A method and system for tracking the trajectory of urban buses based on BeiDou positioning

Abstract

This application relates to the field of vehicle control technology, and provides a method and system for tracking the trajectory of urban buses based on BeiDou positioning. The method includes: acquiring vehicle driving motion information, vehicle BeiDou positioning information, and vehicle IMU positioning information; identifying multiple candidate road segments from preset map data based on the vehicle's BeiDou positioning information and IMU positioning information, and acquiring road geometric feature information for each candidate road segment; determining the vehicle's expected road segment from the multiple candidate road segments based on the vehicle's driving motion information and the road geometric feature information of the multiple candidate road segments, and determining the vehicle's drive control command based on the road geometric feature information of the expected road segment. This solution can improve the accuracy of vehicle control.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and in particular to a method and system for tracking the trajectory of urban buses based on BeiDou positioning. Background Technology

[0002] Modern city buses, especially those with distributed electric drive systems, generally rely on high-precision positioning and map information for intelligent control to achieve high efficiency, energy saving, and improved passenger comfort. However, in some complex urban road sections, such as under tunnels or viaducts, positioning signals may be briefly interrupted, causing the system to switch to inertial estimation mode. In this mode, errors accumulate over time, causing a discrepancy between the vehicle's reported position and its actual position after it leaves the signal-blocked area. This discrepancy can mislead the energy-saving control system's judgment of the road conditions ahead, leading to inappropriate drive commands. This not only affects energy efficiency but may also reduce the passenger experience and even require manual intervention from the driver.

[0003] In energy-saving control strategies based on high-precision positioning and mapping applied to distributed electric city buses, positioning information is briefly interrupted when the vehicle travels through areas where satellite signals are blocked or severely interfered with, such as tunnels, urban canyons, or multi-level overpasses. Although the system can switch to inertial calculation mode, inherent cumulative errors cause a deviation between the calculated position and the actual position after the vehicle leaves the signal-blocked area. This deviation misleads the energy-saving control system, causing it to match incorrect road condition information from the high-precision map, for example, misclassifying a steep uphill section as a gentle slope. Based on this erroneous judgment, the system outputs driving torque commands that do not match actual needs, such as incorrectly reducing the climbing torque, resulting in insufficient vehicle power and abnormal deceleration.

[0004] This situation forces the driver to intervene manually by significantly increasing the accelerator pedal opening to compensate for the power loss, thus triggering a sudden large current discharge. This not only contradicts the original intention of energy-saving control but also reduces ride comfort due to vehicle jerking and jerkiness. Therefore, the current technical challenge is to design a control method that can effectively identify and correct misjudgments of vehicle driving status caused by positional deviations when positioning information is temporarily lost and accumulates errors. This ensures that the energy-saving control system can output smooth and demand-matched drive commands even during the transition phase of positioning recovery, avoiding performance degradation and energy waste caused by erroneous control outputs. Summary of the Invention

[0005] This application provides a method and system for tracking the trajectory of urban buses based on BeiDou positioning. It aims to solve the problem of how to effectively identify and correct the misjudgment of vehicle driving status caused by position deviation when the positioning signal is briefly interrupted and accumulates errors in the intelligent control of urban buses. This ensures that the energy-saving control system can output smooth and matching drive commands even during the transition phase of positioning recovery, and avoids the problem of vehicle performance degradation and energy waste caused by incorrect control output.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] Firstly, a method for tracking the trajectory of urban buses based on BeiDou positioning is provided, comprising: acquiring vehicle driving motion information, vehicle BeiDou positioning location information, and vehicle IMU positioning information; identifying multiple candidate road segments from preset map data based on the vehicle's BeiDou positioning location information and vehicle IMU positioning information, and acquiring road geometric feature information of each candidate road segment; determining the vehicle's expected road segment among the multiple candidate road segments based on the vehicle's driving motion information and the road geometric feature information of the multiple candidate road segments, so as to determine the vehicle's drive control command based on the road geometric feature information of the vehicle's expected road segment.

[0008] Through this technical solution, this application can effectively integrate Beidou positioning and IMU positioning information, combine vehicle driving motion information and preset map data, intelligently identify the vehicle's expected road segment, and generate accurate drive control commands accordingly. This solves the problem of misjudging the vehicle's driving status when positioning information is temporarily lost and accumulates errors, ensuring that the energy-saving control system can output smooth and matching drive commands during the positioning recovery transition phase, thus avoiding vehicle performance degradation and energy waste.

[0009] Furthermore, based on the vehicle's BeiDou positioning information and the vehicle's IMU positioning information, multiple candidate road segments are identified from the preset map data, including: using the midpoint of the line connecting the vehicle's BeiDou positioning information and the vehicle's IMU positioning information as the target location information; and identifying multiple candidate road segments from the preset map data based on the vehicle's drive control accuracy and the target location information.

[0010] This technical solution uses the midpoint of the line connecting BeiDou positioning information and IMU positioning information as the target location, and combines the vehicle's drive control accuracy to identify candidate road segments. This allows for more accurate positioning of the vehicle's current location, and based on the vehicle's actual control capabilities, it filters out candidate road segments from the map that better match the actual driving conditions, thereby improving the accuracy of road matching.

[0011] Furthermore, based on the vehicle's drive control accuracy and target location information, multiple candidate road segments are identified from the preset map data, including: determining the target radius based on the vehicle's drive control accuracy; selecting multiple roads in the target area of ​​the preset map data whose driving direction is the same as the vehicle's driving direction as multiple candidate road segments; the center of the target area is the target location information, and the radius of the target area is the target radius.

[0012] This technical solution determines the target radius based on the vehicle's drive control precision, and then selects roads within the target area that are consistent with the vehicle's driving direction as candidate road segments. This effectively narrows the search range, improves the efficiency and accuracy of candidate road segment identification, and ensures that the selected road segments are highly matched with the vehicle's actual driving intention.

[0013] In some preferred embodiments, determining the target radius based on the vehicle's drive control precision includes: obtaining a first correspondence; the first correspondence includes a one-to-one correspondence between multiple drive control precision ranges and multiple radii; and using the radius corresponding to the drive control precision range in which the vehicle's drive control precision is located in the first correspondence as the target radius.

[0014] By establishing a correspondence between the driving control accuracy range and the radius, this application can dynamically adjust the target radius according to the actual driving control accuracy of the vehicle, making the identification of candidate road segments more flexible and accurate, thereby adapting to the control needs of different vehicles or different working conditions.

[0015] As an optional approach, the vehicle's motion information includes accelerometer and gyroscope data from the IMU, yaw rate data from the IMU, and steering wheel angle sensor data. Based on the vehicle's motion information and road geometry information from multiple candidate road segments, the vehicle's expected road segment among multiple candidate road segments is determined. This includes: determining the vehicle's gradient change information based on the IMU's accelerometer and gyroscope data; determining the vehicle's steering angle change information based on the IMU's yaw rate and steering wheel angle sensor data, and using the vehicle's gradient change information and steering angle change information as the vehicle's motion characteristic information; and determining the vehicle's expected road segment among multiple candidate road segments based on the vehicle's motion characteristic information and road geometry information from multiple candidate road segments.

[0016] This technical solution utilizes data from the IMU's accelerometer, gyroscope, yaw rate, and steering wheel angle sensor to accurately acquire information on vehicle gradient and steering angle changes, forming comprehensive driving motion characteristics. This allows for a more accurate matching of the vehicle's actual motion state with road geometry when determining the expected road segment, thus improving the accuracy and reliability of trajectory tracking.

[0017] To enhance functionality, road geometric feature information includes road slope variation information and road curve curvature variation information. Based on vehicle driving motion feature information and road geometric feature information of multiple candidate road segments, the expected road segment for the vehicle is determined from multiple candidate road segments. This includes: for each candidate road segment, determining the similarity between the road geometric feature information of the candidate road segment and the vehicle driving motion feature information; determining the positional distance between the vehicle's BeiDou positioning information and the vehicle's IMU positioning information; when the positional distance is greater than a preset positional distance threshold, selecting the candidate road segment with a similarity greater than a preset similarity threshold as the vehicle's expected road segment; when the positional distance is less than or equal to a preset positional distance threshold, selecting the candidate road segment with the highest similarity as the vehicle's expected road segment.

[0018] This technical solution introduces the location distance between BeiDou positioning and IMU positioning as a criterion, and combines it with a similarity threshold. It can flexibly select the determination strategy of the expected road segment according to the size of the positioning error. When the positioning accuracy is high, the road with the highest similarity is selected first, and when the positioning accuracy is low, the road with similarity that meets the threshold is selected. This effectively improves the robustness and accuracy of expected road segment identification under different positioning conditions.

