Vehicle positioning method and device based on ground slope and electronic equipment

Abstract

The application provides a vehicle positioning method and device based on ground slope and an electronic device. The method comprises the following steps: acquiring a vehicle pitch angle sequence, a road slope angle sequence and a current predicted position of a vehicle in a current time period, the vehicle pitch angle sequence being used to represent a pitch angle sequence of the vehicle relative to a horizontal plane, and the road slope angle sequence being road slope data extracted from a map according to a road on which the vehicle travels; performing feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence to obtain a pitch angle feature sequence and a slope angle feature sequence; performing feature matching processing on the pitch angle feature sequence and the slope angle feature sequence to calculate a matching score; and performing correction processing on the current predicted position according to the matching score to obtain a current driving position of the vehicle. The application solves the problem of poor vehicle positioning accuracy in the prior art in a scenario where GNSS signals are weakened or disappeared due to obstruction.

Description

Technical Field

[0001] This application relates to the field of vehicle positioning technology based on ground slope, and more specifically, to a vehicle positioning method, apparatus, computer-readable storage medium, and electronic device based on ground slope. Background Technology

[0002] In the fields of autonomous driving and vehicle positioning, vehicle position is typically determined by a combination of GNSS (Global Navigation Satellite System) and IMU (Inertial Measurement Unit) navigation (such as ESKF, Error State Kalman Filter).

[0003] However, in scenarios where GNSS signals are weakened or disappear due to obstruction, such as tunnels, underground parking garages, or densely built-up "urban canyons," some existing correction methods rely mainly on IMU integration to calculate the position. Due to the zero bias and noise of the IMU, the ground slope calculation deviates over time, and the vehicle's longitudinal (direction of travel) position error accumulates continuously, resulting in drift. Therefore, existing technologies suffer from poor vehicle positioning accuracy in scenarios where GNSS signals are weakened or disappear due to obstruction. Summary of the Invention

[0004] The main objective of this application is to provide a vehicle positioning method, device, computer-readable storage medium, and electronic device based on ground slope, so as to at least solve the problem of poor vehicle positioning accuracy in existing solutions when GNSS signals are weakened or lost due to obstruction.

[0005] To achieve the above objectives, according to one aspect of this application, a vehicle positioning method based on ground slope is provided, comprising: acquiring a vehicle pitch angle sequence, a road slope angle sequence, and the vehicle's current predicted position within a current time period, wherein the vehicle pitch angle sequence is acquired by vehicle sensors and is used to characterize the vehicle's pitch angle sequence relative to a horizontal plane, and the road slope angle sequence is road slope data extracted from a map based on the road the vehicle is traveling on; performing feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain a pitch angle feature sequence and a slope angle feature sequence, and performing feature matching processing on the pitch angle feature sequence and the slope angle feature sequence to calculate a matching score between the pitch angle feature sequence and the slope angle feature sequence; and correcting the current predicted position based on the matching score to obtain the vehicle's current driving position.

[0006] Optionally, feature matching processing is performed on the pitch angle feature sequence and the slope angle feature sequence to calculate the matching score between the pitch angle feature sequence and the slope angle feature sequence, including: constructing a candidate displacement set, and translating the slope angle feature sequence according to each candidate displacement in the candidate displacement set to obtain the translated slope angle feature sequence; performing feature matching processing on the pitch angle feature sequence and the translated slope angle feature sequence to obtain the matching score corresponding to each candidate displacement.

[0007] Optionally, feature matching processing is performed on the pitch angle feature sequence and the translated slope angle feature sequence to obtain the matching score corresponding to each candidate displacement, including: calculating the similarity of the overlapping area between the pitch angle feature sequence and each translated slope angle feature sequence, and using each similarity as the matching score corresponding to each candidate displacement.

[0008] Optionally, the current predicted position is corrected based on the matching score to obtain the current driving position of the vehicle, including: converting the matching score corresponding to each candidate displacement into a probability distribution, wherein the probability distribution represents the probability that the vehicle is located at each candidate displacement; performing a weighted average calculation on each candidate displacement based on the probability of each candidate displacement in the probability distribution to obtain a target displacement; and correcting the current predicted position based on the target displacement to obtain the current driving position of the vehicle.

[0009] Optionally, after calculating the target displacement by performing a weighted average calculation on each candidate displacement according to the probability of each candidate displacement in the probability distribution, the method further includes: determining the dispersion of the probability distribution according to the probability of each candidate displacement in the probability distribution, and evaluating the confidence level of the target displacement according to the dispersion.

[0010] Optionally, feature extraction processing is performed on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain pitch angle feature sequences and slope angle feature sequences, including: using a feature extraction model to perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain pitch angle feature sequences and slope angle feature sequences, wherein the feature extraction model is a convolutional neural network model or a recurrent neural network model.

[0011] Optionally, after obtaining the vehicle pitch angle sequence, road slope angle sequence, and the vehicle's current predicted position within the current time period, the method further includes: performing sequence normalization operations on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain normalized vehicle pitch angle sequence and road slope angle sequence, wherein the sequence normalization operation includes at least one of mean removal operation and scale normalization operation.

[0012] According to another aspect of this application, a vehicle positioning device based on ground slope is provided, comprising: an acquisition unit, configured to acquire a vehicle pitch angle sequence, a road slope angle sequence, and the vehicle's current predicted position within a current time period, wherein the vehicle pitch angle sequence is acquired by vehicle sensors and is used to characterize the vehicle's pitch angle sequence relative to a horizontal plane, and the road slope angle sequence is road slope data extracted from a map based on the road the vehicle is traveling on; a feature extraction processing unit, configured to perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain a pitch angle feature sequence and a slope angle feature sequence, and perform feature matching processing on the pitch angle feature sequence and the slope angle feature sequence to calculate a matching score between the pitch angle feature sequence and the slope angle feature sequence; and a correction processing unit, configured to correct the current predicted position based on the matching score to obtain the vehicle's current driving position.

