Vehicle positioning method, track prediction method, equipment and readable storage medium

Abstract

One or more embodiments of the present specification provide a vehicle positioning method, a trajectory prediction method, a device and a readable storage medium, the vehicle positioning method comprising: acquiring vehicle motion data of a vehicle at each first positioning point in a target road section, the vehicle motion data is determined based on the attitude and / or speed information of the vehicle at the first positioning point; vehicle positioning data of the vehicle at a second positioning point in a connection road section of the target road section is obtained, and the connection road section at least comprises a subsequent road section of the target road section; and calculating the vehicle positioning data of the vehicle at each first positioning point in the target road section at least according to the vehicle movement data of the first positioning points and the vehicle positioning data of the second positioning points.

Description

Technical Field

[0001] One or more embodiments of the present specification relate to the field of vehicle positioning technology, and in particular, to a vehicle positioning method, trajectory prediction method, device, and readable storage medium. Background Art

[0002] In the process of realizing operations such as spatial positioning or map construction and position calibration in the navigation system, it is necessary to analyze and process the vehicle positioning data of vehicles and other vehicles on the corresponding road sections. However, in actual applications, due to factors such as the environment or building obstruction and signal interference along the way, the vehicle may pass through sections where positioning data such as Global Positioning System (GPS) data cannot be received during driving, resulting in the vehicle being unable to obtain accurate positioning data on these sections.

[0003] In related technologies, the dead reckoning (DR) method is usually used, that is, the relative motion data detected by the vehicle is used to predict the driving trajectory. Although this method can make up for the data gap during the period when the vehicle positioning data is missing to a certain extent, the vehicle positioning data predicted based on the relative motion data has inherent data errors. This error will cause the predicted vehicle positioning data to gradually deviate from the actual driving path, and the error size will increase over time, eventually making the driving trajectory of the target section unreliable. Summary of the invention

[0004] In view of this, one or more embodiments of this specification provide the following technical solutions:

[0005] According to a first aspect of one or more embodiments of this specification, a vehicle positioning method is proposed, including:

[0006] Acquire vehicle motion data of the vehicle at each first positioning point in the target road section, wherein the vehicle motion data is determined based on the posture and / or speed information of the vehicle at the first positioning point; and,

[0007] Acquire vehicle positioning data of a second positioning point of the vehicle in a connecting section of the target section, wherein the connecting section at least includes a subsequent section of the target section;

[0008] The vehicle positioning data of each first positioning point of the vehicle in the target road section is calculated at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point.

[0009] Optionally, the second positioning point includes at least a third positioning point and / or a fourth positioning point, the third positioning point is the last positioning point that the vehicle passes through in the preceding road section, and the fourth positioning point is the first positioning point that the vehicle passes through in the succeeding road section; the preceding road section is a connecting road section of the target road section;

[0010] and / or

[0011] The target road section includes a road section where the signal quality of the positioning signal of the vehicle meets the abnormal quality requirement, and the positioning signal is used to parse the vehicle positioning data;

[0012] and / or

[0013] The vehicle motion data includes at least one of the displacement data, total acceleration, linear acceleration, angular velocity, rotation rate, gravity vector, attitude angle, and speed of the vehicle.

[0014] Optionally, the calculating the vehicle positioning data of each first positioning point of the vehicle in the target road section includes:

[0015] Acquire vehicle motion data of a second positioning point of the vehicle in the connecting section of the target section;

[0016] The vehicle positioning data of each first positioning point of the vehicle in the target road section is calculated according to the vehicle motion data of the first positioning point, the vehicle positioning data of the second positioning point, and the vehicle motion data of the second positioning point.

[0017] Optionally, when the second positioning points include at least a third positioning point and a fourth positioning point, calculating the vehicle positioning data of each first positioning point of the vehicle in the target road section at least based on the vehicle motion data of the first positioning points and the vehicle positioning data of the second positioning points includes:

[0018] Obtain vehicle motion data of the vehicle at the third positioning point; and, based on the vehicle motion data of the first positioning point, the vehicle positioning data of the third positioning point, the vehicle motion data of the third positioning point, and the vehicle positioning data of the fourth positioning point, calculate the vehicle positioning data of each first positioning point of the vehicle in the target road section; or obtain vehicle motion data of the vehicle at the fourth positioning point; and, based on the vehicle motion data of the first positioning point, the vehicle positioning data of the third positioning point, the vehicle positioning data of the fourth positioning point, and the vehicle motion data of the fourth positioning point, calculate the vehicle positioning data of each first positioning point of the vehicle in the target road section.

[0019] Optionally, calculating the vehicle positioning data of each first positioning point of the vehicle in the target road section at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point includes:

[0020] Generate initial vehicle positioning data of each first positioning point of the vehicle in the target road section based at least on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point;

[0021] An error fitting function for the initial vehicle positioning point is determined at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point, the initial vehicle positioning data is corrected based on the error fitting function, and the corrected initial vehicle positioning data is used as the vehicle positioning data of each first positioning point of the vehicle in the target road section.

[0022] Optionally, determining the error fitting function for the initial vehicle positioning point at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point includes:

[0023] At least the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point are substituted into a preset posture error divergence model to generate an error fitting function for the initial vehicle positioning point.

[0024] According to a second aspect of one or more embodiments of this specification, a vehicle trajectory prediction method is proposed, comprising:

[0025] Based on the method of the first aspect, obtaining vehicle motion data of each first positioning point of the vehicle in the target road section;

[0026] A driving trajectory of the vehicle on the target road section is generated according to the vehicle positioning data of each first positioning point of the vehicle on the target road section.

[0027] Optionally, generating a driving trajectory of the vehicle on the target road section according to the vehicle positioning data of each first positioning point of the vehicle on the target road section includes:

[0028] constructing an absolute position constraint factor according to the vehicle positioning data of the first positioning point, and constructing a relative motion constraint factor according to the vehicle motion data of the first positioning point;

[0029] The factor graph composed of the absolute position constraint factor and the relative motion constraint factor is optimized to obtain an optimal solution of the factor graph, and the driving trajectory of the vehicle in the target road section is generated based on each positioning point represented by the optimal solution.

[0030] Optionally, constructing the relative motion constraint factor according to the vehicle motion data of the first positioning point includes: performing pre-integration processing on the target relative motion data in the vehicle motion data of the first positioning point according to the sequence of the first positioning points corresponding to the vehicle motion data of the first positioning point, and using the pre-integration factor obtained by the processing as the relative motion constraint factor;

[0031] and / or

[0032] The factor graph is provided with a sliding window of a preset length; the factor graph composed of the absolute position constraint factor and the relative motion constraint factor is optimized, including: in the process of the sliding window sliding in sequence according to the preset step length, each time the sliding window slides, each node in the sliding window is optimized once.

