly improving the model's prediction accuracy.

[0109] Based on the calculated offset variable, the system can predictively correct the positioning point when the vehicle is traveling on the next preset route to obtain a third corrected positioning point. Specifically, based on the driving direction of the first corrected positioning point, combined with the offset variable and the first offset distance, the position of the next positioning point is predicted and corrected. The coordinate calculation expression is:

[0110] ,

[0111] ,

[0112] ,

[0113] in, and are the horizontal and vertical coordinates of the second correction positioning point respectively; and Indicates the horizontal and vertical coordinates of the first correction positioning point; and are the horizontal and vertical coordinates of the initial positioning point; is the offset variable just calculated; the “%” symbol represents the remainder operation in mathematics; is the first time interval in seconds; is the offset distance calculated based on the offset direction and vehicle speed of the first corrected positioning point; is the vehicle speed in kilometers per hour.

[0114] This complete set of coordinate transformation formulas enables the system to accurately predict and correct the positions of subsequent positioning points, achieving continuous and smooth trajectory correction. Notably, the same calculation principle can be recursively applied to subsequent positioning points, forming a continuous correction chain to ensure consistent positioning accuracy throughout the entire driving process.

[0115] In order to express the location of the positioning point more accurately, the present invention also provides a detailed geographic coordinate expression, taking into account the influence of various geographical factors. The longitude calculation formula is:

[0116] ,

[0117] The latitude calculation formula is:

[0118] ,

[0119] In this set of formulas, and The latitude and longitude values ​​representing the vehicle's initial position; and is the latitude and longitude value of the positioning point; The value obtained by converting seconds, minutes and seconds into hours is calculated as follows: ,in is the first time (seconds); is the speed of the vehicle (km / h); The time zone offset value. For example, for the East 8 zone where China is located, the value is 8. The longitude ratio corresponding to time, usually 15 degrees / hour; Indicates whether it is a daylight saving time offset, 1 if yes, 0 if no; Indicates whether it is east longitude, 1 if yes, 0 if no; Indicates whether it is east latitude, 1 if yes, 0 if no; Indicates whether it is a small local sutra, 1 if yes, 0 if no; This is an error correction parameter, usually ranging from 0.8 to 1.2, and is dynamically adjusted according to the local GPS signal quality.

[0120] This formula takes into account the influence of various factors on positioning, such as the Earth's rotation, time zone differences, and seasonal changes, making coordinate calculations more accurate. Adjusting these parameters can effectively improve positioning accuracy, especially when traveling across time zones or in areas with unique terrain.

[0121] Traditional positioning correction methods often use fixed thresholds to determine the degree of offset, making them difficult to adapt to complex and changing environmental conditions. To address this issue, this paper introduces an innovative adaptive offset threshold adjustment mechanism, enabling the system to dynamically adjust the judgment criteria based on environmental conditions and historical performance, significantly improving the system's adaptability and robustness.

[0122] The core of adaptive threshold adjustment is a computational model based on historical data and environmental factors. Figure 3 As shown in Figure 2, the model takes into account multiple factors and forms a complete adaptive adjustment closed loop. Its mathematical expression is:

[0123] ,

[0124] In this formula, represents the adaptive threshold at time t; Is the basic threshold, usually set to 5 meters. This value is based on a large amount of experimental data statistics and represents the average error of ordinary GPS receivers under standard conditions; To adjust the sensitivity coefficient, the range is between 0.5 and 1.5, and it can be adjusted dynamically according to needs. For example, a larger value such as 1.2 can be taken in an urban environment, while a smaller value such as 0.8 can be taken in an open area; The default number of historical samples is the data of the last 10 positioning points. This value balances the computational burden and sample representativeness. is the weight of the i-th historical sample. The more recent the sample, the greater the weight, which reflects the time correlation; is the offset distance of the i-th historical positioning point; Represents the weighted average of historical offset distances; is the standard deviation of the historical offset distance, reflecting the degree of offset fluctuation; It is the environmental factor, which takes into account the impact of the current environment on the GPS signal.