[0013] According to another aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform any of the described vehicle positioning methods based on ground slope.

[0014] According to another aspect of this application, an electronic device is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including methods for performing any of the described vehicle positioning methods based on ground slope.

[0015] By applying the technical solution of this application, the pitch angle sequence collected by the vehicle's onboard sensors during the current time period and the ground slope angle sequence extracted from the high-precision map based on the current predicted position are obtained. Features are extracted from the pitch angle sequence and the ground slope angle sequence to form pitch angle feature sequences and slope angle feature sequences with temporal structure and terrain semantic information. The similarity score is calculated through feature matching to establish a direct correlation between the vehicle's motion posture and the road terrain morphology. In the GNSS signal occlusion environment, the matching score can effectively reflect the consistency between the vehicle's actual driving trajectory and the map terrain model, thereby dynamically correcting the longitudinal positioning error caused by IMU integral drift. This achieves highly robust positioning without relying on external visual or lidar features, based solely on the vehicle's own sensors and map slope information. Thus, the problem of longitudinal positioning error accumulation caused by IMU drift and lack of effective shape constraints in the GNSS signal occlusion environment in the prior art is solved, achieving the effect of continuously and stably constraining positioning drift and improving the vehicle's longitudinal positioning accuracy and system reliability even in the absence of satellite signals. This solves the problem of poor vehicle positioning accuracy in existing solutions when GNSS signals are weakened or lost due to obstruction. Attached Figure Description

[0016] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0017] Figure 1 A hardware structure block diagram of a mobile terminal for performing a vehicle positioning method based on ground slope, according to an embodiment of this application, is shown.

[0018] Figure 2 A schematic flowchart of a vehicle positioning method based on ground slope according to an embodiment of this application is shown.

[0019] Figure 3 A schematic flowchart of a specific vehicle positioning method based on ground slope provided according to an embodiment of this application is shown;

[0020] Figure 4 A schematic diagram comparing vehicle pitch angle sequences with road slope angle sequences provided according to embodiments of this application is shown.

[0021] Figure 5 A structural block diagram of a vehicle positioning device based on ground slope provided according to an embodiment of this application is shown. Detailed Implementation

[0022] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0024] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] As described in the background section, existing solutions have poor vehicle positioning accuracy in scenarios where GNSS signals are weakened or disappear due to obstruction. To address this issue, embodiments of this application provide a vehicle positioning method, apparatus, computer-readable storage medium, and electronic device based on ground slope.

[0026] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0027] The methods and embodiments provided in this application can be executed on a mobile terminal, computer terminal, or similar computing device. Taking running on a mobile terminal as an example, Figure 1 This is a hardware structure block diagram of a mobile terminal for a vehicle positioning method based on ground slope according to an embodiment of the present invention. Figure 1 As shown, a mobile terminal may include one or more ( Figure 1Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. The mobile terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the mobile terminal described above. For example, the mobile terminal may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0028] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the vehicle positioning method based on ground slope in this embodiment of the invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. The transmission device 106 is used to receive or send data via a network. Specific examples of the aforementioned networks may include wireless networks provided by the mobile terminal's communication provider. In one example, the transmission device 106 includes a network interface controller (NIC), which can be connected to other network devices via a base station to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0029] This embodiment provides a vehicle positioning method based on ground slope that runs on a mobile terminal, computer terminal or similar computing device. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Also, although the logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0030] Figure 2 This is a flowchart of a vehicle positioning method based on ground slope according to an embodiment of this application. Figure 2 As shown, the method includes the following steps:

[0031] Step S201: Obtain the vehicle pitch angle sequence, road slope angle sequence, and the current predicted position of the vehicle in the current time period. The vehicle pitch angle sequence is collected by the vehicle sensor and is used to characterize the pitch angle sequence of the vehicle relative to the horizontal plane. The road slope angle sequence is road slope data extracted from the map based on the road the vehicle is traveling on.

[0032] Specifically, firstly, the pitch angle sequence of the vehicle collected by sensors within the current time period is acquired. This sequence reflects the continuous value of the instantaneous pitch attitude of the vehicle relative to the horizontal plane over time, serving as real-time observation data of the longitudinal attitude during vehicle movement. Simultaneously, the road slope angle sequence extracted from a high-precision map and associated with the vehicle's current predicted position is acquired. This sequence characterizes the longitudinal slope variation features of the ground preset in the map, and its spatial distribution corresponds to the vehicle's driving path. By aligning the vehicle's real-time pitch angle sequence with the road slope angle sequence at the corresponding position in the map, raw input data is provided for subsequent longitudinal positioning correction based on sequence features. Both are constructed based on vehicle position information within the same spatiotemporal reference frame, forming a comparable observation and reference signal pair.

[0033] Step S202: Perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain pitch angle feature sequence and slope angle feature sequence, and perform feature matching processing on the pitch angle feature sequence and the slope angle feature sequence to calculate the matching score of the pitch angle feature sequence and the slope angle feature sequence.

[0034] Specifically, firstly, the pitch angle sequence acquired in real time during vehicle movement is processed by feature extraction. The original angle change information is transformed into a discriminative feature representation through a specific encoding mechanism, forming a pitch angle feature sequence. Simultaneously, the ground slope angle sequence of the corresponding road segment is extracted from the high-precision map and converted into a slope angle feature sequence using the same feature extraction method, ensuring that the two are comparable in the feature space. Subsequently, the pitch angle feature sequence and the slope angle feature sequence are aligned and matched. By calculating their spatial similarity, a matching score reflecting their geometric consistency is obtained. This score directly characterizes the degree of morphological fit between the vehicle's current position and the map reference sequence, providing a quantitative basis for subsequent positioning correction.