[0019] As a further improvement, the vehicle's motion characteristic information includes the vehicle's slope change information and steering angle change information, and the road geometric characteristic information includes the road's slope change information and the road's curvature change information. Based on the vehicle's motion characteristic information and the road geometric characteristic information of multiple candidate road segments, the expected road segment of the vehicle is determined from among multiple candidate road segments. This includes: for each candidate road segment, determining the similarity between the road geometric characteristic information of the candidate road segment and the vehicle's motion characteristic information; determining whether the difference between the largest and second largest similarity among multiple similarities is greater than a preset similarity difference threshold; when the difference between the largest and second largest similarity among multiple similarities is greater than the preset similarity difference threshold, the candidate road segment corresponding to the largest similarity is taken as the vehicle's expected road segment; when the difference between the largest and second largest similarity among multiple similarities is less than or equal to the preset similarity difference threshold, the candidate road segment with a similarity greater than the preset similarity threshold is taken as the vehicle's expected road segment.

[0020] This technical solution effectively determines the uniqueness and confidence level of road matching by comparing the difference between the maximum and second-highest similarity scores. When the difference is large, it indicates the existence of a significantly better matching road, which is then directly selected as the expected road segment. When the difference is small, it indicates the existence of multiple roads with high similarity scores. In this case, a similarity threshold is used for filtering, avoiding misjudgments caused by close similarity scores and improving the accuracy and reliability of expected road segment selection.

[0021] To optimize the structure, the vehicle's drive control command is determined based on the road geometry information of the expected road segment. This includes: determining whether the number of expected road segments is one; if the number of expected road segments is one, calling a preset drive control model; the preset drive control model is used to determine the vehicle's drive control command based on the road geometry information of the expected road segment and the vehicle's driving motion information; and inputting the road geometry information of the expected road segment and the vehicle's driving motion information into the preset drive control model to obtain the vehicle's drive control command.

[0022] By using this technical solution, this application determines the number of expected road segments. When the expected road segment is unique, it directly uses a preset drive control model, combined with the geometric features of the road and the vehicle's driving motion information, to generate precise drive control commands, thereby simplifying the control logic and improving the efficiency and accuracy of control command generation.

[0023] Based on this, the vehicle's drive control commands are determined according to the road geometric feature information of the expected road segments. This also includes: when there are multiple expected road segments, determining the road distance between the expected road segment and the target location information for each expected road segment; the target location information is the midpoint of the line connecting the vehicle's BeiDou positioning information and the vehicle's IMU positioning information; normalizing the multiple road distances and using the normalization result as the weight value corresponding to each expected road segment; based on the weight values ​​corresponding to the multiple expected road segments, the weighted sum of the road geometric feature information of the multiple expected road segments is used as the target road geometric feature information; calling a preset drive control model; and inputting the target road geometric feature information and the vehicle's driving motion information into the preset drive control model to obtain the vehicle's drive control commands.

[0024] This technical solution calculates the road distance between each expected road segment and the target location when multiple expected road segments exist, performs normalization processing to obtain a weight value, and then weights and fuses the geometric feature information of multiple roads to form a comprehensive target road geometric feature information. This processing method can fully consider all possible expected road segments and assign different weights according to their proximity to the target location, making the final generated drive control command smoother and more robust, effectively avoiding control uncertainty in multi-road selection scenarios.

[0025] Secondly, this application also discloses a BeiDou-based urban bus trajectory tracking system, comprising: an acquisition device and a processing device; the acquisition device is used to acquire vehicle driving motion information, vehicle BeiDou positioning location information, and vehicle IMU positioning information; the processing device is used to identify multiple candidate road segments from preset map data based on the vehicle's BeiDou positioning location information and vehicle IMU positioning information, and acquire road geometric feature information of each candidate road segment; the processing device is used to determine the vehicle's expected road segment among the multiple candidate road segments based on the vehicle's driving motion information and the road geometric feature information of the multiple candidate road segments, so as to determine the vehicle's drive control command based on the road geometric feature information of the vehicle's expected road segment.

[0026] Beneficial effects

[0027] This application discloses a method for tracking the trajectory of urban buses based on BeiDou positioning. By fusing vehicle motion information, BeiDou positioning information, and IMU positioning information, it effectively solves the problem of accumulated errors in traditional inertial estimation methods causing vehicle position deviations and thus misleading the energy-saving control system when positioning signals are interrupted or interfered with in complex urban road sections. Specifically, this application first acquires multi-source information, and then, based on BeiDou positioning and IMU positioning information, intelligently identifies multiple candidate road segments and their road geometric features from preset map data. On this basis, combining the vehicle's motion information and the geometric features of the candidate road segments, the expected road segment for the vehicle is accurately determined. Finally, based on the road geometric feature information of the expected road segment, drive control commands that highly match the actual road conditions are generated.

[0028] Through the above technical solution, this application overcomes the shortcomings of existing technologies where positioning errors lead to misjudgments of road conditions and inappropriate output of drive commands by the energy-saving control system. In situations where positioning information temporarily fails and accumulates errors, this application can effectively identify and correct misjudgments of vehicle driving status caused by position deviations. This ensures that even during the transition phase of positioning recovery, the energy-saving control system can output smooth drive commands that match actual needs, avoiding problems such as insufficient vehicle power, abnormal deceleration, driver intervention, instantaneous high-current discharge, and decreased ride comfort caused by erroneous control outputs. Therefore, the technical solution of this application significantly improves the intelligent control accuracy and robustness of urban buses in complex environments, achieving a dual optimization of energy saving and ride comfort. Attached Figure Description

[0029] Figure 1 A flowchart illustrating a method for tracking the trajectory of urban buses based on BeiDou positioning provided in this application;

[0030] Figure 2 A flowchart illustrating a method for tracking the trajectory of urban buses based on BeiDou positioning provided in this application;

[0031] Figure 3 This application provides a schematic diagram of the architecture of an urban bus trajectory tracking system based on BeiDou positioning. Detailed Implementation

[0032] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0033] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0034] Traditional city buses may experience brief interruptions in positioning signals when traveling on complex urban roads, such as tunnels or under overpasses, causing the system to switch to inertial estimation mode. In this mode, errors accumulate over time, resulting in a discrepancy between the vehicle's reported position and its actual position after it leaves the signal-blocked area. This discrepancy can mislead the energy-saving control system's judgment of the road conditions ahead, leading to inappropriate driving commands. This not only affects fuel efficiency but may also reduce passenger comfort and even require driver intervention. Failure to address these issues will result in degraded vehicle performance and energy waste.

[0035] In this regard, such as Figure 1 As shown, this application proposes a method for tracking the trajectory of urban buses based on BeiDou positioning, including:

[0036] S101. Obtain vehicle driving motion information, vehicle Beidou positioning information, and vehicle IMU positioning information.

[0037] S102. Based on the vehicle's BeiDou positioning information and the vehicle's IMU positioning information, identify multiple candidate road segments from the preset map data and obtain the road geometric feature information of each candidate road segment.

[0038] S103. Based on the vehicle's driving motion information and the road geometry feature information of multiple candidate road segments, determine the vehicle's expected road segment among the multiple candidate road segments, so as to determine the vehicle's drive control command based on the road geometry feature information of the vehicle's expected road segment.

[0039] This application, by comprehensively utilizing BeiDou positioning information, IMU positioning information, and vehicle driving motion information, combined with preset map data, can effectively identify the vehicle's expected road segment and thus determine accurate drive control commands. Therefore, even in complex urban road sections with limited positioning signals, it can ensure the accuracy of vehicle trajectory tracking and the smoothness of drive control, thereby improving energy efficiency and ride comfort, and reducing manual intervention.

[0040] To better understand the technical solution proposed in this application, it is necessary to explain some key terms and implementation environments involved. The "vehicle" referred to in this application specifically refers to a city bus, which is typically equipped with a distributed electric drive system and relies on high-precision positioning and map information for intelligent control. "BeiDou positioning information" refers to the vehicle's geographical location data obtained through the BeiDou satellite navigation system, typically including longitude, latitude, and altitude. "IMU positioning information" refers to the vehicle's position information obtained by the inertial measurement unit (IMU) through measurement of the vehicle's angular velocity and acceleration, and subsequent integration calculations. The IMU typically includes accelerometers and gyroscopes, providing information on the vehicle's attitude and motion state in three-dimensional space.