[0125] Environmental factors It is further refined into the product of three sub-factors:

[0126] ,

[0127] in, is the satellite signal strength factor, ranging from 0.8 to 1.2, with smaller values ​​for stronger signals; is the terrain complexity factor, ranging from 0.9 to 1.5, and the more complex the terrain, the larger the value; The building occlusion factor ranges from 1.0 to 2.0, with higher values ​​for more severe occlusion. These three sub-factors can be obtained from vehicle sensor data or map database information, or can be derived based on historical performance statistics.

[0128] It is worth mentioning the calculation method of historical sample weights. This invention uses a time decay function to make recent samples have a higher reference value:

[0129] ,

[0130] in, is the time attenuation coefficient, the default value is 0.1, and it can be adjusted according to the actual application scenario; Indicates the current timestamp (seconds); This exponential decay model ensures that recent data has a higher reference value while not completely ignoring historical information, thus maintaining the continuity and stability of the prediction.

[0131] In practice, the adaptive mechanism enables the system to maintain optimal performance in diverse environments. For example, when a vehicle is traveling on a city street surrounded by tall buildings, the system automatically identifies building occlusions and appropriately raises the offset threshold to avoid overcorrection and resulting in trajectory jitter. On open highways, the system lowers the threshold to achieve higher accuracy. This intelligent adjustment significantly improves the system's versatility and user experience.

[0132] In the actual navigation process, GPS signal loss is a common problem, especially in tunnels, underground garages or high-rise areas. Traditional navigation systems often cannot provide reliable location information when the signal is lost, which seriously affects the user experience. To address this problem, the present invention designs a special GPS signal loss processing mechanism, such as Figure 4 As shown, it ensures that the system can provide continuous and reliable location services even when the signal is unavailable.

[0133] When the system detects that the GPS signal is lost, it automatically starts to estimate the vehicle's position based on speed and direction. The mathematical expression is:

[0134] ,

[0135] In this formula, express Estimated location coordinates at the moment; Is the last valid GPS positioning point coordinate; For the moment The vehicle speed in meters per second; Indicates time The vehicle's direction of travel, in radians; The default value is 0.1 seconds, which is a good balance between computational accuracy and computational burden.

[0136] This dead reckoning method, based on the principles of inertial navigation, estimates position changes by integrating the vehicle's velocity vector. In practice, velocity information can be obtained from the vehicle's CAN bus, while direction information can be derived from the gyroscope or the last known direction. Notably, this dead reckoning method provides relatively accurate position estimates within a short period of time (typically no more than 30 seconds), with relatively slow error growth. For example, in a typical urban road scenario, the estimated position error within 15 seconds typically does not exceed 10 meters, meeting basic navigation requirements.

[0137] To further improve estimation accuracy, the present invention also incorporates road topology information from electronic maps to constrain the estimated trajectory. Specifically, the system attaches the estimated location points to the most likely road, preventing the trajectory from straying from the actual road. This map-matching technology is particularly suitable for areas with a clear road network, such as urban areas or highways.

[0138] For extended periods of signal loss (over 30 seconds), the system reduces the confidence level in the position estimate and increases the error margin for navigation prompts. For example, when approaching an intersection, the system will issue a prompt indicating that the intersection is approaching, rather than accurately predicting that the intersection will arrive 50 meters later, to avoid incorrect guidance caused by position estimation errors.

[0139] Through this complete signal loss handling mechanism, the present invention can provide continuous and reliable navigation services even when GPS is unavailable, significantly improving the user experience, especially in special scenarios such as tunnels and underground garages.

[0140] In addition to the complete loss of GPS signals, positioning data may also be interfered with by various factors, such as multipath effects, ionospheric anomalies, etc., resulting in mutations or abnormal offsets. If these abnormal data are not processed, they will cause navigation instructions to jump, seriously affecting the user experience. To address this problem, the present invention has designed a complete abnormal offset detection and processing process, such as Figure 5 As shown, the robustness of the system is further improved.

[0141] The first step in anomaly detection is to calculate the offset rate, which is the change in offset distance per unit time:

[0142] ,

[0143] in, Indicates the offset rate in meters per second; The offset distance of the current positioning point, in meters; is the offset distance of the previous fixed point, in meters; It is the time interval between two positioning points in seconds.