[0033] According to a third aspect of one or more embodiments of this specification, an electronic device is proposed, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor implements the steps of the method described in the first aspect or the second aspect by running the executable instructions.

[0034] According to a third aspect of one or more embodiments of the present specification, a computer-readable storage medium is provided, on which computer instructions are stored. When the instructions are executed by a processor, the steps of the method described in the first aspect or the second aspect are implemented.

[0035] According to a fourth aspect of one or more embodiments of this specification, a computer program product is proposed, comprising a computer program / instruction, which implements the steps of the method described in the first aspect or the second aspect when executed by a processor.

[0036] In the technical solution provided in this specification, in addition to obtaining the vehicle motion data determined based on the posture and / or speed information at each first positioning point of the vehicle in the target section, the vehicle positioning data at the second positioning point of the vehicle in the connecting section is also obtained in a targeted manner, wherein the connecting section may at least include the subsequent section of the above-mentioned target section. Based at least on the vehicle motion data of the above-mentioned first positioning point and the vehicle positioning data of the above-mentioned second positioning point, the vehicle target positioning data of the above-mentioned vehicle at each first positioning point in the above-mentioned target section can be calculated. Among them, for the vehicle positioning data of the above-mentioned first positioning point, in addition to the vehicle motion data of each first positioning point in the target section, the vehicle positioning data of the second positioning point is also considered to be determined, thereby avoiding the data error caused by relying only on the relative motion data in the target section in the traditional dead reckoning method, so that the vehicle positioning data determined for the target section is more reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments or prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.

[0038] Figure 1 is a flow chart of a vehicle positioning method provided by an exemplary embodiment;

[0039] Figure 2 is a schematic diagram of an error fitting model provided by an exemplary embodiment;

[0040] Figure 3 is a flow chart of a vehicle trajectory prediction method provided by an exemplary embodiment;

[0041] Figure 4 is a schematic diagram of a factor graph provided by an exemplary embodiment;

[0042] Figure 5 is a flowchart of an optimization method based on a factor graph provided by an exemplary embodiment;

[0043] Figure 6 is a schematic diagram of a delayed output optimization result provided by an exemplary embodiment;

[0044] Figure 7 is a schematic diagram of a semi-real-time output trajectory and a real-time output trajectory provided by an exemplary embodiment;

[0045] Figure 8 is a schematic structural diagram of an electronic device provided by an exemplary embodiment;

[0046] Fig. 9 is a structural schematic diagram of a vehicle positioning device provided by an exemplary embodiment;

[0047] Fig.10 It is a schematic structural diagram of a vehicle trajectory prediction device provided by an exemplary embodiment. DETAILED DESCRIPTION

[0048] Here, exemplary embodiments will be described in detail, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are only examples of devices and methods consistent with some aspects of the present invention.

[0049] The terms used in the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The singular forms "a", "the" and "the" used in the present invention and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0050] It should be understood that although the terms first, second, third, etc. may be used in the present invention to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present invention, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0051] Figure 1 FIG. 1 is a flow chart of a vehicle positioning method provided by an exemplary embodiment. Figure 1 As shown, the above method includes:

[0052] Step S102: acquiring vehicle motion data of the vehicle at each first positioning point in the target road section, wherein the vehicle motion data is determined based on the posture and / or speed information of the vehicle at the first positioning point.

[0053] When it is necessary to locate the vehicle to determine the vehicle positioning data of each first positioning point in the target road section, the vehicle motion data of each first positioning point of the vehicle in the target road section can be first obtained, wherein the above-mentioned vehicle motion data can be used to characterize at least one of the posture, motion trend, speed information, etc. of the vehicle at the corresponding first positioning point, and the vehicle motion data itself can be determined based on at least one of the posture and motion information of the vehicle at the first positioning point. Specifically, the vehicle motion data may include the displacement data of the vehicle at the first positioning point, the total acceleration, the linear acceleration, the angular velocity, the rotation rate, the gravity vector, the attitude angle (such as including but not limited to any one or any combination of the pitch angle, the roll angle, and the yaw angle), and the speed or any combination thereof. Among them, a part of the vehicle motion data can be directly obtained by vehicle sensors such as the inertial measurement unit (IMU), odometer (Odometry, ODO), and visual odometry, such as the total acceleration and gravity component detected by the accelerometer in the inertial measurement unit, the angular velocity and rotation rate detected by the gyroscope in the inertial measurement unit, and the displacement data detected by the odometer. Another part of the vehicle motion data can be indirectly calculated based on the vehicle motion data directly obtained by the above-mentioned vehicle sensors, such as the attitude angle determined by fusing the data of the accelerometer and the gyroscope, the linear acceleration obtained by removing the gravity component from the above-mentioned total acceleration through a software algorithm, and the speed determined by methods such as integrating the linear acceleration once.

[0054] The so-called positioning point can be understood as the path information dynamically generated according to the preset trigger conditions when the vehicle is traveling along the road section, that is, each positioning point can correspond to a data set with a specific time point or spatial position record for applications such as navigation, positioning and path planning. The data set corresponding to each positioning point can usually include two parts of data: vehicle motion data and vehicle positioning data. Among them, the vehicle positioning data can represent the exact position of the vehicle in the geographic coordinate system, and can be provided by any one or more of the external positioning systems such as the Global Navigation Satellite System (GNSS), high-precision map matching, and laser radar SLAM. Among them, GNSS can further include satellite navigation systems operated by different countries or regions such as the Global Positioning System (GPS), BeiDou Satellite Navigation System (BeiDou, BDS), and Galileo Satellite Navigation System (Galileo). The specific content of the above vehicle positioning data includes but is not limited to any one or any combination of longitude coordinates, latitude coordinates, height / altitude, and positioning timestamp.

[0055] For the above-mentioned target road section, the influence of factors such as the environment along the way, building shielding or signal interference is obvious, such as tunnels, viaducts with complex spatial structures, etc., so it is difficult to obtain effective and reliable vehicle positioning data, resulting in the abandonment or neglect of unavailable vehicle positioning data at the first positioning point of the above-mentioned target road section, and only the corresponding vehicle motion data is included. In this regard, this specification also provides a method for determining the above-mentioned target road section.