[0041] "Preset map data" refers to a pre-stored digital map containing detailed geometric features and topological relationships of urban roads, such as a high-precision map, which may include road slope, curve curvature, lane line information, etc. "Candidate road segments" refer to several road segments that may match the vehicle's actual driving path, selected from the preset map data based on the vehicle's current location information. "Expected road segments" refer to the road segments considered to best match the vehicle's current driving state and future driving intentions after further judgment and selection among multiple candidate road segments. "Drive control commands" refer to commands generated by the control system based on the vehicle's expected road segment information, used to adjust the vehicle's power output, steering, and other actuators, such as throttle opening, braking intensity, and steering angle. The implementation environment of this application is typically an intelligent driving or assisted driving system for urban buses, which can acquire and process the above-mentioned information in real time and precisely control the vehicle based on the processing results.

[0042] The core of the urban bus trajectory tracking method based on BeiDou positioning proposed in this application lies in achieving accurate identification and drive control of vehicle driving trajectory through multi-source information fusion.

[0043] First, it is necessary to acquire the vehicle's motion information, its BeiDou positioning information, and its IMU positioning information. The vehicle's motion information can be acquired in various ways. For example, various sensors installed on the vehicle, such as wheel speed sensors, steering angle sensors, accelerometers, and gyroscopes, can be used to collect data on the vehicle's speed, acceleration, angular velocity, and steering angle in real time. This data reflects the vehicle's current motion state and trend. The vehicle's BeiDou positioning information can be obtained through the onboard BeiDou positioning module, which receives BeiDou satellite signals and calculates the vehicle's real-time geographic coordinates. The vehicle's IMU positioning information is provided by the onboard inertial measurement unit (IMU). The IMU measures the vehicle's angular velocity and acceleration, and through integration calculations, obtains the vehicle's relative position and attitude changes.

[0044] Secondly, based on the vehicle's BeiDou positioning information and IMU positioning information, multiple candidate road segments are identified from the preset map data, and the road geometric feature information of each candidate road segment is obtained. In practical applications, BeiDou positioning information may drift or be interrupted due to obstruction from urban buildings, tunnels, and other environmental factors, while IMU positioning information, although continuous, accumulates errors over time. Therefore, it is necessary to fuse these two types of positioning information to obtain a more reliable vehicle position estimate. For example, algorithms such as Kalman filtering or particle filtering can be used to combine the absolute accuracy of BeiDou positioning with the relative continuity of IMU to obtain a fused target position information. Based on this target position information, road matching can be performed from the preset map data. For example, a search radius can be set with the target position information as the center, and all road segments in the preset map data that meet certain conditions can be searched within this radius as candidate road segments. These conditions may include road type (such as urban arterial roads, secondary arterial roads), the matching degree between road direction and vehicle driving direction, etc. Once multiple candidate road segments are identified, it is necessary to extract the road geometry features of each candidate road segment from the preset map data, such as road slope changes, curve curvature changes, lane width, etc.

[0045] Finally, based on the vehicle's motion information and the road geometry features of multiple candidate road segments, the vehicle's expected road segment is determined from among the candidate road segments. This allows for the determination of the vehicle's drive control commands based on the road geometry features of the expected road segment. After acquiring the vehicle's motion information and the geometric features of the candidate road segments, matching and judgment are required to determine the expected road segment where the vehicle is most likely to travel. For example, the vehicle's current motion information (such as slope, steering angle, etc.) can be compared with the geometric features of each candidate road segment to calculate their similarity. The road segment with the highest similarity, or one with a similarity threshold, can be selected as the expected road segment. In some cases, multiple road segments may have high similarity, requiring further judgment mechanisms. Once the expected road segment is determined, its road geometry features, combined with the vehicle's current motion information, can be used to generate corresponding drive control commands through a pre-defined drive control model or algorithm. For example, if the expected road segment indicates an uphill section ahead, the system may increase the drive torque; if it indicates a curve ahead, the system may adjust the steering angle and vehicle speed.

[0046] The method proposed in this application acquires vehicle motion information, BeiDou positioning information, and IMU positioning information. Based on this information, it identifies candidate road segments and their road geometric features from preset map data. Then, based on the vehicle's motion information and the road geometric features of the candidate road segments, it determines the vehicle's intended road segment. Finally, it determines the vehicle's drive control commands based on the road geometric features of the intended road segment. This series of steps forms a complete trajectory tracking and control closed loop.

[0047] The core innovation of this application lies in providing a method that can effectively identify the vehicle's trajectory and generate precise drive control commands when urban buses face uncertainties in positioning signals. Traditional methods often rely on pure inertial calculations when positioning signals are interrupted or drift, leading to accumulated errors, misjudging road conditions ahead, and outputting inappropriate drive commands, thus affecting energy efficiency and passenger comfort.

[0048] Compared to existing technologies, this application has the following advantages: First, by fusing BeiDou positioning information and IMU positioning information, this application can obtain a more stable and reliable vehicle position estimate than a single positioning source, effectively overcoming the problem of unstable positioning signals in complex urban environments. Second, this application introduces a matching mechanism between vehicle motion information and the geometric feature information of candidate road segments. This allows the system to accurately identify the vehicle's expected driving path from multiple possible road segments based on the vehicle's actual motion state, even when there are deviations in the positioning information. For example, when the vehicle is driving in a tunnel, the BeiDou signal is interrupted, and the IMU's accumulated error causes positioning drift, this application can still match the vehicle's motion information, such as slope and steering angle, with the geometric features of each candidate road segment in the preset map, such as slope and curvature, to accurately determine the actual road segment where the vehicle is located. Finally, based on the accurately identified expected road segment, this application can generate drive control commands that highly match the actual road conditions, avoiding energy waste and a decline in passenger experience caused by misjudgment, thereby significantly improving the energy-saving control effect and passenger comfort of urban buses. This strategy of multi-source information fusion and intelligent matching enables the application to maintain smoothness and accuracy of control during the transition phase of positioning recovery, effectively solving the problem of misjudgment of vehicle driving status caused by position deviation in the prior art.

[0049] This application further proposes an optimization scheme for more accurately identifying multiple candidate road segments. This scheme fuses two types of positioning information to more accurately determine the target position of the vehicle and combines the vehicle's drive control accuracy for identification.

[0050] Specifically, such as Figure 2 As shown, the steps for identifying multiple candidate road segments from preset map data based on the vehicle's BeiDou positioning information and IMU positioning information include:

[0051] S201. The midpoint of the line connecting the vehicle's BeiDou positioning information and the vehicle's IMU positioning information is taken as the target location information.

[0052] S202. Identify multiple candidate road segments from preset map data based on the vehicle's drive control accuracy and target location information.

[0053] The vehicle's BeiDou positioning information refers to the vehicle's geographic coordinate data acquired through the BeiDou satellite navigation system, while the vehicle's IMU positioning information refers to the positioning data obtained by integrating the vehicle's relative position or attitude information acquired through the inertial measurement unit (IMU). The midpoint of the line connecting these two types of positioning information is used as the target position information. The aim is to obtain a more representative and robust estimate of the vehicle's current position by fusing positioning data from two different sources. This fusion method effectively reduces the impact of single sensor errors on the positioning results, providing a smoother and more reliable vehicle position reference point.

[0054] Furthermore, after determining the target location information, multiple candidate road segments are identified from preset map data based on the vehicle's drive control accuracy and the target location information. The vehicle's drive control accuracy can be understood as the degree of deviation between the vehicle's actual driving trajectory and the desired trajectory during operation, reflecting the vehicle's handling performance and the accuracy requirements of the positioning system. By combining drive control accuracy, the range or strategy for identifying candidate road segments can be dynamically adjusted, making the identified road segments more consistent with the vehicle's actual driving capabilities and current operating conditions.

[0055] This application's solution effectively integrates two different types of positioning data by using the midpoint of the line connecting the vehicle's BeiDou positioning information and its IMU positioning information as the target position information. The BeiDou positioning system provides absolute position information, but it may be affected by environmental factors such as signal blockage and multipath effects; while the IMU positioning system provides relative position information, boasting a high update rate and short-term accuracy, but suffers from cumulative errors. By taking the midpoint, the long-term stability of BeiDou positioning can be used to correct the cumulative errors of the IMU, while the short-term accuracy of the IMU can compensate for potential instantaneous jumps or accuracy fluctuations in BeiDou positioning, thereby obtaining a more stable, accurate, and representative current target position of the vehicle.

