Methods and devices for determining the road segment where the vehicle is located, fusion positioning module and map engine

Abstract

This specification discloses a method, apparatus, fusion positioning module, and high-precision map engine for determining the road segment where a vehicle is located. The scheme may include: acquiring vehicle driving information of a target vehicle at a first moment; and based on the vehicle driving information, determining a set of candidate matching road segments for the target vehicle in map data at the first moment; then, for each road segment in the candidate matching road segment set, determining first matching degree information between the target vehicle and each road segment at the first moment; further, based on the first matching degree information between the target vehicle and each road segment at the first moment, determining the matching probability between the target vehicle and each road segment; finally, determining the road segment corresponding to the maximum matching probability as the road segment where the target vehicle is located.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electronic maps, and in particular to a method and device for determining a link in which a vehicle is located, a fusion positioning module, and a high-precision map engine. BACKGROUND

[0002] With the rise of autonomous driving, map matching technology for matching vehicle positioning data with map data is playing an increasingly important role.

[0003] After sensor data is processed by a fusion positioning algorithm, a latitude and longitude positioning result can be obtained, and after the latitude and longitude positioning result is matched with map data, road positioning information or lane positioning information of the vehicle can be obtained. The process of matching the latitude and longitude positioning result with the map data is a map matching process.

[0004] A conventional map matching method determines a link in which a vehicle is located only based on position information and a heading angle of the vehicle. However, in a complex road scene, the matching accuracy is not high only based on the position information and the heading angle. SUMMARY

[0005] The embodiments of the present specification provide a method and device for determining a link in which a vehicle is located, a fusion positioning module, and a high-precision map engine to solve the problem of low matching accuracy when matching a vehicle with a link in map data.

[0006] To solve the above technical problems, the embodiments of the present specification are implemented as follows:

[0007] The method for determining a link in which a vehicle is located provided by the embodiments of the present specification comprises:

[0008] Obtaining vehicle driving information of a target vehicle at a first time; the vehicle driving information comprises at least one of latitude and longitude information and driving direction information; determining a set of candidate matching links in map data in which the target vehicle is located at the first time based on the vehicle driving information; determining first matching degree information of the target vehicle and each link in the set of candidate matching links at the first time; determining a matching probability of the target vehicle and each link based on the first matching degree information of the target vehicle and each link at the first time; and determining a link corresponding to a maximum value of the matching probability as a link in which the target vehicle is located.

[0009] The device for determining a link in which a vehicle is located provided by the embodiments of the present specification comprises:

[0010] The information acquisition module is used to acquire the vehicle driving information of the target vehicle at a first moment; the vehicle driving information includes at least one of latitude and longitude information and driving direction information.

[0011] The alternative matching road segment determination module is used to determine the set of alternative matching road segments of the target vehicle in the map data at the first moment based on the vehicle driving information.

[0012] The matching degree information determination module is used to determine the first matching degree information between the target vehicle and each road segment in the candidate matching road segment set at the first time.

[0013] The matching probability determination module is used to determine the matching probability between the target vehicle and each road segment based on the first matching degree information between the target vehicle and each road segment at the first time.

[0014] The road segment determination module is used to determine the road segment corresponding to the maximum matching probability as the road segment where the target vehicle is located.

[0015] This specification provides an embodiment of a fusion positioning module, including the aforementioned device for determining the road segment where a vehicle is located. The device for determining the road segment where the vehicle is located is used to determine the vehicle's location on a high-precision map, assist in cross-validation with data from other vehicle sensors, achieve high-precision fusion positioning, and obtain the vehicle's precise location. The other vehicle sensors include at least one of inertial navigation, GNSS / RTK, vision, and lidar.

[0016] This specification provides an embodiment of a high-precision map engine, comprising:

[0017] The aforementioned fusion positioning module;

[0018] The electronic horizon module is used to receive external high-precision vehicle location information and match it to a map, providing a functional interface for autonomous driving applications to make control and judgments.

[0019] In addition, at least one of the following modules: autonomous driving design and operation domain judgment module, map update module, crowdsourcing preprocessing and feedback module, path intersection association module, and lane-level path planning module;

[0020] The autonomous driving design operation domain judgment module is used to configure the autonomous driving area and judgment requirements.

[0021] The map update module is used to obtain map data update information of a high-precision map based on the vehicle's location and the planned route.

[0022] The crowdsourcing preprocessing and feedback module is used to preprocess UGC visual vector data by filtering and fusion, and then feed it back to the cloud and update the map data center.

[0023] The path cross-association module is used to synchronize the global path planning results initiated by the user to the autonomous driving system, and obtain the matching path of the navigation path on the high-precision map by cross-associating with the high-precision map.

[0024] The lane-level path planning module is used to output lane-level local path planning within a certain length range in front of the vehicle based on the results of navigation path matching and route correction.

[0025] One embodiment of this specification can achieve at least the following beneficial effects: by treating the state of the vehicle during its driving process as a sequence, when calculating the matching probability of the vehicle with each road segment at the current moment, the matching probability of the vehicle with each road segment at the previous moment is referenced, thereby improving the accuracy of calculating the matching probability of the vehicle with each road segment, and thus improving the accuracy of determining the matching result of the road segment where the vehicle is located. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 A flowchart illustrating a method for determining the road segment where a vehicle is located, provided in an embodiment of this specification;

[0028] Figure 2 A schematic diagram of a branching road application scenario provided in the embodiments of this specification;

[0029] Figure 3 This is a schematic diagram of an application scenario of upper and lower level roads provided in the embodiments of this specification;

[0030] Figure 4 This is a schematic diagram illustrating another branching path application scenario provided in the embodiments of this specification;

[0031] Figure 5 This is a flowchart illustrating the training method for a deep learning model used to predict the road where a vehicle is located, as provided in the embodiments of this specification.