[0056] In one embodiment, the target road section may include a road section where the signal quality of the positioning signal of the vehicle meets the abnormal quality requirement, and the positioning signal may be a signal used to parse the positioning data of the vehicle; the abnormal quality requirement may include at least one of the signal strength of the positioning signal being lower than a preset minimum valid signal strength threshold, the number of interruptions of the positioning service for receiving the positioning signal being greater than a preset interruption threshold, and the duration of the positioning service being shorter than a preset duration threshold. On this basis, additional requirements may be made for the abnormal quality requirement based on the time dimension, for example, when the signal quality meets the abnormal quality requirement, the corresponding satisfaction duration needs to be greater than the preset target duration. Specifically, during the driving process of the vehicle, if the signal strength of the positioning signal is lower than the preset minimum valid signal strength threshold for a continuous first preset duration, the last corresponding positioning point received with a signal strength greater than the minimum valid signal strength threshold is determined as the starting point of the target road section; and / or, during the driving process of the vehicle, if the signal strength of the positioning signal is higher than the preset minimum valid signal strength threshold for a continuous second preset time, the first corresponding positioning point received with a signal strength greater than the minimum valid signal strength threshold is determined as the end point of the target road section. In this embodiment, on the one hand, the starting point and / or end point of the target road section are not affected by the fixed position information, but are determined according to the actual reception of the vehicle positioning data by the vehicle. Therefore, even in the same geographical location, vehicles with different data reception performance can obtain their own accurate target road section judgments, thereby improving the robustness and flexibility when obtaining corresponding data. On the other hand, by introducing the first preset time length and / or the second preset time length, the misjudgment phenomenon of the starting point and / or end point of the above-mentioned target road section is reduced, and the following situations are avoided: for example, the vehicle fails to receive valid vehicle positioning data in a short period of time due to abnormal reasons such as signal interference in the middle of a normal road section, and then misjudges this place as the starting point of the target road section; or the vehicle accidentally receives valid vehicle positioning data in a short period of time in the middle of the target road section, and then misjudges this place as the end point of the target road section. As a result, the accuracy of identifying the target road section is improved.

[0057] Step S104, obtaining vehicle positioning data of a second positioning point of the vehicle in a connecting section of the target section, wherein the connecting section at least includes a subsequent section of the target section.

[0058] The connecting section of the above-mentioned target section can be regarded as a section connected to the target section, and the section for the vehicle to drive out from the target section can be called the above-mentioned successor section, from which the vehicle positioning data of at least one second positioning point can be selected for use in the subsequent steps to calculate the vehicle positioning data of each first positioning point of the vehicle in the above-mentioned target section. In addition, the section used to enter the target section can also be called a predecessor section, and the above-mentioned second positioning point can be selected from it. In other words, the above-mentioned second positioning point can include the positioning points in the successor section and the positioning points in the predecessor section. Among them, the above-mentioned predecessor section, the above-mentioned target section and the above-mentioned successor section are connected in sequence, and the section formed after the connection can be used to characterize a continuous section of the vehicle.

[0059] It is understandable that, unlike the above-mentioned target road section, the second positioning point in the above-mentioned connecting road section corresponds to valid vehicle positioning data, and a positioning point that is closer to the above-mentioned target road section in terms of spatial position and generation time can be selected from the above-mentioned connecting road section as the above-mentioned second positioning point, thereby ensuring that the vehicle positioning data of each first positioning point in the target road section is more in line with the actual situation. Taking the above-mentioned predecessor road section and successor road section as an example, the second positioning point can be set to include at least the third positioning point and / or the fourth positioning point, and the last positioning point that the vehicle passes through in the predecessor road section is used as the third positioning point included in the second positioning point, and / or the first positioning point that the vehicle passes through in the said successor road section is used as the fourth positioning point included in the second positioning point.

[0060] It should be noted that, although in the case where, for example, the connecting section only includes the above-mentioned successor section, the vehicle motion data of each first positioning point in the target section and the vehicle positioning data of the second positioning point in the connecting section do have a fixed generation order, the generated vehicle motion data and vehicle positioning data can be stored through data caching and other technologies, and then the vehicle motion data of each first positioning point and the vehicle positioning data of the second positioning point can be obtained respectively at any time point after recording according to actual conditions, so as to calculate the vehicle positioning data of the above-mentioned vehicle at each first positioning point in the above-mentioned target section. Therefore, the specific execution order of step S102 and step S104 is not limited in this specification.

[0061] Step S106, calculating the vehicle positioning data of each first positioning point of the vehicle in the target road section at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point.

[0062] After obtaining the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point, the initial vehicle positioning data of each first positioning point of the vehicle in the target road section can be calculated. The initial vehicle positioning data of each first positioning point can actually be generated based on the vehicle positioning data of the second positioning point. Specifically, assuming that the connecting road section only includes the subsequent road section, since the vehicle motion data of each first positioning point provides at least one of the posture, motion trend, speed information, etc. of the vehicle after entering the target road section and before entering the subsequent road section from the target road section, and the vehicle positioning data of the second positioning point also includes the exact position of the second positioning point of the subsequent road section, the initial vehicle positioning data of each previous first positioning point can be reversed from the back to the front based on the vehicle's driving direction according to the vehicle positioning data of the second positioning point of the vehicle in the subsequent road section and the vehicle motion data of each previous first positioning point. If there are multiple second positioning points, the first positioning point that the vehicle passes through in the subsequent road section can be selected as the starting positioning point of the above reverse operation to improve the reverse efficiency. In fact, this is the reverse dead reckoning technology. Correspondingly, assuming that the above-mentioned connecting section includes a predecessor section in addition to the successor section, then, for example, forward dead reckoning technology can also be used, that is, based on the vehicle positioning data of the second positioning point of the vehicle in the predecessor section and the vehicle motion data of the previous first positioning points, and based on the vehicle's driving direction, the initial vehicle positioning data of each first positioning point in the target section is calculated in sequence from front to back.

[0063] As mentioned above, considering that the initial vehicle positioning data of each first positioning point obtained by dead reckoning has inherent data errors, the trajectory generated based on it will also have corresponding drift phenomena, resulting in a large discrepancy between the actual driving trajectory of the vehicle in the target section and the driving trajectory corresponding to the above initial positioning data. This specification provides a set of corresponding correction methods based on the initial vehicle positioning data, so that the corrected initial vehicle positioning data can be used as the vehicle positioning data of each first positioning point of the above-mentioned vehicle in the above-mentioned target section.