[0056] Therefore, when identifying candidate road segments from the preset map data, a more reliable center point can be used for searching, improving the effectiveness and accuracy of the search. Furthermore, combining the vehicle's drive control precision with the identification process allows the identification of candidate road segments to take into account the vehicle's own motion characteristics and control capabilities, avoiding the identification of road segments that are difficult for the vehicle to actually drive on, further enhancing the practicality of trajectory tracking.

[0057] In this application, "the direction of travel is the same as the direction of travel of the vehicle" does not specifically refer to the direction of the vehicle's straight-line travel, but rather to the vehicle's current instantaneous direction of travel (i.e., the vehicle's heading angle or front direction) matching the local tangent direction of the road segment in the preset map data. In other words, when the vehicle turns, its direction of travel will continuously change.

[0058] Meanwhile, the road geometric feature information stored in the preset map data, such as the curvature change information of road curves, includes the local tangent direction of the road at different points. Therefore, even when the vehicle is turning, the system will still search for curved road segments within the target area whose local tangent direction is similar to the vehicle's current driving direction as candidate road segments, based on the vehicle's current driving direction. Furthermore, in subsequent steps, this application will also utilize the vehicle's driving motion feature information, including the vehicle's steering angle change information, to perform similarity matching with the road geometric feature information of the candidate road segments, further accurately determining the vehicle's expected road segment, thereby effectively handling trajectory tracking when the vehicle is turning.

[0059] Furthermore, even on straight sections of road where vehicles do not need to turn, there may be "multiple roads" with characteristics similar to the direction of vehicle travel; these "multiple roads" refer to "multiple candidate road segments." This mainly includes the following situations:

[0060] 1. Multi-lane roads: In urban roads, a physically straight road may contain multiple parallel lanes. In high-precision map data, these lanes are typically modeled as independent road segments. These parallel road segments all share characteristics with the same direction of vehicle travel within a local area.

[0061] 2. Parallel Roads: In some urban areas, there may be two or more very close parallel roads, such as a main road and a side road, or an adjacent elevated bridge and a ground-level road. If these road segments all fall within the "target area" and their travel direction is consistent with the vehicle's current travel direction, they may be identified as candidate road segments.

[0062] 3. Complex intersections or forks: At some complex intersections or road forks, there may be multiple paths that initially have similar directions of travel, but then gradually diverge. Before a vehicle has fully entered a particular path, these paths may all be considered as candidate road segments with the same direction of travel.

[0063] Therefore, "multiple roads" refers to multiple potential road segments available for vehicle selection within the current search area, obtained based on preliminary screening conditions (driving direction matching). This application will subsequently use more refined matching and judgment mechanisms (such as similarity calculation and location distance judgment) to determine the "expected road segment" that best matches the vehicle's actual situation from these multiple candidate road segments. Through the above technical solution, this application can effectively solve the positioning inaccuracies or fluctuations that may exist when using BeiDou positioning information and IMU positioning information alone or independently. By fusing the two positioning information and taking the midpoint of their connecting line as the target location information, a more stable, accurate, and representative estimate of the vehicle's current position can be obtained, thereby significantly improving the accuracy and robustness of identifying candidate road segments from preset map data. Furthermore, incorporating the vehicle's drive control precision into the candidate road segment identification process makes the identification results more consistent with the vehicle's actual driving capabilities and control characteristics, avoiding unrealistic road selections, further enhancing the practicality and safety of the trajectory tracking method, and providing a more reliable foundation for subsequent drive control command determination.

[0064] This application further proposes a step for identifying multiple candidate road segments from preset map data based on the vehicle's drive control accuracy and target location information, including:

[0065] The target radius is determined based on the vehicle's drive control precision; multiple roads in the target area within the preset map data with the same driving direction as the vehicle's driving direction are selected as multiple candidate road segments; the center of the target area is the target location information, and the radius of the target area is the target radius.

[0066] Specifically, the target radius refers to the radius of the circular area used to define the search range. This target radius is determined based on the vehicle's drive control precision, meaning different levels of drive control precision will correspond to different target radii. For example, when the vehicle's drive control precision is high, a smaller target radius may be needed to focus on a more precise local area; conversely, when the drive control precision is low, a larger target radius can be used to cover a wider potential road area. The target area is a circular region centered on the target location information and with the target radius as its radius. This target area defines a specific search range within the preset map data.

[0067] In practical applications, when identifying multiple candidate road segments from preset map data, all roads within the target area are first filtered out. Furthermore, to ensure that the identified candidate road segments better match the vehicle's actual driving intention or current state, only roads within the target area whose direction of travel is the same as the vehicle's direction of travel are included as the final multiple candidate road segments. This aims to exclude roads that do not match the vehicle's current direction of travel, such as oncoming lanes or irrelevant roads at intersections, thereby improving the quality of candidate road segment selection.

[0068] This application's solution effectively addresses the aforementioned issues of insufficient recognition efficiency and accuracy by introducing the concepts of target radius and target area, combined with matching the driving direction. Specifically, firstly, the target radius is dynamically determined based on the vehicle's drive control precision, allowing the search range to be adaptively adjusted. This avoids the computational burden of an excessively large search range and prevents the oversight of potentially correct roads due to an excessively small search range. Secondly, a target area is constructed with the target location information as the center and the target radius as the radius, providing a clear and focused spatial boundary for the identification of candidate road segments. Finally, by using roads within the target area whose driving direction is the same as the vehicle's driving direction as candidate road segments, the selection criteria are further refined, ensuring a high correlation between the identified candidate road segments and the vehicle's actual driving trajectory, thereby significantly improving the accuracy and efficiency of candidate road segment identification.

[0069] Through the above technical solution, this application can adaptively determine a reasonable search range based on the vehicle's drive control precision, and through a combination of target area and driving direction matching, make the candidate road segments identified from the preset map data more accurate and relevant. This not only effectively reduces unnecessary computation and improves recognition efficiency, but also significantly enhances the accuracy and robustness of subsequent vehicle trajectory tracking, providing a more reliable road matching foundation for intelligent driving of urban buses.

[0070] Specifically, the steps described above for determining the target radius based on the vehicle's drive control precision can be further refined.

[0071] According to the above method, the steps for determining the target radius based on the vehicle's drive control accuracy include:

[0072] Obtain the first correspondence relationship; the first correspondence relationship includes a one-to-one correspondence between multiple drive control precision ranges and multiple radii; take the radius corresponding to the drive control precision range in which the vehicle's drive control precision is located in the first correspondence relationship as the target radius.

[0073] The first correspondence can be understood as a pre-established mapping table or functional relationship, used to associate the driving control accuracy exhibited by the vehicle under different operating conditions with a suitable search radius. This correspondence can be trained and optimized based on a large amount of actual driving data, vehicle performance test data, and the accuracy requirements of the map matching algorithm. For example, when the vehicle's driving control accuracy is high, the deviation between its actual position and the expected position on the road is small. In this case, a smaller target radius can be set to narrow the search range and improve matching efficiency. Conversely, when the driving control accuracy is low, a larger target radius needs to be set to ensure that the vehicle's possible actual position is included within the target area.

[0074] Specifically, the drive control accuracy range refers to dividing the vehicle's drive control accuracy into several intervals, each corresponding to a specific radius value. For example, drive control accuracy can be divided into multiple levels such as "high accuracy," "medium accuracy," and "low accuracy," with a preset target radius for each level. The vehicle's drive control accuracy can be obtained in various ways, such as by real-time estimation using vehicle sensor data (e.g., GPS, inertial measurement unit, wheel speed sensors) combined with the vehicle's dynamics model, or by judgment based on factors such as vehicle type, current speed, and road type.

[0075] The solution proposed in this application achieves dynamic adaptive determination of the target radius by introducing a first correspondence relationship. When the system obtains the current drive control accuracy of the vehicle, this accuracy value is used to find the drive control accuracy range to which it belongs in the preset first correspondence relationship, and thus obtains the target radius that uniquely corresponds to that range. This mechanism of determining the target radius based on the actual drive control accuracy ensures that the size of the determined target area matches the actual positioning and control capabilities of the vehicle. Therefore, it avoids the problem of the search range being too large or too small that may occur when using a fixed radius, thereby improving the accuracy and efficiency of subsequent candidate road segment identification.

[0076] The above technical solution enables dynamic adjustment of the target radius based on the actual driving control precision of the vehicle, making the target area setting more reasonable and accurate. This not only helps improve the accuracy of identifying multiple candidate road segments from preset map data and reduces the possibility of misidentification, but also optimizes the use of computing resources, avoids unnecessary broad searches, and thus improves the robustness and real-time performance of the entire trajectory tracking method.