[0035] Step S203: Correct the current predicted position based on the matching score to obtain the current driving position of the vehicle.

[0036] Specifically, the current predicted position is corrected based on the matching score to obtain the vehicle's current driving position. This means that the system determines the vehicle's relative offset in the longitudinal direction based on the similarity score calculated during sequence matching, and adjusts its previously estimated position accordingly. The matching score reflects the degree of fit between the real-time observed slope sequence and the map reference slope sequence under different displacement assumptions; a higher score indicates that the corresponding displacement is closer to the true deviation. Through this scoring result, the system directly applies a correction to the predicted position consistent with the displacement indicated by the score, thereby outputting an updated vehicle driving position. This position no longer depends solely on inertial estimation but is driven by the matching results of terrain slope features.

[0037] In this embodiment, by applying the above steps S201, S202, and S203, the pitch angle sequence collected by the vehicle's onboard sensors during the current time period and the ground slope angle sequence extracted from the high-precision map based on the current predicted position are obtained. Features are extracted from the pitch angle sequence and the ground slope angle sequence to form pitch angle feature sequences and slope angle feature sequences with temporal structure and terrain semantic information. The similarity score is calculated through feature matching to establish a direct correlation between the vehicle's motion posture and the road terrain morphology. In the GNSS signal occlusion environment, the matching score can effectively reflect the consistency between the vehicle's actual driving trajectory and the map terrain model, thereby dynamically correcting the longitudinal positioning error caused by IMU integral drift. This achieves highly robust positioning without relying on external visual or lidar features, based solely on the vehicle's own sensors and map slope information. Thus, the problem of longitudinal positioning error accumulation caused by IMU drift and lack of effective shape constraints in the GNSS signal occlusion environment in the prior art is solved, achieving the effect of continuously and stably constraining positioning drift and improving the vehicle's longitudinal positioning accuracy and system reliability even in the absence of satellite signals. This solves the problem of poor vehicle positioning accuracy in existing solutions when GNSS signals are weakened or lost due to obstruction.

[0038] In the specific implementation process, feature matching processing is performed on the above-mentioned pitch angle feature sequence and the above-mentioned slope angle feature sequence to calculate the matching score of the above-mentioned pitch angle feature sequence and the above-mentioned slope angle feature sequence. This includes: constructing a candidate displacement set, and translating the above-mentioned slope angle feature sequence according to each of the above-mentioned candidate displacements in the above-mentioned candidate displacement set to obtain the translated above-mentioned slope angle feature sequence; performing feature matching processing on the above-mentioned pitch angle feature sequence and the translated above-mentioned slope angle feature sequence to obtain the matching score corresponding to each of the above-mentioned candidate displacements.

[0039] The candidate displacement set represents the possible longitudinal offset (in meters) of the vehicle's current actual position relative to its estimated position on the map; that is, candidate points in the longitudinal direction where the vehicle is "suspected" of deviating. The construction of the candidate displacement set must meet the following requirements:

[0040] 1) Cover the maximum expected drift: The set range must cover the maximum longitudinal cumulative error of inertial navigation in typical scenarios (such as the error range of driving 1km in a tunnel).

[0041] 2) Supports sub-meter resolution: Achieves continuous correction output at the sub-step level by using a sufficiently small step size;

[0042] 3) Controllable computational efficiency: The size of the dataset should be moderate to ensure real-time inference on the vehicle-mounted embedded system (<10ms).

[0043] 4) Consistent with spatial sampling rate: The discrete points of the set must be aligned with the spatial sampling interval of the input sequence (e.g., 0.5m / point) to avoid errors introduced by interpolation.

[0044] For example: to set a vertical search range R=10m and a step size d=0.5m, the formula can be used:

[0045] S = {Si = -R + (i-1)d, i = 1, 2, 3 ... N}, construct a candidate displacement set S, where Si is the i-th candidate displacement and N is the total number of candidate displacements.

[0046] In this embodiment, by constructing a candidate displacement set covering multiple discrete longitudinal displacements, and aligning the ground slope angle feature sequence extracted from the map according to each candidate displacement, dynamic multi-position matching of the ground slope angle feature sequence collected by the vehicle sensor in the longitudinal space is achieved, thereby generating a series of matching scores corresponding to different displacement assumptions. This process systematically searches and evaluates the positioning deviation that may be caused by IMU longitudinal drift, avoiding the risk of misjudgment caused by a single matching point, and making the final position correction based on the optimal matching candidate rather than a coarse estimate. This significantly improves the accuracy and robustness of vehicle longitudinal positioning in GNSS occlusion environment, achieves sub-meter level positioning correction effect, and effectively suppresses the error accumulation problem caused by IMU drift and lack of shape constraints.

[0047] Specifically, feature matching processing is performed on the above-mentioned pitch angle feature sequence and the above-mentioned translated slope angle feature sequence to obtain the above-mentioned matching score corresponding to each of the above-mentioned candidate displacements, including: calculating the similarity of the overlapping area between the above-mentioned pitch angle feature sequence and each of the above-mentioned translated slope angle feature sequences, and using each of the above-mentioned similarities as the above-mentioned matching score corresponding to each of the above-mentioned candidate displacements.

[0048] The similarity can be determined by calculating the negative mean square error of the overlapping area between the pitch angle feature sequence and each translated slope angle feature sequence; alternatively, the similarity between sequences can be determined by correlation, distance metric, and similarity score.