[0064] In one embodiment, after generating the initial vehicle positioning data of each first positioning point of the vehicle in the target section at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point, the error fitting function for the initial vehicle positioning data can be determined at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point, and the initial vehicle positioning data can be corrected based on the error fitting function, and the corrected vehicle positioning data can be determined as the target positioning data of the vehicle during the driving process of the target section. Among them, the error fitting function can be used to characterize the error generated by the initial positioning data of each first positioning point in the target section as the driving time in the target section increases. Different sensors for detecting vehicle motion data or different driving states in the target section of the vehicle may cause differences in the error fitting function. In this regard, a corresponding posture error divergence model can be constructed according to the type of the above-mentioned sensor, and at least the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point are substituted into the preset posture error divergence model, so as to generate an error fitting function for the initial positioning point. Of course, in addition to the error fitting function for the above-mentioned initial vehicle positioning data, the solution of this specification can actually also determine the error fitting function for the vehicle motion data of each first positioning point that has been acquired, thereby further reducing the error between the vehicle motion data of each first positioning point and the actual situation.

[0065] The following takes the inertial measurement unit of a vehicle as an example. Assume that the inertial measurement unit is used to detect the vehicle motion data of each first positioning point of the vehicle on the target road section. Assume that the inertial measurement unit includes a gyroscope and an accelerometer. Then, the gyroscope error model can be determined for these two sensors:

[0066]

[0067] And, the accelerometer error model:

[0068]

[0069] In formula (1), for the gyroscope error model, is the measured angular velocity, I is the unit matrix, s g is the scale factor, is the true angular velocity, b g is zero bias, i.e. constant bias, n g is white noise, ignoring the scale factor s g After the influence of It can be determined by the following formula (3):

[0070]

[0071] The main error of the gyroscope is It comes from the zero bias and white noise of the gyroscope. In particular, since the value of white noise has positive and negative differences, after performing mathematical processing such as integration, it can be transformed into a random walk in angle.

[0072] In formula (2), for the accelerometer error model, is the measured acceleration, I is the unit matrix, s a is the scale factor, f b is the true acceleration, b a is the bias, i.e. the constant deviation, n a is white noise, ignoring the scale factor s a After the influence of b The source can be determined by the following formula (4):

[0073] δf b =b a +n a Formula (4)

[0074] That is, the main error of the accelerometer δf b It comes from the bias and white noise of the accelerometer.

[0075] It is understood by those skilled in the art that when calibrating or initially aligning a sensor, a static measurement averaging method can be used to keep the sensor static, i.e., not moving or rotating, and record the measurement values ​​over a period of time, and reduce the influence of white noise by calculating the average value of these static measurement values. However, in practical applications, especially during long-term operation, white noise will still accumulate and gradually affect the accuracy of the system. In addition, even if a relatively accurate zero bias b is obtained through static measurement, g or bias b a ,These parameters may also drift with changes in environmental factors such as time and temperature, resulting in new errors in subsequent measurements. In short, the errors of gyroscopes and accelerometers exist naturally, and will still enter the attitude error divergence model below and cause error accumulation.

[0076] Based on the above gyroscope error model and accelerometer error model, the following posture error divergence model can be determined:

[0077]

[0078]

[0079]

[0080]

[0081]

[0082] In formula (5), As the rate of change of attitude error, it can be obtained by the transformation matrix from the navigation coordinate system n to the body coordinate system (i.e., the coordinate system of the inertial measurement unit) b The matrix is ​​used to convert the error from one coordinate system to another. The error of the gyroscope zero bias is δb g And the angular velocity error δarw caused by the gyroscope white noise is determined. In formula (6), As the rate of change of the velocity error, the actual acceleration f in the inertial coordinate system i can be obtained by converting the vector into a skew-symmetric matrix through the operation function askew. i , the acceleration error δvrw caused by the accelerometer white noise is determined. In formula (7), As the rate of change of position error, the velocity error δv can be obtained. In formula (8), As the rate of change of the angle random walk caused by the gyroscope white noise, it can be expressed by the time constant τ of the gyroscope white noise. gyro , and the angular velocity error δarw caused by the gyroscope white noise. In formula (9), As the rate of change of the rate random walk caused by the acceleration white noise, it can be expressed by the time constant τ of the accelerometer white noise. acc , and the accelerometer error δvrw caused by the gyroscope white noise.

[0083] In summary, the corresponding equations in the above attitude error divergence model describe how the attitude, velocity and position errors change over time in the inertial navigation system corresponding to the inertial measurement unit, and take into account the errors of the gyroscope and accelerometer. Through these models, the accumulation of the above errors can be better understood and predicted. For example, in the pre-integration process for the inertial measurement unit, the drift law of the position error can be fitted by, for example, a quadratic curve about time, thereby obtaining an error fitting function about the position error, such as formula (10) and formula (11) as follows:

[0084] Δr i =0.5·a·t i 2 , i∈[t 0 ,t k ] Formula (10)

[0085] a=2*Δr k / t k 2 Formula (11)

[0086] And the error fitting function about the speed error, such as formula (12) and formula (13) are as follows:

[0087] Δv i = b·t i , i∈[t 0 ,t k ] Formula (12)

[0088] b=Δv k / t k Formula (13)

[0089] Among them, in formula (10) and formula (11), for the error fitting function about the position error, it can be combined with Figure 2 The lower half of Δr i As the three-dimensional position compensation from the i-th state value of the vehicle to t0, a is the fitting parameter of the error curve, t i is the time interval from the i-th state value to t0, where, taking forward dead reckoning as an example, the index value of i at the starting point of dead reckoning (i.e., the second positioning point in the predecessor section) is t0, and at the time of high confidence observation (i.e., the second positioning point in the successor section) is tk. Then, the vehicle can start from, for example, the third positioning point in the predecessor section, pass through the first positioning points of the target section, and end at the fourth positioning point in the successor section. The corresponding relationship between the position error and time in this trajectory can be based on Figure 2 If the reverse dead reckoning is taken as an example, and the connecting section only includes the subsequent section, the index value of i at the starting point of the dead reckoning (i.e., the first positioning point in the target section) is t0, and at the time of the high confidence observation (i.e., the second positioning point in the subsequent section) is tk, then the vehicle can start from, for example, the first first positioning point in the target section, pass through the subsequent first positioning points of the target section, and end at the third positioning point of the subsequent section. The corresponding relationship between the position error and time in this trajectory can be based on Figure 2 The above error fitting function for the position error may correspond to the error fitting function for the vehicle positioning data of each first positioning point in the above text.

[0090] Similarly, in formula (12) and formula (13), for the error fitting function about the speed error, we can combine Figure 2 In the upper part, Δvi is used as the three-dimensional speed compensation from the vehicle's i-th state value to t0, and b is the curve coefficient. Taking forward dead reckoning as an example, the vehicle can start from the third positioning point in the front drive section, pass through the first positioning points in the target section, and end at the fourth positioning point in the subsequent section. The corresponding relationship between the speed error and time in this trajectory can be based on Figure 2 If the reverse dead reckoning is used as an example, and the connecting segment only includes the subsequent segment, then the vehicle can start from, for example, the first first positioning point in the target segment, pass through each subsequent first positioning point of the target segment, and end at the third positioning point of the subsequent segment. The corresponding relationship between the speed error and time in this trajectory can be based on Figure 2 In fact, the above error fitting function about the speed error corresponds to the error fitting function for the vehicle motion data of each first positioning point in the above text.