[0077] This application further proposes the above-mentioned method, wherein the vehicle's driving motion information includes IMU accelerometer and gyroscope data, IMU yaw rate and steering wheel angle sensor data, and based on the vehicle's driving motion information and road geometric feature information of multiple candidate road segments, the expected road segment of the vehicle among multiple candidate road segments is determined, including:

[0078] The vehicle's gradient change information is determined based on the accelerometer and gyroscope data from the IMU; the vehicle's steering angle change information is determined based on the yaw rate and steering wheel angle sensor data from the IMU, and the vehicle's gradient change information and steering angle change information are used as the vehicle's driving motion characteristic information; the vehicle's expected road segment among multiple candidate road segments is determined based on the vehicle's driving motion characteristic information and the road geometric characteristic information of multiple candidate road segments.

[0079] Specifically, vehicle motion information refers to a set of data reflecting the vehicle's current motion state and trend. In this embodiment, this information is specifically defined as accelerometer data, gyroscope data, yaw rate data from the IMU (Inertial Measurement Unit), and steering wheel angle sensor data. The IMU's accelerometer data measures the vehicle's linear acceleration in various directions, while the gyroscope data measures the vehicle's angular velocity along various axes. The combination of these two data points can be used to calculate changes in the vehicle's attitude, such as changes in gradient. The IMU's yaw rate directly reflects the vehicle's rotational speed around its vertical axis and is closely related to the vehicle's steering behavior. The steering wheel angle sensor data directly provides the driver's steering intention and is a direct input to the vehicle's steering behavior.

[0080] Furthermore, based on the accelerometer and gyroscope data from the IMU, the vehicle's gradient change information can be calculated. For example, by analyzing the accelerometer's component in the direction of gravity and combining it with the angular velocity integral from the gyroscope, the vehicle's pitch and roll angles during travel can be accurately estimated, thus obtaining the vehicle's gradient change. The purpose is to capture the vehicle's vertical motion state on inclines and declines.

[0081] Furthermore, based on the yaw rate from the IMU and data from the steering wheel angle sensor, information about changes in the vehicle's steering angle can be determined. For example, the yaw rate directly reflects the vehicle's actual steering rate, while the steering wheel angle sensor data provides the driver's steering input. Combining the two provides a more comprehensive description of the vehicle's steering dynamics, such as the actual turning radius and steering tendency. The aim is to accurately reflect the vehicle's horizontal steering state when cornering.

[0082] Therefore, the information on changes in vehicle gradient and steering angle is integrated to form the vehicle's motion characteristic information. This motion characteristic information is a comprehensive reflection of the vehicle's dynamic behavior, and can more comprehensively and accurately describe the vehicle's motion state under complex road conditions.

[0083] Ultimately, based on the vehicle's motion characteristics and the road geometry of multiple candidate road segments, the vehicle's intended road segment can be determined more accurately from among the multiple candidate road segments.

[0084] This application's solution incorporates accelerometer and gyroscope data from the IMU, along with yaw rate and steering wheel angle sensor data, and processes them into vehicle gradient change information and steering angle change information, thereby constructing more representative vehicle motion characteristic information. Traditionally, relying solely on simple speed or heading information may not accurately capture the subtle dynamics of a vehicle on slopes or curves. For example, in the operation of a city bus, gradient change information reflects the longitudinal movement trend of the vehicle when climbing or descending a slope, while steering angle change information accurately describes the lateral movement state of the vehicle when turning or changing lanes.

[0085] Because these detailed and dynamic driving motion characteristics are taken into account, a more refined and accurate comparison can be made when matching them with the road geometry features of candidate road segments in the preset map data. For example, when a vehicle is on an uphill curve, its gradient and steering angle changes will closely match the gradient and curvature information of the corresponding road segment in the map, thus significantly improving the recognition accuracy of the expected road segment and effectively avoiding misjudgments at multi-intersection or complex road sections.

[0086] Through the above technical solution, this application can obtain more comprehensive and accurate information on the vehicle's driving motion characteristics. Compared to using only basic driving motion information, this solution determines the vehicle's slope change information by combining IMU accelerometer and gyroscope data, and determines the vehicle's steering angle change information by combining IMU yaw rate and steering wheel angle sensor data, thus fully reflecting the vehicle's dynamic attitude and steering intention. Therefore, when determining the vehicle's expected road segment, it can more accurately match the vehicle's actual motion state with the geometric features of the candidate road segment, significantly improving the accuracy of expected road segment identification. Especially when urban buses face complex road conditions (such as slopes, sharp bends, and multiple intersections), it can effectively reduce trajectory tracking errors and improve the reliability and safety of autonomous driving or assisted driving systems.

[0087] This application further proposes a method for determining the expected road segment of a vehicle among multiple candidate road segments. This method improves the accuracy and robustness of the expected road segment selection by comprehensively considering the consistency of the vehicle's positioning information and feature similarity.

[0088] Specifically, road geometric feature information includes road slope variation information and road curve curvature variation information. Based on vehicle driving motion feature information and road geometric feature information of multiple candidate road segments, the expected road segment for the vehicle is determined from multiple candidate road segments, including:

[0089] For each candidate road segment among multiple candidate road segments, the similarity between the road geometric feature information of the candidate road segment and the driving motion feature information of the vehicle is determined; the positional distance between the vehicle's BeiDou positioning information and the vehicle's IMU positioning information is determined; when the positional distance is greater than a preset positional distance threshold, the candidate road segment with a similarity greater than the preset similarity threshold is taken as the vehicle's expected road segment; when the positional distance is less than or equal to the preset positional distance threshold, the candidate road segment with the highest similarity is taken as the vehicle's expected road segment.

[0090] Road geometric feature information can be understood as a description of the physical form of the road stored in the pre-defined map data, specifically including information on road slope changes and road curvature changes. Slope changes describe the longitudinal undulations of the road, while curvature changes describe the lateral curvature of the road. This information is crucial for vehicle trajectory tracking and path planning.

[0091] Vehicle motion characteristics include information on changes in vehicle gradient and changes in vehicle steering angle. Gradient changes reflect the vehicle's current longitudinal gradient, while steering angle changes reflect the vehicle's current lateral steering tendency.

[0092] Similarity refers to the degree of matching between the road geometric features of candidate road segments and the driving motion features of vehicles. This similarity can be calculated using various algorithms, such as those based on Euclidean distance, cosine similarity, correlation coefficient, or dynamic time warping (DTW). Its purpose is to assess the degree of agreement between the vehicle's current motion trend and the geometric features of each candidate road segment.

[0093] Location distance refers to the spatial distance between a vehicle's BeiDou positioning information and its IMU positioning information. This distance is used to measure the degree of consistency or difference between location data provided by two different positioning sources. When the location distance is small, it indicates that the two positioning information are highly consistent and the positioning reliability is high; when the location distance is large, it indicates that the two positioning information have a large deviation and the positioning reliability may be low.

[0094] The preset location distance threshold is a pre-defined distance value used to differentiate the consistency of location information. When the location distance exceeds this threshold, the system considers the location information to have significant uncertainty; when the location distance is less than or equal to the threshold, the system considers the location information to be relatively reliable.

[0095] The preset similarity threshold is a pre-defined similarity value used to filter out candidate road segments that have a sufficiently high degree of matching with the vehicle's motion characteristics. Only when the similarity of a candidate road segment exceeds this threshold is the road segment considered to have a significant matching relationship with the vehicle's motion trend.

[0096] This application's solution dynamically adjusts the selection strategy for expected road segments by introducing a judgment on the consistency of vehicle positioning information. Specifically, when the positional distance between the vehicle's BeiDou positioning information and IMU positioning information is small (i.e., less than or equal to a preset positional distance threshold), it indicates that the vehicle's positioning accuracy is high. In this case, the result of feature matching can be trusted more, so the candidate road segment with the highest similarity is directly selected as the vehicle's expected road segment to achieve the most accurate road matching. However, when the positional distance is large (i.e., greater than the preset positional distance threshold), it indicates that the vehicle's positioning information has significant uncertainty or error. If only the road segment with the highest similarity is selected in this case, incorrect road matching may occur due to positioning errors. Therefore, in this case, this application's solution adopts a more cautious strategy, only selecting candidate road segments with similarity greater than a preset similarity threshold as the vehicle's expected road segment. This approach effectively avoids misjudgment caused by a single high similarity when positioning uncertainty is high, thereby improving the robustness of road matching.