[0049] In this embodiment, by constructing a candidate displacement set and performing multi-scale translation processing on the ground slope angle feature sequence, a series of translated versions with potential temporal or spatial offsets from the vehicle pitch angle feature sequence are generated. Then, the similarity between each translated version and the pitch angle feature sequence in the overlapping area is calculated, which is used as a matching score to quantify the degree of local waveform alignment. This mechanism can effectively overcome the longitudinal position accumulation error caused by IMU drift, especially in complex scenarios where the road slope features are unevenly distributed or have local offsets. By discretely traversing potential displacements and using the similarity of the overlapping area as the matching criterion, accurate local alignment of the pitch angle and slope angle sequences is achieved, thereby significantly improving the accuracy and stability of the predicted position correction based on map slope information. Finally, it effectively suppresses and highly reliable corrects the longitudinal positioning error of the vehicle under GNSS occlusion.

[0050] Further, the current predicted position is corrected based on the matching score to obtain the current driving position of the vehicle, including: converting the matching score corresponding to each candidate displacement into a probability distribution, wherein the probability distribution represents the probability that the vehicle is located at each candidate displacement; performing a weighted average calculation on each candidate displacement based on the probability of each candidate displacement in the probability distribution to obtain a target displacement; and correcting the current predicted position based on the target displacement to obtain the current driving position of the vehicle.

[0051] In this embodiment, by constructing a candidate displacement set and performing multi-scale translation on the ground slope angle feature sequence, combined with the matching score of the vehicle pitch angle feature sequence, the originally discrete displacement matching results are transformed into a continuous probability distribution, thereby spatially expressing the matching confidence of the vehicle at each candidate displacement in probabilistic form. On this basis, a weighted average calculation is performed according to the weight of each candidate displacement in the probability distribution to output the target displacement at the sub-step level, realizing continuous and high-precision correction of the vehicle's current predicted position. This not only breaks through the positioning accuracy bottleneck caused by the step size limitation of traditional discrete candidate sets, but also makes the entire matching and correction process differentiable, providing a mathematical basis for subsequent end-to-end neural network training, effectively suppressing the error accumulation of IMU drift in GNSS occlusion environment, and significantly improving the stability and accuracy of vehicle longitudinal positioning.

[0052] Furthermore, after calculating the target displacement by performing a weighted average calculation on each of the candidate displacements according to the probabilities of each of the candidate displacements in the above probability distribution, the method further includes: determining the dispersion of the probability distribution according to the probabilities of each of the candidate displacements in the above probability distribution, and evaluating the confidence level of the target displacement based on the dispersion.

[0053] The dispersion of the probability distribution can be determined by calculating the variance of the probability distribution.

[0054] In this embodiment, by acquiring the vehicle's pitch angle sequence and ground slope angle sequence within the current time period, a matching score reflecting the similarity between the two is generated through feature extraction and matching, and then transformed into a probability distribution characterizing the vehicle's possible displacement probability. Further, by calculating the variance of this probability distribution as the dispersion of the distribution, only when the variance is below a preset threshold, indicating that the probability distribution of the candidate displacement has a significant central tendency and the positioning confidence is high, is a probability-based weighted average calculation performed to obtain the target displacement, and the vehicle's current predicted position is corrected accordingly. When the variance is below the preset threshold, the confidence of the target displacement is considered low, the road surface is flat, and the target displacement is not used to correct the vehicle's positioning. This mechanism effectively avoids the problem of introducing erroneous displacements due to blind correction when the terrain is flat and the features are not significant, leading to a uniform probability distribution. It ensures that position correction is only implemented when the positioning status is reliable, thereby significantly improving the accuracy and stability of the vehicle's longitudinal positioning in GNSS obstruction environments, mitigating the error growth caused by IMU drift accumulation, and enhancing the system's robustness in complex road environments.

[0055] Specifically, feature extraction processing is performed on the vehicle pitch angle sequence and the road slope angle sequence to obtain pitch angle feature sequence and slope angle feature sequence, including: using a feature extraction model to perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence to obtain pitch angle feature sequence and slope angle feature sequence, wherein the feature extraction model is a convolutional neural network model or a recurrent neural network model.

[0056] In this embodiment, the feature extraction network is a lightweight one-dimensional convolutional neural network. Its training process is end-to-end supervised learning, which aims to enable the network to learn robust feature representations that can discriminate changes in terrain slope, thereby improving the accuracy and generalization ability of sequence matching.

[0057] In this embodiment, by acquiring the vehicle's pitch angle sequence and ground slope angle sequence within the current time period, and combining them with the vehicle's current predicted position, a convolutional neural network model or a recurrent neural network model is used to extract features from these two sets of sequences. This transforms the original sensor data into a high-dimensional feature sequence with nonlinear pattern capture capability and noise interference resistance, thereby effectively alleviating the feature mismatch problem caused by sensor drift, local terrain disturbance, or sequence temporal deformation. Subsequently, a similarity score is calculated by matching the two sets of feature sequences, and the predicted position is dynamically corrected based on this score. This significantly improves the stability and accuracy of vehicle longitudinal positioning in GNSS signal obstruction environments, solving the technical defects of traditional methods that suffer from low matching success rate and severe accumulation of correction errors due to direct comparison of the original angle sequences. Ultimately, this achieves a highly robust and accurate vehicle positioning effect.

[0058] More specifically, after obtaining the vehicle pitch angle sequence, road slope angle sequence, and the current predicted position of the vehicle in the current time period, the method further includes: performing sequence normalization operations on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain normalized vehicle pitch angle sequence and road slope angle sequence, wherein the sequence normalization operation includes at least one of mean removal operation and scale normalization operation.