[0091] Of course, in addition to the quadratic curve and the straight line, other different fitting methods can be used to fit the position error and the speed error according to actual conditions, and this is not limited in this specification.

[0092] It is understandable that in addition to the above-mentioned inertial measurement unit, other posture error divergence models similar to the above-mentioned posture error divergence model can be generated based on the same idea according to the inherent characteristics of sensors such as odometers and the acquired vehicle motion data, and this is not limited in this specification.

[0093] In addition, the above error fitting function may require the vehicle motion data of each positioning point including the target section and the connecting section, which means that the second positioning point will lack the corresponding vehicle motion data when it only contains vehicle positioning data. In view of this situation, a mathematical model of vehicle motion based on the principles of classical mechanics or an artificial intelligence model based on machine learning can be established, and the vehicle motion data of the second positioning point can be calculated based on the vehicle motion data of each first positioning point and the initial vehicle positioning data, and the second positioning point only contains vehicle positioning data, or, when the second positioning point is the third positioning point and / or the fourth positioning point closest to the first positioning point in the target section, the vehicle motion data of the second positioning point is approximately determined as the vehicle motion data of the closest first positioning point in the target section, which is not limited in this specification.

[0094] Of course, if the second positioning point in the connecting section itself corresponds to vehicle motion data, then the vehicle motion data of the second positioning point of the vehicle in the connecting section of the above-mentioned target section can be obtained, and the above-mentioned error fitting function can be directly determined based on the vehicle motion data of the first positioning point, the vehicle positioning data of the second positioning point, and the vehicle motion data of the second positioning point, thereby calculating the vehicle positioning data of each first positioning point of the above-mentioned vehicle in the target section.

[0095] Further, in the case where the second positioning point at least includes the above-mentioned third positioning point and the above-mentioned fourth positioning point, assuming that the third positioning point and the above-mentioned fourth positioning point both correspond to vehicle motion data, then, in the first case, the vehicle motion data of the vehicle at the above-mentioned third positioning point can be obtained, and the vehicle positioning data of each first positioning point of the above-mentioned vehicle in the target road section can be calculated based on the vehicle motion data of the above-mentioned first positioning point, the vehicle positioning data of the third positioning point, the vehicle positioning data of the third positioning point, and the vehicle positioning data of the fourth positioning point; or, in the second case, the vehicle motion data of the vehicle at the above-mentioned fourth positioning point can be obtained, and the vehicle positioning data of each first positioning point of the above-mentioned vehicle in the target road section can be calculated based on the vehicle motion data of the above-mentioned first positioning point, the vehicle positioning data of the third positioning point, the vehicle positioning data of the fourth positioning point, and the vehicle motion data of the fourth positioning point. Among them, compared with calculating the vehicle positioning data of each first positioning point based on the positioning data and motion data of the positioning points in the predecessor and successor sections at the same time, the above two situations respectively give up obtaining the vehicle motion data of the positioning points in the predecessor or successor section to participate in the above calculation process. This method can reduce the amount of data that needs to be prepared in advance without significantly affecting the accuracy of the calculation results, thereby helping to improve the utilization of computing resources.

[0096] After determining the error fitting function, the initial vehicle positioning data may be corrected based on the error fitting function, wherein the correction process may be understood as compensating the vehicle positioning data of each first positioning point according to the corresponding error fitting function to make it closer to the actual situation.

[0097] After obtaining the corrected vehicle positioning data of each first positioning point, the driving trajectory of the vehicle in the target section can be generated to characterize the driving trajectory of the vehicle in the target section. In addition to simply connecting the first positioning points in order according to the vehicle positioning data, the factor graph optimization algorithm (FGO) can also be introduced to fuse the vehicle motion data including at least each first positioning point in the target section with the vehicle positioning data to obtain a driving trajectory that is closer to the actual situation. Figure 3 The above vehicle trajectory prediction method is described. Figure 1 As shown, the above method includes:

[0098] Step S302: Based on the vehicle positioning method, the vehicle motion data of each first positioning point of the vehicle in the target road section is obtained.

[0099] As mentioned above, according to the vehicle positioning method described in the above embodiments of the specification, the vehicle motion data of each first positioning point of the vehicle in the target road section can be accurately obtained, so it will not be repeated here in this specification.

[0100] Step S304: generating a driving trajectory of the vehicle on the target road section according to the vehicle positioning data of each first positioning point of the vehicle on the target road section.

[0101] In one embodiment, the absolute position constraint factor in the factor graph can be constructed based on the vehicle positioning data of the first positioning point, and the relative motion constraint factor can be constructed based on the vehicle motion data of the first positioning point. Then, the factor graph composed of the absolute position constraint factor and the relative motion constraint factor can be optimized to obtain the optimal solution of the factor graph, and the driving trajectory of the vehicle in the target road section can be generated based on each positioning point represented by the optimal solution. Figure 4 For example, it is assumed that the inertial measurement unit detects the vehicle motion data of each first positioning point of the vehicle in the target section and the vehicle positioning data of the second positioning point in the connecting section, and each positioning point can maintain the positioning state information recording the corresponding vehicle motion data and vehicle positioning data, wherein the state information can include the position information, speed and attitude in the world coordinate system, and the constant deviation of the gyroscope and acceleration, and as Figure 4 The content of each location state node in the factor graph is n, and the window length of the sliding window of the factor graph is n. Then the location state node x at the i-th moment (i∈[0,n]) in the sliding window i It can be expressed by formula (14):

[0102] x i =[p n ,v n ,q n ,b g ,b a ,s odo ] Formula (14)

[0103] In formula (14), p n is the position state of the navigation system, v n is the 3D velocity in the navigation system, q n is the 3D posture of the navigation system vehicle, b g ,b a are the constant bias of the gyroscope and the constant bias of the accelerometer, s odo It is the vehicle speed proportional factor error state.

[0104] The vehicle positioning data for the second positioning point in the connecting section can be used as Figure 4 The GNSS observation nodes in the link are then used as the GNSS observation factors of each positioning state node in the link section, i.e., the absolute position constraint factors. In an open environment, GNSS can provide high-precision absolute position information. The residual expression of the GNSS factor is gnss (y gnss, X) is expressed by formula (15) as follows:

[0105]

[0106] Among them, the noise variance matrix of GNSS is N gnss , r gnss is the gnss residual at time i in the window, Measure location information for GNSS, is the lever arm from the GNSS to the IMU center of mass in the b frame. is the rotation matrix at time i.