[0144] The system then compares the calculated offset rate with the dynamic threshold to determine whether there is an anomaly. It is not a fixed value, but is dynamically adjusted according to vehicle speed:

[0145] ,

[0146] in, is the dynamic offset rate threshold, in meters per second; The default value is 2 meters per second, which is determined based on the common mutation characteristics of GPS. is the current vehicle speed in kilometers per hour. This dynamic threshold design takes into account the rapid changes in position when driving at high speeds and avoids misjudgment.

[0147] When the system detects a deviation rate exceeding a threshold, it marks that point as an outlier. To prevent misjudgments caused by occasional noise, the system employs a continuous outlier detection mechanism: only after three consecutive outliers are detected will the outlier handling process be initiated. This three-point detection rule balances sensitivity and reliability, enabling timely detection of true anomalies while effectively filtering out occasional noise.

[0148] Once an abnormal offset is confirmed, the system will start the Kalman filter for smoothing. Kalman filtering is a classic state estimation algorithm that is particularly suitable for processing continuous time series data containing noise. In the present invention, the state vector of the Kalman filter is defined as:

[0149] ,

[0150] This vector contains four elements: and Indicates the location coordinates (meters), and represents the velocity component (m / s). By estimating both position and velocity simultaneously, the Kalman filter can provide a smoother trajectory.

[0151] The state transition matrix is ​​defined as:

[0152] ,

[0153] in is the time step in seconds. This transition matrix describes how the system state evolves over time, based on the classical physical equation of motion: the new position equals the old position plus the velocity times the time.

[0154] Kalman filtering also requires two key parameters: the measurement noise covariance matrix and the process noise covariance matrix The settings of these two parameters directly affect the filtering effect. In the preferred embodiment of the present invention, based on a large amount of experimental data analysis, Set as a diagonal matrix with diagonal elements of [25,25,4,4], with units of square meters and square meters per second, respectively, reflecting the uncertainty of GPS measurements; Set it as a diagonal matrix with diagonal elements of [0.1, 0.1, 0.01, 0.01], with the same units as above, representing the uncertainty of the internal state changes of the system.

[0155] It is worth noting that the Kalman filter parameters are not fixed, but can be dynamically adjusted according to the specific application scenario and GPS receiver performance. For example, on a high-precision GPS receiver, the Kalman filter parameters can be appropriately reduced. The element values ​​of the matrix; in complex urban environments, it may be necessary to increase matrix elements to accommodate more frequent speed changes.

[0156] Through Kalman filtering, the system effectively smooths out abnormal deviations, maintaining the continuity and rationality of the navigation trajectory. Actual tests have shown that this method can eliminate most sudden changes while preserving the trajectory's true direction, significantly improving the navigation experience.

[0157] In terms of specific implementation, the vehicle route positioning point offset correction system of the present invention adopts a modular design, such as Figure 6 As shown, each functional module works together to form a complete technical solution.

[0158] The core of the system is a series of processing modules with clear functions. The initial positioning point acquisition module 1 is responsible for acquiring the vehicle's original GPS position data; the first correction positioning point acquisition module 2 performs preliminary corrections to the initial point; the first offset information acquisition module 3 calculates and records offset information; the second correction positioning point acquisition module 4 processes the offset data when the vehicle reaches the next location; the offset variable calculation module 5 implements the core algorithm of the present invention; and the third correction positioning point acquisition module 6 corrects subsequent points based on the offset prediction.

[0159] In addition to the basic processing flow, the system also includes several enhancement modules: an adaptive threshold adjustment module 7 dynamically optimizes the judgment criteria; a GPS signal loss processing module 8 provides position estimation when the signal is unavailable; an abnormal offset detection module 9 identifies and processes sudden changes in data; a data storage module 10 saves historical information for system learning; and a data output module 11 distributes the corrected positioning data through multiple channels.

[0160] From a data flow perspective, the system's operating process can be roughly divided into five phases. The first phase is initialization, during which the system boots up and loads configuration parameters, including default threshold settings and historical data caching. Next, the data acquisition phase occurs, where the GPS module collects real-time location data, along with auxiliary information such as vehicle speed and direction. The third phase involves offset detection and calculation, including data preprocessing, coordinate conversion, and offset calculation. This is followed by the correction calculation phase, where the system predicts and corrects subsequent positioning points based on historical offset trends. Finally, the correction results are distributed to the vehicle terminal and mobile devices via the data output module and stored in a database as historical records.