[0059] In this embodiment, after acquiring the vehicle pitch angle sequence and the ground slope angle sequence, mean centering and variance scaling are performed on both sequences for sequence normalization. This effectively eliminates the systematic offset of the pitch angle measurement system caused by sensor installation deviation, as well as the interference of absolute numerical errors in map slope data on feature matching. This ensures that the subsequently extracted pitch angle feature sequences and slope angle feature sequences retain only relative shape and trend information. Thus, the feature matching process focuses on the true consistency of the two sequences in terms of temporal shape, rather than the false similarity affected by absolute numerical deviations. This significantly improves the accuracy and robustness of the matching score. Furthermore, in GNSS signal obstruction environments, the IMU predicted position is reliably corrected based on the high-precision matching results, suppressing the accumulation of longitudinal positioning errors caused by IMU drift. Ultimately, this achieves stable and high-precision vehicle positioning in complex environments.

[0060] To enable those skilled in the art to better understand the technical solution of this application, the implementation process of the vehicle positioning method based on ground slope of this application will be described in detail below with reference to specific embodiments.

[0061] Existing solutions typically rely on a combination of GNSS (Global Navigation Satellite System) and IMU (Inertial Measurement Unit) navigation to determine vehicle position. In scenarios where GNSS signals are weakened or absent due to obstruction, such as tunnels, underground parking garages, or densely built-up "urban canyons," the positioning system primarily relies on IMU integration to estimate position. Because IMUs have zero bias and noise, the vehicle's longitudinal (direction of travel) position error accumulates over time, causing drift. Additionally, existing solutions also use lidar or visual matching to determine vehicle position. While lidar or visual matching can provide longitudinal positioning to some extent in areas with obstructed or absent signals, these solutions have the following drawbacks:

[0062] 1. Global point cloud matching or complex image feature extraction places high demands on the computing power of the vehicle-mounted computing platform, resulting in high consumption of computing resources;

[0063] 2. In scenarios with longitudinal degradation, such as tunnels (lacking longitudinal geometric features), it is difficult for vision and LiDAR to extract effective features for longitudinal constraints, causing the localization to easily drift along the road direction;

[0064] 3. The map features used by LiDAR or vision have poor universality. LiDAR or vision map features need to be specially exported and have different formats.

[0065] To address the aforementioned technical problems, this embodiment relates to a specific vehicle positioning method based on ground slope. It utilizes the unique sequential characteristics of road slope (pitch angle) changes traversed by the vehicle, matching them with slope data in a high-precision map to calculate the vehicle's longitudinal position deviation. For example... Figure 3 As shown, the specific processing flow is as follows:

[0066] 1) Data input and preprocessing:

[0067] Input A (real-time observation): Acquires the real-time pitch angle sequence of the vehicle over a period of time (the vehicle pitch angle sequence, which is usually derived from the vehicle sensor data fusion positioning output).

[0068] Among them, the real-time observed "Pitch angle sequence" refers to a part of the vehicle attitude obtained by the fusion of IMU / IMU wheel speed / IMU + other sensors, which is the pitch angle of the vehicle coordinate system relative to the local horizontal coordinate system (which can be obtained by quaternion or Euler angle decomposition, depending on the way the vehicle attitude is represented).

[0069] Input B (Map Reference): Extracts the corresponding road slope angle sequence from the high-precision map based on the current estimated location.

[0070] The diagram showing the comparison between the vehicle pitch angle sequence and the road slope angle sequence is as follows: Figure 4 As shown, the road slope angle sequence referenced in the map refers to the slope angle of a local road tangent relative to the horizontal plane (i.e., the ground slope angle). The road slope angle sequence is bound to the lane geometry data; for example, for a specific lane, there will be slope data every 1 meter or x meters. This slope data can be generated from the lane centerlines of the base map. High-precision maps typically use the UTM coordinate system, WGS84 coordinate system, or GCJ02 coordinate system.

[0071] The mechanism for extracting and generating road slope angle sequences includes: obtaining the vehicle's current estimated position (the prior position calculated from the previous moment), finding the lane closest to that position on the map, and finding the road slope angle sequence along the front and rear directions of the lane centerline with the current estimated position as the center.

[0072] 2) Sequence normalization: To eliminate sensor installation angle bias or systematic bias in map data, the input sequence is mean-centered (mean subtracted) and variance-scaled to ensure that the matching focuses only on waveform trends rather than absolute values, thus improving robustness.

[0073] 3) Feature Encoding:

[0074] Instead of directly using the original angle data for matching, a convolutional neural network is used to extract deep features from the vehicle pitch angle sequence and the road slope angle sequence to obtain pitch angle feature sequences and slope angle feature sequences. In addition to using a convolutional neural network to extract features, other sequence feature advance networks can also be used to extract deep features from the vehicle pitch angle sequence and the road slope angle sequence.

[0075] The encoder consists of two layers of one-dimensional convolution (Conv1d), combined with instance normalization (InstanceNorm1d) and the ReLU activation function. This allows for the extraction of more robust local slope variation features.

[0076] In addition, the convolutional neural network (CNN) part can be replaced with a recurrent neural network (RNN / LSTM) or a Transformer structure to extract sequence features.

[0077] 4) Shift Scoring:

[0078] Set a longitudinal search range (e.g., + / - 10 meters) and a step size (e.g., 0.5 meters).

[0079] A discrete set of displacement candidate values ​​is constructed, and the slope angle feature sequence is translated according to the set of displacement candidate values ​​to obtain multiple translated slope angle feature sequences.