[0097] Through the above technical solution, this application can adaptively adjust the strategy for determining the expected road segment based on the reliability of vehicle positioning information. When the positioning information is reliable, the road segment with the highest feature matching degree is selected first to ensure high-precision tracking; when the positioning information has significant uncertainty, stricter similarity screening conditions are adopted to avoid mismatches caused by positioning errors. This dynamic adjustment mechanism significantly improves the accuracy and robustness of urban buses' trajectory tracking in complex environments, especially in challenging scenarios such as blocked BeiDou positioning signals or IMU data drift, enabling more reliable determination of the vehicle's expected driving path, thereby providing more stable and precise commands for vehicle drive control.

[0098] This application further proposes a step for determining the vehicle's expected road segment from multiple candidate road segments based on the vehicle's driving motion characteristics and road geometric characteristics of multiple candidate road segments, including:

[0099] For each candidate road segment among multiple candidate road segments, the similarity between the road geometric feature information of the candidate road segment and the driving motion feature information of the vehicle is determined; it is determined whether the difference between the largest and second largest similarity among multiple similarities is greater than a preset similarity difference threshold; when the difference between the largest and second largest similarity among multiple similarities is greater than the preset similarity difference threshold, the candidate road segment corresponding to the largest similarity is taken as the vehicle's expected road segment; when the difference between the largest and second largest similarity among multiple similarities is less than or equal to the preset similarity difference threshold, the candidate road segment with a similarity greater than the preset similarity threshold is taken as the vehicle's expected road segment.

[0100] Specifically, vehicle motion characteristic information can be understood as a set of parameters describing the vehicle's current motion state and trend, including information on changes in vehicle gradient and changes in vehicle steering angle. The gradient change information reflects the vehicle's vertical motion trend, such as the degree of uphill or downhill movement; the steering angle change information reflects the vehicle's horizontal steering trend, such as the degree of left or right turn. This information collectively characterizes the vehicle's real-time driving characteristics.

[0101] Meanwhile, road geometric feature information refers to the set of parameters describing the physical shape and orientation of a road, including information on road slope changes and road curvature changes. Road slope changes describe the longitudinal undulations of the road, while road curvature changes describe the lateral curvature. By comparing the vehicle's motion characteristics with the road geometric feature information of candidate road segments, their similarity can be calculated. This similarity measure quantifies the degree of matching between the vehicle's current driving state and the geometric features of a specific road segment.

[0102] Furthermore, the preset similarity difference threshold is a key parameter used to distinguish between "obvious best match" and "multiple approximate matches." When the difference between the largest and second largest similarities is greater than this threshold, it indicates that a candidate road segment has a much higher match degree with the vehicle's current driving state than other candidate road segments. In this case, the candidate road segment corresponding to the largest similarity can be confidently regarded as the vehicle's expected road segment. Conversely, when the difference is less than or equal to the preset similarity difference threshold, it means that multiple candidate road segments have very close matches with the vehicle's current driving state. In this case, more careful selection is needed, including all candidate road segments with similarities greater than the preset similarity threshold in the expected road segment category to avoid prematurely eliminating potentially correct road segments. The preset similarity threshold is used to filter out candidate road segments that have a certain degree of match with the vehicle's driving characteristics and exclude obviously irrelevant road segments.

[0103] This application's solution effectively addresses the problem of inaccurate road segment selection that can occur with traditional methods when multiple candidate road segments share similar characteristics by introducing a judgment on the difference between the maximum and second-highest similarity scores. Specifically, when a vehicle travels to a complex road environment, such as a multi-lane road section or a continuous curve area, the geometric features of multiple candidate road segments may exhibit similarities to the vehicle's motion characteristics. In this case, simply selecting the road segment with the highest similarity score carries the risk of misselection if the difference between the maximum and second-highest similarity scores is not significant.

[0104] This application intelligently distinguishes between "explicit matching" and "fuzzy matching" by determining whether the difference value exceeds a preset similarity difference threshold. When the difference is large, it indicates the existence of a significantly better match, and this best match is selected as the expected road segment, ensuring the accuracy of the selection. When the difference is small, it is assumed that there are multiple potential matching road segments. In this case, all candidate road segments with similarity higher than the preset similarity threshold are selected as expected road segments, thus avoiding making a single and potentially erroneous decision when information is insufficient or ambiguous, and providing more comprehensive road information for subsequent drive control.

[0105] Through the above technical solution, this application can significantly improve the accuracy and robustness of trajectory tracking for urban buses in complex road environments. Especially when faced with similar road features, errors in positioning information, or fluctuations in vehicle driving status, this method effectively avoids misjudgment of expected road segments due to close similarity by finely evaluating the matching degree between candidate road segments and vehicle driving characteristics. Therefore, the vehicle's drive control commands can be generated based on more accurate expected road segment information, thereby improving the safety, stability, and comfort of vehicle driving and reducing the risk of deviation or accidents in complex traffic scenarios.

[0106] It should be noted that the optimal value of each threshold mentioned in this application often depends on various factors such as the specific application scenario, vehicle type, sensor accuracy, map data resolution, and desired system performance. However, the methods for obtaining and determining these thresholds are conventional techniques well-known to those skilled in the art, mainly including the following methods:

[0107] 1. Empirical method and expert knowledge: Based on extensive practical driving experience and domain experts' understanding of vehicle behavior and road feature matching, a reasonable initial threshold range is determined through repeated trials and adjustments.

[0108] 2. Data-Driven Approach and Machine Learning: Collect a large amount of real-world driving data, including vehicle motion characteristics, high-precision positioning information, and corresponding actual driving roads. Using this data, train models with machine learning algorithms (such as classifiers and regression models) or optimization algorithms to automatically learn and determine thresholds that maximize matching accuracy and robustness. For example, cross-validation, grid search, and other methods can be used to find the optimal combination of thresholds within a certain range.

[0109] 3. Simulation and Testing: In a high-fidelity simulation environment, various complex driving scenarios are simulated. By systematically changing the threshold parameters, the system's performance in different scenarios is observed, and the optimal threshold is selected based on the expected performance indicators (such as matching accuracy, misjudgment rate, control smoothness, etc.).

[0110] Therefore, although this application does not provide specific numerical values, its description of "preset" indicates that these thresholds are pre-designed and determined, and that the method for obtaining them can be implemented by those skilled in the art through conventional technical means. This application further proposes a scheme for generating drive control commands by determining the number of expected road segments when determining the vehicle's drive control commands, and by invoking a preset drive control model when the number is one.

[0111] Specifically, when determining the vehicle's drive control commands based on the road geometry information of the intended road segment, the process includes:

[0112] Determine if the number of expected road segments for the vehicle is 1; if the number of expected road segments for the vehicle is 1, invoke the preset drive control model; the preset drive control model is used to determine the vehicle's drive control command based on the road geometry feature information of the vehicle's expected road segment and the vehicle's driving motion information; input the road geometry feature information of the vehicle's expected road segment and the vehicle's driving motion information into the preset drive control model to obtain the vehicle's drive control command.

[0113] The determination of whether the number of expected road segments for the vehicle is one refers to the system counting the set of determined expected road segments after identifying multiple candidate road segments and determining the vehicle's expected road segment. If the set contains only one road segment, the number of expected road segments is considered to be one. The preset drive control model can be understood as a pre-trained algorithm or program module designed to calculate the required drive control commands, such as steering, acceleration, or braking, based on the vehicle's current state (represented by driving motion information) and the characteristics of the road it will be traveling on (represented by the road geometry information of the expected road segment). This model can be built based on various control theories, such as PID control, model predictive control (MPC), or other machine learning methods.

[0114] In practical applications, inputting the road geometry information of the expected road segment and the vehicle's driving motion information into a preset drive control model means using these data as input parameters for the model. After internal calculation and processing, the model outputs corresponding drive control commands. For example, road geometry information may include road curvature, slope, lane width, etc., while driving motion information may include vehicle speed, acceleration, yaw rate, steering wheel angle, etc.

[0115] The proposed solution first determines the number of expected road segments for the vehicle. If only one expected road segment is identified, a pre-defined drive control model is directly invoked. This mechanism ensures that when the vehicle's travel path is highly deterministic, drive control commands can be generated quickly and accurately using an optimized control model. The pre-defined drive control model fully utilizes the vehicle's motion information and the road geometry of the expected road segment to perform refined control calculations, thereby providing the vehicle with stable and efficient trajectory tracking capabilities.