[0161] In practical applications, system efficiency is a key consideration, especially given the limited computing resources of onboard terminals. To this end, the present invention employs a variety of optimization strategies to improve computational efficiency. For example, quadtree indexing is used to accelerate spatial queries, increasing query speeds by 5-8 times in testing. An LRU cache mechanism is introduced to reduce recalculation, achieving a cache hit rate of over 85% on regular commuting routes. Dynamic parameter adjustment is also implemented to adaptively optimize computation accuracy and frequency based on device performance and battery status.

[0162] In addition to optimizing efficiency, the system also implements multiple precision-enhancing strategies. Multi-source fusion technology combines multiple positioning data sources, including GPS, base stations, and Wi-Fi, to improve positioning accuracy by 15-20% in urban environments. A historical trajectory learning function records and leverages historical deviation patterns of specific road sections, achieving increasing accuracy over time. Road feature recognition enables the system to apply specialized correction strategies for specific road structures (such as roundabouts and overpasses), further improving accuracy.

[0163] To fully verify the effectiveness of the present invention, we conducted system tests in a variety of typical scenarios. The following details the experimental results of two representative scenarios.

[0164] First, consider the highway scenario. GPS signals are generally relatively stable on highways, but vehicle speeds are relatively high, placing correspondingly higher demands on positioning accuracy and real-time performance. In this scenario, the base offset threshold was set to 8 meters, the speed impact factor was 0.6, the road type coefficient was 1.2, the road condition adjustment coefficient was 0.8, the offset direction impact factor was 0.4, and the time decay coefficient was 0.15. These parameter settings take into account the characteristics of highways: faster position updates at higher speeds, relatively stable road conditions, and predominantly straight-line driving.

[0165] The test used a test vehicle equipped with a medium-precision GPS receiver, driving about 100 kilometers along the Beijing-Hong Kong-Macao Expressway at an average speed of 100 kilometers per hour, collecting positioning data once per second. Figure 7 As shown in the test results, the average deviation was 12.6 meters before using this solution, but it was reduced to 3.2 meters after using this solution, with a deviation improvement rate of 74.6%. It is particularly worth mentioning that this solution performed even better in complex sections such as highway entrances and exits, with positioning accuracy increased by over 85%, greatly improving the accuracy and timeliness of navigation instructions.

[0166] Another typical test scenario is a complex urban intersection. This type of environment, often characterized by obstructions from tall buildings and multiple intersecting roads, presents one of the most challenging scenarios for GPS positioning. Parameter settings were adjusted for this scenario: the basic offset threshold was lowered to 5 meters, the speed impact factor was increased to 0.9, the road type coefficient was 0.9, the building obstruction factor was set to 1.8, the terrain complexity factor was 1.4, the road condition adjustment coefficient was increased to 1.2, and the offset direction influence factor was increased to 0.7. These parameter changes reflect the characteristics of urban environments: low speeds, complex road conditions, frequent turns, and significant building obstruction.

[0167] The test was conducted at five typical complex intersections near Zhongguancun in Haidian District, Beijing. The same test vehicle was used and the test was repeated five times at each intersection with an average speed of about 30 km / h. Figure 8 As shown in the test results, the average deviation was 18.3 meters before using this solution, but it was reduced to 5.7 meters with this solution, a deviation improvement of 68.9%. Particularly notable was the excellent performance of this solution even in the most complex intersection conditions, reducing the maximum deviation of 24.5 meters to 7.2 meters, an improvement of 70.6%. This stable performance under extreme conditions fully demonstrates the robustness and adaptability of this invention.

[0168] In addition to the two main scenarios mentioned above, we also designed tests specifically for GPS signal loss and abnormal drift. On a test route that included an 800-meter-long tunnel (travel time of approximately 45 seconds), our proposed signal loss handling mechanism performed exceptionally well: within the first 15 seconds of signal loss, the average position estimation error was only 6.3 meters; between 15 and 30 seconds, the average error was kept to 11.5 meters; and even after prolonged signal loss exceeding 30 seconds, the average error did not exceed 18.7 meters, maintaining basic navigation service.