[0080] For each candidate displacement k, the pitch angle feature sequence is aligned with the translated slope angle feature sequence, and the similarity of the overlapping area (Negative MSE) is calculated as the matching score. This means that the more similar the two curves are, the higher the score.

[0081] Alternatively, traditional cross-correlation analysis can be used instead of deep learning feature matching, and dynamic time warping (DTW) algorithms can be used to handle nonlinear stretching of sequences.

[0082] 5) Differentiable soft grid search and positioning correction:

[0083] After dividing the above scores by a preset coefficient, they are converted into a probability distribution (probs) using the Softmax function. This represents the probability that the vehicle is located at each candidate displacement.

[0084] 6) Position Correction Output: Instead of directly taking the point with the highest probability (Argmax), it calculates the mathematical expectation of the probability distribution, i.e., pred_offset = sum(probs). This approach makes the output continuously differentiable, enabling sub-step level continuous displacement correction accuracy.

[0085] Alternatively, hard classification (directly taking the Bin with the highest probability) can be used instead of the mathematical expectation of the probability distribution.

[0086] 7) Uncertainty estimation:

[0087] The model simultaneously calculates the variance (second moment) of the predicted probability distribution and outputs the standard deviation of the predicted displacement. The system determines the confidence level of the predicted displacement based on the standard deviation. If the standard deviation is less than a standard deviation threshold, it is determined that the current terrain has sufficient undulation features to support positioning, and the predicted displacement is reliable. If the standard deviation is greater than or equal to the standard deviation threshold, it is determined that the current terrain does not have sufficient undulation features to support positioning, and the predicted displacement is unreliable. This can determine whether the current terrain has sufficient undulation features to support positioning (if it is a flat road, the variance will be very large, and the system can choose not to accept this correction).

[0088] The variance of the predicted probability distribution is also called the second central moment, and its calculation formula is as follows:

[0089] Let xi be the i-th candidate displacement value, and pi be the corresponding probability output by the model.

[0090] Predicted mean (expected value): μ = ∑(pi) xi);

[0091] Second-order moment at the origin: E[X2]=∑(pi) xi2);

[0092] Variance: Var(X) = E[X²] - μ²;

[0093] Uncertainty standard deviation: ;

[0094] Physical meaning: This value reflects the model's "confidence" in the current localization result. If the terrain is flat and featureless, the probability distribution pi will tend to be uniformly distributed, leading to a larger variance; if the terrain features are significant, the probability distribution will be concentrated at a certain point, and the variance will be smaller.

[0095] This embodiment specifically achieves the following technical effects:

[0096] 1) High robustness: Through sequence normalization and feature encoding, this embodiment eliminates the influence of sensor installation errors and map absolute value deviations, making it more reliable than directly matching the original values.

[0097] 2) No expensive sensors required: It can correct longitudinal positioning in scenarios without GPS, such as tunnels, by relying only on IMU attitude and map data, without the need for LiDAR or cameras.

[0098] 3) Differentiability and high precision: The “Soft Grid Search” mechanism is adopted, which transforms discrete search steps into continuous correction outputs, improving positioning accuracy and supporting end-to-end model training.

[0099] 4) Possesses self-evaluation capabilities: It can output the uncertainty standard deviation (Uncertainty Std), allowing downstream fusion algorithms to dynamically adjust weights based on the richness of terrain features.

[0100] This application also provides a vehicle positioning device based on ground slope. It should be noted that the vehicle positioning device based on ground slope in this application can be used to execute the vehicle positioning method based on ground slope provided in this application. This device is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0101] The following describes the vehicle positioning device based on ground slope provided in the embodiments of this application.

[0102] Figure 5 This is a schematic diagram of a vehicle positioning device based on ground slope according to an embodiment of this application. Figure 5 As shown, the device includes:

[0103] The acquisition unit 51 is used to acquire the vehicle pitch angle sequence, road slope angle sequence and the current predicted position of the vehicle in the current time period. The vehicle pitch angle sequence is acquired by the vehicle sensor and is used to characterize the pitch angle sequence of the vehicle relative to the horizontal plane. The road slope angle sequence is road slope data extracted from the map based on the road the vehicle is traveling on.

[0104] The feature extraction processing unit 52 is used to perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain the pitch angle feature sequence and the slope angle feature sequence, and to perform feature matching processing on the pitch angle feature sequence and the slope angle feature sequence to calculate the matching score of the pitch angle feature sequence and the slope angle feature sequence.

[0105] The correction processing unit 53 is used to correct the current predicted position based on the matching score to obtain the current driving position of the vehicle.

[0106] In this embodiment, the acquisition unit is used to acquire the vehicle pitch angle sequence, road slope angle sequence, and the vehicle's current predicted position within the current time period. The vehicle pitch angle sequence is acquired by vehicle sensors and represents the vehicle's pitch angle sequence relative to the horizontal plane. The road slope angle sequence is a sequence of road slopes extracted from the map based on the vehicle's current predicted position on the map. The feature extraction processing unit is used to perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain pitch angle feature sequences and slope angle feature sequences. It then performs feature matching processing on the pitch angle feature sequences and slope angle feature sequences to calculate a matching score. The correction processing unit is used to correct the current predicted position based on the matching score to obtain the vehicle's current driving position. By acquiring the pitch angle sequence collected by the vehicle's onboard sensors within the current time period and the ground slope angle sequence extracted from a high-precision map based on the current predicted location, features are extracted from the pitch angle sequence and the ground slope angle sequence to form pitch angle feature sequences and slope angle feature sequences with temporal structure and terrain semantic information. Similarity scores are calculated through feature matching to establish a direct correlation between vehicle motion attitude and road terrain morphology. In GNSS signal obstruction environments, this matching score can effectively reflect the consistency between the vehicle's actual driving trajectory and the map terrain model, thereby dynamically correcting longitudinal positioning errors caused by IMU integral drift. This achieves highly robust positioning without relying on external visual or lidar features, based solely on the vehicle's own sensors and map slope information. Thus, it solves the problem of longitudinal positioning error accumulation caused by IMU drift and lack of effective shape constraints in GNSS signal obstruction environments in existing technologies, achieving the effect of continuously and stably constraining positioning drift and improving vehicle longitudinal positioning accuracy and system reliability even in the absence of satellite signals. This solves the problem of poor vehicle positioning accuracy in existing solutions when GNSS signals are weakened or lost due to obstruction.