[0107] For the target road section and the connecting road section, considering that both the vehicle motion data of each first positioning point and the vehicle motion data of the second positioning point (if not available, it can be inferred according to the method described above) can be detected by the inertial measurement unit, all the above vehicle motion data can be pre-integrated according to the order of each first positioning point and the second positioning point, wherein the pre-integration processing is particularly suitable for processing a large amount of continuous data from high-frequency sensors, such as vehicle motion data detected by the inertial measurement unit, and the processed pre-integration factor is used as the relative motion constraint factor between the corresponding positioning state node and the previous positioning state node. Among them, the pre-integration processing is intended to simplify and optimize the processing of the raw data provided by the inertial measurement unit. It reduces the workload of real-time processing and improves the efficiency and accuracy of the algorithm by accumulating and calculating the data detected by the inertial measurement unit in advance. Of course, for sensors such as visual odometers whose first relative motion data and second relative motion data are not high-frequency or continuous, the above pre-integration processing can be omitted and directly used as the above relative motion constraint factor. On the basis of the above pre-integration, the above positioning state nodes can also obtain corresponding support, such as b g ,b a That is, it can be estimated directly through the pre-integration factor of the IMU.

[0108] The actual construction process of the pre-integration factor based on the inertial measurement unit can be understood as: calculating the state change of the inertial measurement unit between two calculation cycle intervals to form the pre-integration factor of the inertial measurement unit, wherein the estimated state at the current moment is obtained based on the previous state, which can maintain a high accuracy in a short time. The residual expression of the pre-integration factor is expressed as follows through formula (16):

[0109]

[0110] in, represents the pre-integrated residual, Represents the attitude rotation matrix from the b system to the world system coordinate w at time tk-1, and are the position states at time tk and time tk-1, respectively, and are used as optimization parameters to participate in the residual calculation of the pre-integrated position difference. are the speed states at time tk and time tk-1, respectively, and are used as optimization parameters to participate in the residual calculation of the pre-integrated speed estimation. is the carrier attitude quaternion at time k, is the change of vehicle posture during the integration period, The position increment constraint position state increment calculated by the vehicle speed integral within the integration period, s odok is the odometer error proportional factor, which participates in the vehicle speed compensation constraint. In this residual expression, the position and velocity interference term caused by the earth rotation term between tk-1 and tk is simplified, and the influence under the use conditions of medium and low precision inertial sensors can be controlled.

[0111] In addition, if other sensors besides the inertial measurement unit can also be used to detect the vehicle motion data of each first positioning point and the vehicle motion data of the second positioning point, such as an odometer, then the detected data can be added to the factor graph as an additional observation factor to enhance the robustness of positioning.

[0112] In short, after constructing the absolute position constraint factor and the relative motion constraint factor, we can construct the following based on the preset priori factors and the above positioning state nodes: Figure 4 The complete factor graph is shown in Figure 1, and then the optimization process is performed. The main optimization objective function is shown in formula (17):

[0113]

[0114] Among them, the first term is the marginalized residual, r p and H p It is the prior information in edge optimization. The second term r pre is the pre-integrated residual, and the third term is r gnss The gnss factor residual m is the number of GNSS factors.

[0115] Based on the objective function, an optimal solution of the factor graph can be obtained, and based on each positioning point represented by the optimal solution, a driving trajectory of the vehicle in the target road section is generated.

[0116] Combine the following Figure 5 The optimization method based on factor graph is further explained. Figure 5 As shown in FIG, the process of the factor graph-based optimization method includes the following steps:

[0117] Step S502: pre-processing the data of GNSS, IMU and ODO.

[0118] In one embodiment, for the preceding trajectory point data, the first relative motion data, and the subsequent trajectory point data originating from the GNSS, the inertial measurement unit, and the odometer, at least any one of the following can be performed: data denoising, data filtering, data time synchronization, anomaly detection, and the detected abnormal data can be processed according to the anomaly handling strategy. Specifically, denoising and filtering techniques can be used to remove obviously abnormal GNSS observations, and the inertial measurement unit data can be filtered to reduce noise interference. Alternatively, ensure that the timestamps of all sensor data are consistent to facilitate subsequent fusion processing. Alternatively, finite interpolation methods can be used to supplement missing or abnormal GNSS and inertial measurement unit data to ensure data continuity.

[0119] Step S504: construct a factor graph including GNSS observation factors, IMU measurement factors and ODO measurement factors according to the preprocessed data.

[0120] In one embodiment, for the GNSS, IMU and ODO data of each first positioning point after preprocessing, the GNSS observation can be constructed as an absolute position constraint factor, the IMU data can be pre-integrated to be constructed as a relative motion constraint factor, and the ODO data can be directly used as a relative motion constraint factor or an absolute position constraint factor, thereby forming the above-mentioned factor graph.

[0121] Step S506: The factor graph is optimized to obtain an optimal solution.

[0122] In one embodiment, the factor graph is iteratively optimized by an optimization algorithm (such as Ceres Solver) to solve the optimal variable value (i.e., position, velocity, attitude, etc.). The specific process may include defining an optimization objective function, including marginalized residuals, pre-integrated residuals, and GNSS factor residuals to optimize the objective function, and using methods such as maximum a posteriori probability estimation or least squares method to gradually approach the optimal solution through multiple iterations to ensure that the output results maintain overall consistency.

[0123] Step S508: output the optimized result as target positioning data.

[0124] In one embodiment, the optimization result of the previous step may be output as target positioning data for characterizing the driving trajectory of the vehicle in the target road section.

[0125] In addition, this solution can also change the optimization processing logic for each node in the sliding window, so that without adopting methods that consume a lot of computing resources such as expanding the window, detecting key frames and adding loop factors, the output results of each section including the target section can still be guaranteed to have high accuracy.

[0126] In one embodiment, the factor graph is provided with a sliding window of a preset length, and in the process of the sliding window sliding in sequence according to the preset step length, each time the sliding window slides once, each node in the sliding window can be optimized once. In other words, for each node in the factor graph, if the node is still in the sliding window after the sliding window slides once, the node is optimized until the node is outside the sliding window after the sliding window slides once, and the optimization of the node is stopped. Of course, in the process of the vehicle entering the target section, the above-mentioned repeated optimization method has a low optimization effect on the positioning point that only generates the first relative motion data in real time, so it can be selectively used in the connecting section to reduce the computing power overhead. The following is Figure 6 For example, assume that there are positioning state nodes y0~yn in the factor graph, n is the total number of all pre-integration cycles, and k is the window length of the sliding window. Then in the time period corresponding to time t2 to time tn-k, the vehicle enters the target section and cannot receive the first absolute position data of each positioning point in the section. Therefore, each node can be temporarily not optimized and output until the target positioning data of the target section driving process is determined after entering the backward connection section and then output; in the time period corresponding to time tn-k to time tk, only the optimization results of the nodes sliding out of the sliding window are output each time.