[0169] In tests of handling abnormal deviations, we artificially introduced sudden deviation points to evaluate the system's detection and smoothing capabilities. The results showed that the system successfully detected 97.5% of these abnormal deviations, reduced trajectory deviation by 82.3% after smoothing, and achieved an average processing delay of only 1.4 seconds, fully meeting the requirements of real-time navigation. These test results fully demonstrate the superior performance of this invention under various complex conditions.

[0170] Compared with the existing technology, the present invention has many significant advantages. The first is the predictive correction capability. The existing technology is mostly reactive correction, that is, correction is made after the offset is discovered; the present invention predicts the future offset trend through offset variables, realizes pre-correction, and greatly improves the timeliness and accuracy of navigation instructions. Secondly, it has strong adaptability. The present invention can dynamically adjust the system parameters according to environmental conditions, while traditional methods often use fixed parameters and have limited adaptability. Thirdly, it is a segmented fine processing strategy. Unlike the traditional method that regards the entire path as a homogeneous body, the present invention adopts differentiated processing according to the characteristics of the road section, especially in special sections such as complex intersections and ramps. In addition, the present invention also has powerful exception handling capabilities and low resource consumption, which is more suitable for implementation in resource-constrained vehicle environments.

[0171] Overall, the algorithm-based method and system for correcting vehicle route positioning point offsets proposed in this paper achieves a technological leap from passive correction to active prediction by introducing innovative technologies such as an offset variable prediction model and an adaptive threshold adjustment mechanism. Tests in various road environments have demonstrated an average improvement in positioning accuracy of over 70%, while also demonstrating excellent robustness and adaptability, demonstrating significant technological advancement and practical value.

[0172] This technology is not only applicable to general navigation scenarios but can also be expanded to areas requiring higher positioning accuracy, such as autonomous driving and logistics. Future research directions include further integrating deep learning technologies to improve adaptability to complex scenarios, and combining them with high-precision maps and V2X communication technologies to establish a more comprehensive location-based service ecosystem.

[0173] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A vehicle route positioning point offset correction method based on an algorithm, characterized in that: include: Obtain the initial positioning point of the vehicle on the preset route; Performing offset correction on the initial positioning point to obtain a first corrected positioning point; Acquire first offset information of the vehicle away from the first correction positioning point, where the first offset information includes a first offset distance; When it is detected that the vehicle travels to the second correction positioning point, a second offset distance is obtained; Calculating an offset variable based on the second offset distance and the first offset distance by using Lie group transformation and geodesic curvature on a Riemannian manifold; Correcting the positioning point of the vehicle on the next preset route according to the offset variable to obtain a third corrected positioning point; The obtaining of first offset information of the vehicle leaving the first correction positioning point further includes: Based on the driving direction of the first corrected positioning location, the second corrected positioning location is obtained according to the offset variable and the first offset distance. The coordinate expression of the second corrected positioning location is: ， ， ， in, The horizontal coordinate of the second corrected positioning location; The vertical coordinate of the second corrected positioning location; is the horizontal coordinate of the first corrected positioning location; The vertical coordinate of the first corrected positioning location; is the horizontal coordinate of the initial positioning location; The vertical coordinate of the initial positioning location, used to calculate navigation parameters; is the offset variable computed by a manifold-preserving homeomorphism in topological space, is the position correction parameter, which is calculated from the offset variable and the distance difference; % is the remainder function in the number, which represents the periodic boundary condition on the compact Lie group; Indicates the first time, in seconds; To correct the offset direction of the positioning location and the speed of the vehicle according to the first The calculated vehicle offset distance of the first corrected positioning location; is the speed of the vehicle in kilometers per hour; is the offset distance of the initial positioning point; is the first offset distance.

2. The method according to claim 1, characterized in that Also includes an adaptive offset threshold adjustment step: The offset determination threshold is dynamically adjusted according to an adaptive threshold calculation model, wherein the adaptive threshold calculation model is: ， in, for Adaptive threshold at the moment; is the basic threshold; To adjust the sensitivity coefficient, the range is is the Lyapunov index, which characterizes the trajectory sensitivity of the system near the chaotic attractor; is the number of historical samples; The first The weight of historical samples; For the The offset distance of each historical positioning point; is the weighted average of historical offset distances; is the standard deviation of the historical offset distance; For environmental factors.