[0107] As an optional approach, the feature extraction processing unit includes a construction module and a feature matching processing module. The construction module is used to construct a candidate displacement set and, based on each candidate displacement in the candidate displacement set, translates the slope angle feature sequence to obtain the translated slope angle feature sequence. The feature matching processing module is used to perform feature matching processing on the pitch angle feature sequence and the translated slope angle feature sequence to obtain the matching score corresponding to each candidate displacement.

[0108] In one optional scheme, the feature matching processing module includes a calculation submodule, which is used to calculate the similarity of the overlapping areas between the pitch angle feature sequence and each translated slope angle feature sequence, and to use each similarity as the matching score corresponding to each candidate displacement.

[0109] In one optional scheme, the correction processing unit includes a conversion module, a weighted average calculation processing module, and a correction processing module; the conversion module is used to convert the matching scores corresponding to each of the above-mentioned candidate displacements into a probability distribution, wherein the probability distribution represents the probability that the vehicle is located at each of the above-mentioned candidate displacements; the weighted average calculation processing module is used to perform weighted average calculation processing on each of the above-mentioned candidate displacements according to the probability of each of the above-mentioned candidate displacements in the above-mentioned probability distribution to obtain the target displacement; the correction processing module is used to correct the current predicted position according to the target displacement to obtain the current driving position of the vehicle.

[0110] In one optional scheme, the weighted average calculation processing module includes an evaluation module, which is used to calculate the variance of the probability distribution based on the probability of each of the candidate displacements in the probability distribution, and when the variance is less than a variance threshold, perform a weighted average calculation processing on each of the candidate displacements based on the probability of each of the candidate displacements in the probability distribution to obtain the target displacement.

[0111] In one optional scheme, the feature extraction processing unit includes a feature extraction processing module, which is used to perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively using a feature extraction model to obtain the pitch angle feature sequence and the slope angle feature sequence, wherein the feature extraction model is a convolutional neural network model or a recurrent neural network model.

[0112] In one optional embodiment, the device further includes a sequence normalization operation unit, which, after acquiring the vehicle pitch angle sequence, road slope angle sequence, and the current predicted position of the vehicle in the current time period, performs sequence normalization operations on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain normalized vehicle pitch angle sequence and road slope angle sequence, wherein the sequence normalization operation includes mean centering operation and variance scaling operation.

[0113] The aforementioned vehicle positioning device based on ground slope includes a processor and a memory. The acquisition unit, feature extraction processing unit, correction processing unit, etc., are all stored as program units in the memory, and the processor executes the program units stored in the memory to achieve the corresponding functions. All of the above modules are located in the same processor; alternatively, the above modules may be located in different processors in any combination.

[0114] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured, and adjusting kernel parameters can address the problem of poor vehicle positioning accuracy in scenarios where GNSS signals are weakened or lost due to obstruction.

[0115] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0116] This invention provides a computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device containing the computer-readable storage medium to perform the vehicle positioning method based on ground slope.

[0117] Specifically, vehicle positioning methods based on ground slope include:

[0118] Step S201: Obtain the vehicle pitch angle sequence, road slope angle sequence, and the current predicted position of the vehicle in the current time period. The vehicle pitch angle sequence is obtained by the vehicle sensor and is used to characterize the vehicle's pitch angle sequence relative to the horizontal plane. The road slope angle sequence is the sequence of road slopes extracted from the map based on the vehicle's current predicted position on the map.

[0119] Step S202: Perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain pitch angle feature sequence and slope angle feature sequence, and perform feature matching processing on the pitch angle feature sequence and the slope angle feature sequence to calculate the matching score of the pitch angle feature sequence and the slope angle feature sequence.

[0120] Step S203: Correct the current predicted position based on the matching score to obtain the current driving position of the vehicle.

[0121] This invention provides a processor for running a program, wherein the program executes the vehicle positioning method based on ground slope.

[0122] Specifically, vehicle positioning methods based on ground slope include:

[0123] Step S201: Obtain the vehicle pitch angle sequence, road slope angle sequence, and the current predicted position of the vehicle in the current time period. The vehicle pitch angle sequence is obtained by the vehicle sensor and is used to characterize the vehicle's pitch angle sequence relative to the horizontal plane. The road slope angle sequence is the sequence of road slopes extracted from the map based on the vehicle's current predicted position on the map.

[0124] Step S202: Perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain pitch angle feature sequence and slope angle feature sequence, and perform feature matching processing on the pitch angle feature sequence and the slope angle feature sequence to calculate the matching score of the pitch angle feature sequence and the slope angle feature sequence.

[0125] Step S203: Correct the current predicted position based on the matching score to obtain the current driving position of the vehicle.

[0126] This invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs at least the following steps:

[0127] Step S201: Obtain the vehicle pitch angle sequence, road slope angle sequence, and the current predicted position of the vehicle in the current time period. The vehicle pitch angle sequence is obtained by the vehicle sensor and is used to characterize the vehicle's pitch angle sequence relative to the horizontal plane. The road slope angle sequence is the sequence of road slopes extracted from the map based on the vehicle's current predicted position on the map.