[0116] By employing the aforementioned technical solution, when the vehicle's intended road segment is uniquely determined, complex decision-making or weighted processing among multiple possibilities can be avoided, thereby simplifying the determination process of drive control commands and improving the system's response speed and real-time performance. Furthermore, by invoking a pre-defined drive control model specifically optimized for such well-defined scenarios, the generated drive control commands can be ensured to be more accurate and reliable, thus enhancing the overall performance and safety of urban bus trajectory tracking.

[0117] This application further proposes a method for determining the vehicle's drive control commands based on the road geometry information of the expected road segments when there are multiple expected road segments for the vehicle, specifically including:

[0118] When there are multiple expected road segments for the vehicle, for each expected road segment, the road distance between the expected road segment and the target location information is determined; the target location information is the midpoint of the line connecting the vehicle's Beidou positioning information and the vehicle's IMU positioning information; the multiple road distances are normalized, and the normalization result is used as the weight value corresponding to each expected road segment; based on the weight values ​​corresponding to the multiple expected road segments, the weighted sum of the road geometric feature information of the multiple expected road segments is used as the target road geometric feature information; a preset drive control model is invoked; the target road geometric feature information and the vehicle's driving motion information are input into the preset drive control model to obtain the vehicle's drive control command.

[0119] Specifically, when the system identifies multiple expected road segments, these segments need to be comprehensively evaluated to generate smoother and more robust drive control commands. The target location information refers to the midpoint of the line connecting the vehicle's BeiDou positioning information and its IMU positioning information. This midpoint is considered the best estimate of the current vehicle position and is used to measure the proximity of the expected road segment to the vehicle's current position. Determining the road distance between the expected road segment and the target location information involves calculating the actual road distance from the nearest point on each expected road segment to the target location information. This reflects the degree of matching between the vehicle and the expected road segment. Normalizing multiple road distances converts distance values ​​of different dimensions or ranges into uniform, comparable weight values. For example, normalization can be performed using the reciprocal of the distance, an exponential function, or a linear mapping, so that road segments that are closer receive higher weights.

[0120] Using the normalized result as the weight value for each expected road segment means that the importance of each expected road segment will be determined by its distance from the target location information; the closer the distance, the greater the weight. Using the weighted sum of the road geometric feature information of multiple expected road segments as the target road geometric feature information means multiplying the road geometric feature information of each expected road segment (such as road slope changes, road curvature changes, etc.) by its corresponding weight value, and then summing all the weighted feature information to obtain a comprehensive target road geometric feature information that represents the characteristics of all expected road segments. Calling the preset drive control model means activating a pre-trained control algorithm or strategy that can generate specific drive control commands based on the input road feature information and vehicle motion information. Inputting the target road geometric feature information and the vehicle's driving motion information into a preset drive control model to obtain the vehicle's drive control commands means inputting the integrated target road geometric feature information and the vehicle's current driving motion information (such as IMU accelerometer, gyroscope data, IMU yaw rate and steering wheel angle sensor data, etc.) into the drive control model, and the model calculates the steering, acceleration or braking commands required by the vehicle.

[0121] In this application, "target location information" refers to "the midpoint of the line connecting the vehicle's BeiDou positioning information and the vehicle's IMU positioning information." For each "expected road segment," the "road distance" between it and the target location information is determined, which typically refers to the shortest path distance from the target location information to the nearest point on the expected road segment. Specific implementation methods may include:

[0122] 1. For each expected road segment, first find the geometric point on the road segment that is closest to the target location information (e.g., by calculating the Euclidean distance from the target location information to all points on the road segment and taking the minimum value). Then, use this shortest Euclidean distance as the "road distance".

[0123] 2. Project the target location information onto the nearest road segment, and then calculate the actual travel distance from that projection point along the road network to a reference point on the expected road segment (e.g., the start of the expected road segment or the nearest point in the vehicle's current direction of travel).

[0124] The geometric feature information of the target road obtained through weighted processing is reasonable and has technical advantages. Its reasonableness is reflected in the following aspects:

[0125] 1. In complex traffic scenarios (such as multi-lane roads, complex intersections, and poor location signals), a vehicle's perception and positioning systems may not be able to definitively determine a single, expected road segment at any given moment. If the system forcibly selects one road for control in such situations, an incorrect selection could lead to sharp turns, accelerations, or braking, impacting driving safety and passenger comfort. Weighted processing mechanisms allow the system to avoid making a single, rigid decision when uncertainty exists, instead comprehensively considering all possibilities.

[0126] 2. Even if the bus ultimately travels only to one of the road segments, weighted processing provides a smooth transition control strategy. By weighted fusion of the geometric features of multiple anticipated road segments, the system generates drive control commands that are an "average" or "combined" reflection of these road features. For example, if there are two similar roads ahead, one slightly to the left and the other slightly to the right, weighted processing might generate a command to make a slight steering wheel adjustment, rather than an immediate large left or right turn. As the vehicle travels and more information is acquired, the weight of a particular anticipated road segment gradually increases, eventually dominating the control command, thus achieving a smooth transition from uncertainty to certainty.

[0127] 3. Weighted processing can improve the system's robustness to sensor noise and positioning errors. Even if there is a slight error in the identification of a certain expected road segment, since its weight may not be 100%, its impact on the final control command will be "diluted" by the characteristics of other expected road segments, thus avoiding the catastrophic impact of a single erroneous message on the control system.

[0128] In summary, although a bus can only travel on one road segment at a time, during the transitional phase when decisions are unclear, weighted processing that integrates the geometric features of multiple anticipated road segments can effectively manage uncertainty and achieve smoother, more robust, and safer trajectory tracking and drive control.

[0129] This application's solution addresses the challenge of effectively fusing information from multiple potential road segments to generate stable and accurate drive control commands when multiple driving paths exist. This is achieved by introducing a weighted processing mechanism for multiple anticipated road segments. When a vehicle is in complex road conditions, such as multi-lane forks or roundabouts, a single anticipated road segment may not accurately reflect the vehicle's true intention or optimal driving path. By calculating the road distance between each anticipated road segment and the vehicle's current target location and normalizing it as a weight value, the system can quantify the influence of each road segment on the vehicle's current driving decision. Road segments closer to the vehicle's current location have a higher weight, indicating a stronger guiding effect on the vehicle's current trajectory.

[0130] Subsequently, a comprehensive target road geometric feature information is generated by weighted summation of the road geometric feature information of all expected road segments. This information integrates the geometric characteristics of all potential paths, thus avoiding control discontinuities or decision-making errors that may result from rigid selection among multiple paths. Finally, this fused target road geometric feature information, along with the vehicle's driving motion information, is input into a preset drive control model. This allows the model to generate drive control commands based on more comprehensive and smoother road information, thereby improving the vehicle's trajectory tracking accuracy and control stability in complex environments.

[0131] Through the above technical solution, this application can effectively address the problem of multiple expected road segments encountered by urban buses in complex traffic scenarios. Compared to controlling only a single expected road segment, this solution uses weighted fusion of multiple expected road segments, enabling the generated drive control commands to more comprehensively consider the impact of all potential paths, thereby significantly improving the robustness and smoothness of trajectory tracking. Especially in areas with road forks, merging, or multi-lane roads, this method can avoid control oscillations or deviations caused by incorrect selection of a single path, ensuring that the vehicle makes more reasonable and safer driving decisions among multiple possible paths, thus improving the adaptability and reliability of the urban bus autonomous driving system.

[0132] This application proposes a BeiDou-based urban bus trajectory tracking system, comprising: an acquisition device and a processing device; wherein, the acquisition device is used to acquire vehicle driving motion information, vehicle BeiDou positioning location information, and vehicle IMU positioning information; the processing device is used to identify multiple candidate road segments from preset map data based on the vehicle's BeiDou positioning location information and vehicle IMU positioning information, and acquire road geometric feature information of each candidate road segment; the processing device is also used to determine the vehicle's expected road segment among the multiple candidate road segments based on the vehicle's driving motion information and the road geometric feature information of the multiple candidate road segments, so as to determine the vehicle's drive control command based on the road geometric feature information of the vehicle's expected road segment.

[0133] In a preferred embodiment, the acquisition device may include, but is not limited to: a BeiDou positioning module for receiving BeiDou satellite signals and calculating the vehicle's real-time geographic coordinates; an inertial measurement unit (IMU) for measuring the vehicle's angular velocity and acceleration, and providing information on the vehicle's relative position and attitude changes; and other vehicle sensors, such as wheel speed sensors, steering angle sensors, accelerometers, gyroscopes, etc., for real-time acquisition of vehicle motion data such as speed, acceleration, angular velocity, and steering angle. These sensors can be used independently or integrated to form the acquisition device, the purpose of which is to comprehensively and accurately perceive the vehicle's motion state and external positioning information.