3. The method according to claim 1, characterized in that The obtaining of first offset information of the vehicle leaving the first correction positioning point further includes: Determining the first offset distance; When the first offset distance is greater than or equal to a first preset value, performing the step of obtaining a second offset distance when detecting that the vehicle has traveled to a second correction positioning point; When the first offset distance is less than the first preset value, a preset first offset distance range is obtained; second offset information of the vehicle leaving the first correction positioning point is obtained; a second offset distance is obtained; an offset difference between the second offset distance and the first offset distance is calculated; when the offset difference is greater than a second preset value, the offset difference is obtained; when the offset difference is less than or equal to the second preset value, the step of obtaining the second offset distance when it is detected that the vehicle has traveled to the second correction positioning point is executed.

4. The method according to claim 3, characterized in that The step of obtaining a second offset distance when detecting that the vehicle has traveled to the second correction positioning point includes: Obtain positioning points on the vehicle's route; Calculate a first distance between the positioning point and a first corrected positioning point based on the positioning point and the initial positioning location; calculating a vehicle speed of the vehicle leaving the first correction positioning point based on the first distance; Calculating a first time when the vehicle leaves the first correction positioning point according to the vehicle speed; Calculate the distance traveled by the vehicle to the positioning point based on the initial positioning location and the positioning point position; Calculating a vehicle speed of the vehicle traveling to the positioning point based on the travel distance and the first time; A second offset distance is calculated based on the first offset distance of the first corrected positioning location and the first time.

5. The method according to claim 1, wherein The geographic coordinate expression of the location is: longitude: latitude: in, 、 is the latitude and longitude of the vehicle's initial position; 、 The latitude and longitude values ​​of the positioning point; For the first time, The value obtained by converting seconds, minutes and seconds into hours; is the speed of the vehicle in kilometers per hour; is the time zone offset value; is the longitude ratio corresponding to the time; Whether it is a DST offset, 1 for yes, 0 for no; Is it the east longitude? 1 means yes, 0 means no. Is it east latitude, 1 for yes, 0 for no; Is it a small local sutra, 1 for yes, 0 for no; is the error correction parameter.

6. The method according to claim 1, characterized in that The environmental factors The calculation formula is: ， in, is the satellite signal strength factor, ranging from is the terrain complexity factor, ranging from ; is the building occlusion factor, ranging from [1.0, 2.0]; The historical sample weight The calculation uses the time decay function: , in, is the time attenuation coefficient, the default value is 0.1; is the current timestamp; For the The timestamp of each historical sample.

7. The method according to claim 1, characterized in that It also includes the steps for handling GPS signal loss: When GPS signal loss is detected, the vehicle position is estimated using dead reckoning based on speed and direction: , in, for Estimated location at the moment; The last valid GPS positioning point; For the moment speed; For the moment Direction of travel; The default sampling time interval is 0.1 seconds.

8. The method according to claim 1, characterized in that It also includes abnormal deviation detection and processing steps: Calculating the drift rate ; Determine whether the offset rate exceeds the threshold ; when When , it is marked as an abnormal point; When three abnormal points are detected continuously, Kalman filtering is used for smoothing, and its state vector , state transition matrix is a 4×4 matrix, where , and the rest of the elements are 0.

9. An algorithm-based vehicle route positioning point offset correction system implementing the method according to any one of claims 1 to 8, characterized in that: include: The initial positioning point acquisition module is used to obtain the initial positioning location of the vehicle when it is traveling on a preset route; A first correction positioning point acquisition module is used to perform offset correction on the initial positioning location to obtain a first correction positioning location; A first offset information acquisition module is configured to acquire first offset information of the vehicle moving away from the first corrected positioning location, wherein the first offset information includes a first offset distance; A second corrected positioning location acquisition module is configured to obtain a second offset distance when detecting that the vehicle has traveled to the second corrected positioning location; an offset variable calculation module, configured to calculate an offset variable of the vehicle traveling between the first corrected positioning location and the second corrected positioning location based on the second offset distance and the first offset distance; The third corrected positioning location acquisition module is used to correct the positioning location of the vehicle when it travels on the next preset route according to the offset variable to obtain a third corrected positioning location.

Images