[0128] Step S202: Perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain pitch angle feature sequence and slope angle feature sequence, and perform feature matching processing on the pitch angle feature sequence and the slope angle feature sequence to calculate the matching score of the pitch angle feature sequence and the slope angle feature sequence.

[0129] Step S203: Correct the current predicted position based on the matching score to obtain the current driving position of the vehicle.

[0130] The devices mentioned in this article can be servers, PCs, tablets, mobile phones, etc.

[0131] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program having at least the following method steps:

[0132] Step S201: Obtain the vehicle pitch angle sequence, road slope angle sequence, and the current predicted position of the vehicle in the current time period. The vehicle pitch angle sequence is obtained by the vehicle sensor and is used to characterize the vehicle's pitch angle sequence relative to the horizontal plane. The road slope angle sequence is the sequence of road slopes extracted from the map based on the vehicle's current predicted position on the map.

[0133] Step S202: Perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain pitch angle feature sequence and slope angle feature sequence, and perform feature matching processing on the pitch angle feature sequence and the slope angle feature sequence to calculate the matching score of the pitch angle feature sequence and the slope angle feature sequence.

[0134] Step S203: Correct the current predicted position based on the matching score to obtain the current driving position of the vehicle.

[0135] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0136] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0137] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0138] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0139] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0140] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0141] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0142] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0143] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0144] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0145] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A vehicle positioning method based on ground slope, characterized in that, include: The vehicle pitch angle sequence, road slope angle sequence, and the vehicle's current predicted position are obtained within the current time period. The vehicle pitch angle sequence is acquired by vehicle sensors and is used to characterize the vehicle's pitch angle sequence relative to the horizontal plane. The road slope angle sequence is road slope data extracted from the map based on the road the vehicle is traveling on. Feature extraction is performed on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain pitch angle feature sequences and slope angle feature sequences; feature matching is then performed on the pitch angle feature sequences and the slope angle feature sequences to calculate the matching score between them, including: constructing a candidate displacement set, and translating the slope angle feature sequence according to each candidate displacement in the candidate displacement set to obtain the translated slope angle feature sequence; feature matching is then performed on the pitch angle feature sequence and the translated slope angle feature sequence to obtain the matching score corresponding to each candidate displacement; The current predicted position is corrected based on the matching score to obtain the current driving position of the vehicle. The current predicted position is corrected based on the matching score to obtain the current driving position of the vehicle, including: converting the matching score corresponding to each candidate displacement into a probability distribution, wherein the probability distribution represents the probability that the vehicle is located at each candidate displacement; performing a weighted average calculation on each candidate displacement based on the probability of each candidate displacement in the probability distribution to obtain a target displacement; and correcting the current predicted position based on the target displacement to obtain the current driving position of the vehicle.

2. The method according to claim 1, characterized in that, The pitch angle feature sequence and the translated slope angle feature sequence are subjected to feature matching processing to obtain the matching score corresponding to each candidate displacement, including: Calculate the similarity of the overlapping regions between the pitch angle feature sequence and each translated slope angle feature sequence, and use each similarity as the matching score corresponding to each candidate displacement.

3. The method according to claim 1, characterized in that, After calculating the target displacement by performing a weighted average calculation on each candidate displacement based on the probability of each candidate displacement in the probability distribution, the method further includes: The dispersion of the probability distribution is determined based on the probability of each candidate displacement in the probability distribution, and the confidence level of the target displacement is evaluated based on the dispersion.

4. The method according to claim 1, characterized in that, Feature extraction processing is performed on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain pitch angle feature sequences and slope angle feature sequences, including: The vehicle pitch angle sequence and the road slope angle sequence are processed by a feature extraction model to obtain the pitch angle feature sequence and the slope angle feature sequence, wherein the feature extraction model is a convolutional neural network model or a recurrent neural network model.

5. The method according to claim 1, characterized in that, After obtaining the vehicle pitch angle sequence, road slope angle sequence, and the vehicle's current predicted position for the current time period, the method further includes: The vehicle pitch angle sequence and the road slope angle sequence are respectively subjected to sequence normalization operation to obtain normalized vehicle pitch angle sequence and road slope angle sequence, wherein the sequence normalization operation includes at least one of mean removal operation and scale normalization operation.

6. A vehicle positioning device based on ground slope, characterized in that, The vehicle positioning device is used to perform the vehicle positioning method based on ground slope as described in any one of claims 1 to 5, and the vehicle positioning device comprises: The acquisition unit is used to acquire the vehicle pitch angle sequence, road slope angle sequence, and the current predicted position of the vehicle in the current time period. The vehicle pitch angle sequence is acquired by the vehicle sensor and is used to characterize the vehicle's pitch angle sequence relative to the horizontal plane. The road slope angle sequence is road slope data extracted from the map based on the road the vehicle is traveling on. The feature extraction processing unit is used to perform feature extraction processing on the vehicle pitch angle sequence and the road slope angle sequence respectively to obtain pitch angle feature sequence and slope angle feature sequence, and to perform feature matching processing on the pitch angle feature sequence and the slope angle feature sequence to calculate the matching score of the pitch angle feature sequence and the slope angle feature sequence. The correction processing unit is used to correct the current predicted position based on the matching score to obtain the current driving position of the vehicle.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the computer-readable storage medium to perform the vehicle positioning method based on ground slope as described in any one of claims 1 to 5.

8. An electronic device, characterized in that, include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including methods for performing the vehicle positioning method based on ground slope as described in any one of claims 1 to 5.

Images