[0134] It is important to emphasize that the processing device can be one or more processors, such as a central processing unit (CPU), microcontroller (MCU), digital signal processor (DSP), field-programmable gate array (FPGA), or application-specific integrated circuit (ASIC), which can be configured to execute preset program instructions to achieve the aforementioned functions. The processing device may also include memory for storing preset map data, program instructions, and various types of data generated during processing. In some embodiments, the processing device can be integrated into the vehicle's central control unit (VCU) or intelligent driving domain controller.

[0135] The above are merely embodiments of this application and are not intended to limit the scope of protection of 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 scope of protection of this application.

Claims

1. A method for tracking the trajectory of urban buses based on BeiDou positioning, characterized in that, include: Acquire vehicle driving motion information, vehicle BeiDou positioning information, and vehicle IMU positioning information; Based on the vehicle's BeiDou positioning information and IMU positioning information, multiple candidate road segments are identified from the preset map data, and the road geometric feature information of each candidate road segment is obtained. Based on the vehicle's driving motion information and the road geometry feature information of multiple candidate road segments, the expected road segment of the vehicle is determined among multiple candidate road segments, so as to determine the vehicle's drive control command based on the road geometry feature information of the expected road segment of the vehicle. Based on the vehicle's BeiDou positioning information and IMU positioning information, multiple candidate road segments are identified from preset map data, including: The midpoint of the line connecting the vehicle's BeiDou positioning information and the vehicle's IMU positioning information is taken as the target location information. Multiple candidate road segments are identified from preset map data based on the vehicle's drive control precision and target location information; The vehicle's drive control commands are determined based on the road geometry information of the intended road segment, including: Determine if the vehicle's expected road segment number is 1; When the number of expected road segments for the vehicle is 1, the preset drive control model is invoked; the preset drive control model is used to determine the vehicle's drive control commands based on the road geometry features of the expected road segments and the vehicle's driving motion information. The road geometry information of the expected road segment and the vehicle's driving motion information are input into the preset drive control model to obtain the vehicle's drive control commands. Determining the vehicle's drive control commands based on road geometry information of the intended road segment also includes: When there are multiple expected road segments for a vehicle, for each expected road segment, the road distance between the expected road segment and the target location information is determined; the target location information is the midpoint of the line connecting the vehicle's BeiDou positioning information and the vehicle's IMU positioning information. The distances to multiple roads are normalized, and the normalization result is used as the weight value corresponding to each expected road segment. Based on the weight values ​​corresponding to multiple expected road segments, the weighted sum of the road geometric feature information of multiple expected road segments is used as the target road geometric feature information; Call the preset drive control model; The target road geometry information and vehicle motion information are input into a preset drive control model to obtain the vehicle's drive control commands.

2. The urban bus trajectory tracking method based on BeiDou positioning according to claim 1, characterized in that, Based on the vehicle's drive control precision and target location information, multiple candidate road segments are identified from preset map data, including: The target radius is determined based on the vehicle's drive control precision; Multiple roads within the target area in the preset map data that have the same driving direction as the vehicle's driving direction are selected as multiple candidate road segments; the center of the target area is the target location information, and the radius of the target area is the target radius.

3. The urban bus trajectory tracking method based on BeiDou positioning according to claim 2, characterized in that, The target radius is determined based on the vehicle's drive control precision, including: Obtain the first correspondence; the first correspondence includes a one-to-one correspondence between multiple drive control accuracy ranges and multiple radii; The radius corresponding to the range of drive control accuracy of the vehicle in the first correspondence is taken as the target radius.

4. The urban bus trajectory tracking method based on BeiDou positioning according to claim 1, characterized in that, The vehicle's motion information includes IMU accelerometer and gyroscope data, IMU yaw rate, and steering wheel angle sensor data. Based on the vehicle's motion information and road geometry features of multiple candidate road segments, the vehicle's expected road segment is determined from among these candidate segments, including: The vehicle's gradient change information is determined based on the accelerometer and gyroscope data from the IMU. The vehicle's steering angle change information is determined based on the yaw rate of the IMU and the steering wheel angle sensor data, and the vehicle's slope change information and steering angle change information are used as the vehicle's driving motion characteristic information. The expected road segment for the vehicle is determined based on the vehicle's driving motion characteristics and the road geometry characteristics of multiple candidate road segments.

5. The urban bus trajectory tracking method based on BeiDou positioning according to claim 4, characterized in that, Road geometric feature information includes road slope variation information and road curve curvature variation information. Based on vehicle driving motion feature information and road geometric feature information of multiple candidate road segments, the expected road segment for the vehicle is determined from multiple candidate road segments, including: For each candidate road segment among multiple candidate road segments, determine the similarity between the road geometric feature information and the vehicle driving motion feature information of the candidate road segment; Determine the location distance between the vehicle's BeiDou positioning information and the vehicle's IMU positioning information; When the location distance is greater than a preset location distance threshold, candidate road segments with a similarity greater than a preset similarity threshold are used as the vehicle's expected road segment; When the location distance is less than or equal to a preset location distance threshold, the candidate road segment with the highest similarity is selected as the vehicle's expected road segment.

6. The urban bus trajectory tracking method based on BeiDou positioning according to claim 1, characterized in that, Vehicle motion characteristics include slope change information and steering angle change information; road geometry characteristics include slope change information and curvature change information. Based on the vehicle motion characteristics and road geometry characteristics of multiple candidate road segments, the expected road segment for the vehicle is determined from among multiple candidate road segments, including: For each candidate road segment among multiple candidate road segments, determine the similarity between the road geometric feature information and the vehicle driving motion feature information of the candidate road segment; Determine whether the difference between the largest and second largest similarity scores among multiple similarity scores is greater than a preset similarity difference threshold; When the difference between the largest and second largest similarity scores among multiple similarity scores is greater than a preset similarity difference threshold, the candidate road segment corresponding to the largest similarity score is taken as the vehicle's expected road segment. When the difference between the largest and second largest similarity scores among multiple similarity scores is less than or equal to a preset similarity difference threshold, the candidate road segment with a similarity score greater than the preset similarity threshold is taken as the vehicle's expected road segment.

7. A city bus trajectory tracking system based on BeiDou positioning, characterized in that, include: Acquisition device and processing device; Acquisition device, used to acquire vehicle driving motion information, vehicle Beidou positioning location information and vehicle IMU positioning information; The processing device is used to identify multiple candidate road segments from preset map data based on the vehicle's Beidou positioning information and the vehicle's IMU positioning information, and to obtain the road geometric feature information of each candidate road segment among the multiple candidate road segments. The processing device is also used to determine the expected road segment of the vehicle among multiple candidate road segments based on the vehicle's driving motion information and the road geometric feature information of multiple candidate road segments, so as to determine the vehicle's drive control command based on the road geometric feature information of the vehicle's expected road segment. The processing unit is used to identify multiple candidate road segments from preset map data based on the vehicle's BeiDou positioning information and the vehicle's IMU positioning information, including: The midpoint of the line connecting the vehicle's BeiDou positioning information and the vehicle's IMU positioning information is taken as the target location information. Multiple candidate road segments are identified from preset map data based on the vehicle's drive control precision and target location information; The processing unit is also used to determine drive control commands for the vehicle based on road geometry information of the expected road segment, including: Determine if the vehicle's expected road segment number is 1; When the number of expected road segments for the vehicle is 1, the preset drive control model is invoked; the preset drive control model is used to determine the vehicle's drive control commands based on the road geometry features of the expected road segments and the vehicle's driving motion information. The road geometry information of the expected road segment and the vehicle's driving motion information are input into the preset drive control model to obtain the vehicle's drive control commands. The processing unit is also used to determine the vehicle's drive control commands based on road geometry information of the expected road segment, and further includes: When there are multiple expected road segments for a vehicle, for each expected road segment, the road distance between the expected road segment and the target location information is determined; the target location information is the midpoint of the line connecting the vehicle's BeiDou positioning information and the vehicle's IMU positioning information. The distances to multiple roads are normalized, and the normalization result is used as the weight value corresponding to each expected road segment. Based on the weight values ​​corresponding to multiple expected road segments, the weighted sum of the road geometric feature information of multiple expected road segments is used as the target road geometric feature information; Call the preset drive control model; The target road geometry information and vehicle motion information are input into a preset drive control model to obtain the vehicle's drive control commands.

Images