[0127] The above process of repeatedly optimizing nodes requires at least one cycle corresponding to the length of the sliding window before being output. Compared with the real-time output method, the above can be called a semi-real-time output. Its advantage is that the accuracy of the corresponding positioning data of each node in the factor graph is maximized with less computing power and time cost. Figure 7 The real-time output trajectory and the semi-real-time output trajectory in the figure show that the positioning points between the target section and the connecting section in the semi-real-time output trajectory are more naturally connected, and the corresponding driving trajectory is more in line with the actual situation.

[0128] Figure 8 is a schematic structural diagram of a device provided by an exemplary embodiment. Figure 8At the hardware level, the device includes a processor 802, an internal bus 804, a network interface 806, a memory 808, and a non-volatile memory 810, and may also include hardware required for other functions. One or more embodiments of this specification may be implemented based on software, such as the processor 802 reading the corresponding computer program from the non-volatile memory 810 into the memory 808 and then running it. Of course, in addition to the software implementation, one or more embodiments of this specification do not exclude other implementations, such as logic devices or a combination of software and hardware, etc., that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0129] Corresponding to the above-mentioned embodiment of the vehicle positioning method, this specification also provides an embodiment of a vehicle positioning device. Fig. 9 ; The device comprises:

[0130] The vehicle motion data acquisition unit 902 is used to acquire the vehicle motion data of the vehicle at each first positioning point in the target road section, wherein the vehicle motion data is determined based on the posture and / or speed information of the vehicle at the first positioning point; and

[0131] A vehicle positioning data acquisition unit 904, configured to acquire vehicle positioning data of a second positioning point of the vehicle in a connecting section of the target section, wherein the connecting section at least includes a subsequent section of the target section;

[0132] The vehicle positioning data calculation unit 906 is used to calculate the vehicle positioning data of each first positioning point of the vehicle in the target road section based on at least the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point.

[0133] Optionally, the second positioning point includes at least a third positioning point and / or a fourth positioning point, the third positioning point is the last positioning point that the vehicle passes through in the preceding road section, and the fourth positioning point is the first positioning point that the vehicle passes through in the succeeding road section; the preceding road section is a connecting road section of the target road section;

[0134] and / or

[0135] The target road section includes a road section where the signal quality of the positioning signal of the vehicle meets the abnormal quality requirement, and the positioning signal is used to parse the vehicle positioning data;

[0136] and / or

[0137] The vehicle motion data includes at least one of the displacement data, total acceleration, linear acceleration, angular velocity, rotation rate, gravity vector, attitude angle, and speed of the vehicle.

[0138] Optionally, the vehicle positioning data calculation unit 906 is specifically used to:

[0139] Acquire vehicle motion data of a second positioning point of the vehicle in the connecting section of the target section;

[0140] The vehicle positioning data of each first positioning point of the vehicle in the target road section is calculated according to the vehicle motion data of the first positioning point, the vehicle positioning data of the second positioning point, and the vehicle motion data of the second positioning point.

[0141] Optionally, when the second positioning point includes at least a third positioning point and a fourth positioning point, the vehicle positioning data calculation unit 906 is specifically used to:

[0142] Obtain vehicle motion data of the vehicle at the third positioning point; and, based on the vehicle motion data of the first positioning point, the vehicle positioning data of the third positioning point, the vehicle motion data of the third positioning point, and the vehicle positioning data of the fourth positioning point, calculate the vehicle positioning data of each first positioning point of the vehicle in the target road section; or obtain vehicle motion data of the vehicle at the fourth positioning point; and, based on the vehicle motion data of the first positioning point, the vehicle positioning data of the third positioning point, the vehicle positioning data of the fourth positioning point, and the vehicle motion data of the fourth positioning point, calculate the vehicle positioning data of each first positioning point of the vehicle in the target road section.

[0143] Optionally, the vehicle positioning data calculation unit 906 is specifically used to:

[0144] Generate initial vehicle positioning data of each first positioning point of the vehicle in the target road section based at least on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point;

[0145] An error fitting function for the initial vehicle positioning point is determined at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point, the initial vehicle positioning data is corrected based on the error fitting function, and the corrected initial vehicle positioning data is used as the vehicle positioning data of each first positioning point of the vehicle in the target road section.

[0146] Optionally, the vehicle positioning data calculation unit 906 is specifically used to:

[0147] At least the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point are substituted into a preset posture error divergence model to generate an error fitting function for the initial vehicle positioning point.

[0148] Corresponding to the above-mentioned embodiment of the vehicle trajectory prediction method, this specification also provides an embodiment of a vehicle trajectory prediction device. Fig.10 ; The device comprises:

[0149] The vehicle motion data acquisition unit 1002 is used to acquire the vehicle motion data of each first positioning point of the vehicle in the target road section according to the vehicle positioning method described in the above embodiment of the specification;

[0150] The driving trajectory generating unit 1004 is used to generate the driving trajectory of the vehicle on the target road section according to the vehicle positioning data of each first positioning point of the vehicle in the target road section.

[0151] Optionally, the driving trajectory generating unit 1004 is specifically used to:

[0152] constructing an absolute position constraint factor according to the vehicle positioning data of the first positioning point, and constructing a relative motion constraint factor according to the vehicle motion data of the first positioning point;

[0153] The factor graph composed of the absolute position constraint factor and the relative motion constraint factor is optimized to obtain an optimal solution of the factor graph, and the driving trajectory of the vehicle in the target road section is generated based on each positioning point represented by the optimal solution.

[0154] Optionally, the driving trajectory generating unit 1004 is specifically used to:

[0155] According to the sequence of the first positioning points corresponding to the vehicle motion data of the first positioning points, pre-integration processing is performed on the target relative motion data in the vehicle motion data of the first positioning points, and the pre-integration factor obtained by the processing is used as the relative motion constraint factor;

[0156] and / or

[0157] In the process that the sliding window slides sequentially according to the preset step length, each time the sliding window slides, each node in the sliding window is optimized.

[0158] Based on the same concept as the above method, this specification also provides an electronic device, including: a processor; a memory for storing processor executable instructions; wherein the processor implements the steps of the method described in any of the above embodiments by running the executable instructions.

[0159] Based on the same concept as the above method, this specification also provides a computer-readable storage medium on which computer instructions are stored. When the instructions are executed by a processor, the steps of the method described in any of the above embodiments are implemented.

[0160] Based on the same concept as the above method, this specification also provides a computer program product, including a computer program / instruction, which implements the steps of the method described in any of the above embodiments when executed by a processor.

[0161] Although the present invention includes many specific implementation details, these should not be interpreted as limiting the scope of any invention or the scope of protection claimed, but are mainly used to describe the features of the specific embodiments of specific inventions. Certain features described in multiple embodiments of the present invention may also be implemented in combination in a single embodiment. On the other hand, the various features described in a single embodiment may also be implemented separately in multiple embodiments or in any suitable sub-combination. In addition, although features may work in certain combinations as described above and even initially claimed as such, one or more features from the claimed combination may be removed from the combination in some cases, and the claimed combination may point to a sub-combination or a variation of a sub-combination.

[0162] Similarly, although operations are depicted in a particular order in the accompanying drawings, this should not be understood as requiring that these operations be performed in the particular order shown or performed sequentially, or requiring that all illustrated operations be performed to achieve the desired results. In some cases, multitasking and parallel processing may be advantageous. In addition, the separation of various system modules and components in the above-described embodiments should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product, or packaged into multiple software products.

[0163] Thus, specific embodiments of the subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the particular order or sequential order shown to achieve the desired results. In some implementations, multitasking and parallel processing may be advantageous.

[0164] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A vehicle positioning method, characterized in that: include: Acquire vehicle motion data of the vehicle at each first positioning point in the target road section, wherein the vehicle motion data is determined based on the posture and / or speed information of the vehicle at the first positioning point; and, Acquire vehicle positioning data of a second positioning point of the vehicle in a connecting section of the target section, wherein the connecting section at least includes a subsequent section of the target section; The vehicle positioning data of each first positioning point of the vehicle in the target road section is calculated at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point.

2. The method according to claim 1, characterized in that The second positioning point includes at least a third positioning point and / or a fourth positioning point, wherein the third positioning point is the last positioning point that the vehicle passes through in the preceding road section, and the fourth positioning point is the first positioning point that the vehicle passes through in the succeeding road section; the preceding road section is a connecting road section of the target road section; and / or The target road section includes a road section where the signal quality of the positioning signal of the vehicle meets the abnormal quality requirement, and the positioning signal is used to parse the vehicle positioning data; and / or The vehicle motion data includes at least one of the displacement data, total acceleration, linear acceleration, angular velocity, rotation rate, gravity vector, attitude angle, and speed of the vehicle.

3. The method according to claim 1, characterized in that The calculating the vehicle positioning data of each first positioning point of the vehicle in the target road section includes: Acquire vehicle motion data of a second positioning point of the vehicle in the connecting section of the target section; The vehicle positioning data of each first positioning point of the vehicle in the target road section is calculated according to the vehicle motion data of the first positioning point, the vehicle positioning data of the second positioning point, and the vehicle motion data of the second positioning point.

4. The method according to claim 2, characterized in that: When the second positioning points include at least a third positioning point and a fourth positioning point, calculating the vehicle positioning data of each first positioning point of the vehicle in the target road section at least based on the vehicle motion data of the first positioning points and the vehicle positioning data of the second positioning points includes: Acquire vehicle motion data of the vehicle at the third positioning point; and calculate the vehicle positioning data of each first positioning point of the vehicle in the target road section based on the vehicle motion data of the first positioning point, the vehicle positioning data of the third positioning point, the vehicle motion data of the third positioning point, and the vehicle positioning data of the fourth positioning point; or Acquire vehicle motion data of the vehicle at the fourth positioning point; and calculate the vehicle positioning data of each first positioning point of the vehicle in the target section based on the vehicle motion data of the first positioning point, the vehicle positioning data of the third positioning point, the vehicle positioning data of the fourth positioning point, and the vehicle motion data of the fourth positioning point.

5. The method according to claim 1, characterized in that The step of calculating the vehicle positioning data of each first positioning point of the vehicle in the target road section at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point comprises: Generate initial vehicle positioning data of each first positioning point of the vehicle in the target road section based at least on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point; An error fitting function for the initial vehicle positioning point is determined at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point, the initial vehicle positioning data is corrected based on the error fitting function, and the corrected initial vehicle positioning data is used as the vehicle positioning data of each first positioning point of the vehicle in the target road section.

6. The method according to claim 5, characterized in that The step of determining the error fitting function for the initial vehicle positioning point at least based on the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point comprises: At least the vehicle motion data of the first positioning point and the vehicle positioning data of the second positioning point are substituted into a preset posture error divergence model to generate an error fitting function for the initial vehicle positioning point.

7. A vehicle trajectory prediction method, characterized in that: include: Based on the vehicle positioning method according to any one of claims 1 to 6, obtaining vehicle motion data of each first positioning point of the vehicle in the target road section; A driving trajectory of the vehicle on the target road section is generated according to the vehicle positioning data of each first positioning point of the vehicle on the target road section.

8. The method according to claim 7, characterized in that The step of generating a driving trajectory of the vehicle on the target road section according to the vehicle positioning data of each first positioning point of the vehicle on the target road section comprises: constructing an absolute position constraint factor according to the vehicle positioning data of the first positioning point, and constructing a relative motion constraint factor according to the vehicle motion data of the first positioning point; The factor graph composed of the absolute position constraint factor and the relative motion constraint factor is optimized to obtain an optimal solution of the factor graph, and the driving trajectory of the vehicle in the target road section is generated based on each positioning point represented by the optimal solution.

9. The method according to claim 8, characterized in that Constructing the relative motion constraint factor according to the vehicle motion data of the first positioning point includes: performing pre-integration processing on the target relative motion data in the vehicle motion data of the first positioning point according to the sequence of the first positioning points corresponding to the vehicle motion data of the first positioning point, and using the pre-integration factor obtained by the processing as the relative motion constraint factor; and / or The factor graph is provided with a sliding window of a preset length; the factor graph composed of the absolute position constraint factor and the relative motion constraint factor is optimized, including: in the process of the sliding window sliding in sequence according to the preset step length, each time the sliding window slides, each node in the sliding window is optimized once.

10. An electronic device, characterized in that: include: processor; A memory for storing processor-executable instructions; wherein the processor implements the steps of the method according to any one of claims 1-6 or 7-9 by running the executable instructions.

11. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the instructions are executed by a processor, the steps of the method as described in any one of claims 1-6 or 7-9 are implemented.

12. A computer program product, characterized in that The method comprises a computer program / instruction which, when executed by a processor, implements the steps of the method as claimed in any one of claims 1 to 6 or 7 to 9.

Images