[0032] Figure 6 The embodiments provided in this specification correspond to Figure 1 A schematic diagram of the structure of a device for determining the road segment where a vehicle is located;

[0033] Figure 7This is a schematic diagram of the structure of a high-precision map engine provided in the embodiments of this specification. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of one or more embodiments of this specification clearer, the technical solutions of one or more embodiments of this specification will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of them. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of one or more embodiments of this specification.

[0035] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another.

[0036] The technical solutions provided in the various embodiments of this specification are described in detail below with reference to the accompanying drawings.

[0037] Sensor data obtained from sensors such as Global Navigation Satellite System (GNSS) and Inertial Measurement Unit (IMU) is processed by a fusion positioning algorithm to obtain a latitude and longitude positioning result. This result can then be matched with map data to obtain the vehicle's road or lane positioning information. This matching process is known as map matching. Map matching technology plays a crucial role as the link between positioning information and map data in the application of map data to driving navigation, including using high-precision map data for vehicle navigation and high-resolution map data for autonomous driving navigation.

[0038] In existing technologies, traditional map matching methods are based on high-precision or standard-precision maps. They determine the current road segment by inputting the angle difference between the vehicle's heading angle and the road segment's extension direction, and calculating the vertical distance from the vehicle's location point to the road segment. The problem with this method is that for complex road scenarios, such as intersections or multi-level roads, relying solely on the vehicle's position and heading angle cannot accurately pinpoint the vehicle's location, thus failing to effectively provide data support for autonomous driving.

[0039] To address the shortcomings of existing technologies, in the embodiments of this specification, considering that the vehicle's location points acquired at a certain frequency are regarded as a sequence, the Viterbi algorithm based on Hidden Markov Model (HMM) is adopted. When determining the road segment where the vehicle is located at the current moment, the road segment matching result of the vehicle at the previous moment can be referenced. That is, the road segment matching result of the vehicle at the current position and the road segment matching result of the vehicle at the previous moment are comprehensively considered, thereby improving the accuracy of road segment matching.

[0040] Figure 1 This is a flowchart illustrating a method for determining the road segment where a vehicle is located, as provided in an embodiment of this specification.

[0041] From a programming perspective, the entity executing the process can be a program hosted on an application server or application terminal. It can be understood that this method can be executed by any device, equipment, platform, or cluster of devices with computing and processing capabilities.

[0042] like Figure 1 As shown, the process may include the following steps:

[0043] Step 102: Obtain the vehicle driving information of the target vehicle at the first moment; the vehicle driving information includes at least one of latitude and longitude information and driving direction information.

[0044] In practical applications, the vehicle's state (including its position and location on the road) constantly changes during its journey. Therefore, the changing states of the target vehicle during its journey can be viewed as a time-varying sequence of states. At different times, the target vehicle corresponds to different states within this sequence. Based on... Figure 1 The method described above can determine the state of a target vehicle at a certain moment (e.g., the first moment), and more specifically, can determine the road segment where the target vehicle is located at a certain moment (e.g., the first moment).

[0045] In the embodiments of this specification, before determining the road segment where the target vehicle is located, that is, before matching the target vehicle with the road segment in the map data, it is necessary to obtain the target vehicle's own driving information. Specifically, the target vehicle's own driving information may include at least one of the target vehicle's latitude and longitude information and the target vehicle's driving direction information. Only then can the target vehicle be matched to the road segment in the map data based on the target vehicle's own driving information.

[0046] The vehicle's driving information can be determined based on sensor data collected by sensors, such as those from Global Navigation Satellite System (GNSS) and Inertial Measurement Unit (IMU). Additionally, vehicle speed information and information obtained from the vehicle's odometer can also be used to determine the vehicle's driving information. Optionally, sensor data collected by other vehicles or roadside equipment can also be used. Any existing method can be used to determine the vehicle's driving information; this application does not impose any specific limitations.

[0047] In a preferred embodiment, the acquired vehicle driving information of the target vehicle can be lane-level vehicle driving information. For example, the accuracy of the latitude and longitude information in the vehicle driving information can be at the centimeter level. Using lane-level vehicle driving information can improve the accuracy of subsequent road segment matching.

[0048] Step 104: Based on the vehicle driving information, determine the set of candidate matching road segments for the target vehicle in the map data at the first moment.

[0049] In the implementation of this specification, determining the road segment where the vehicle is located specifically means identifying the road segment with the highest matching degree to the vehicle in the map data based on the vehicle's driving information, and considering this as the road segment where the vehicle is currently actually traveling. The map data may specifically include standard precision map data or high precision map data.

[0050] In determining the road segment where a vehicle is located, the vehicle's expected forward travel range can be defined based on prior values ​​in the map data. From this expected forward travel range, a set of alternative matching road segments that may be the vehicle's current location can be identified. Specifically, when defining the vehicle's expected forward travel range, if the vehicle's driving information is at the lane level, the expected forward travel range can be defined according to the lane's defined forward travel range (e.g., the forward travel ranges differ for straight lanes, left-turn lanes, right-turn lanes, and U-turn lanes).

[0051] In practical applications, the step of determining candidate matching road segments in step 104 may specifically include: based on the position of the projection point of the target vehicle perpendicular to the map road segment and the projection distance between the target vehicle and the map road segment, determining the map road segments whose projection points fall within the target road segment and whose projection distances meet preset conditions as candidate matching road segments that meet the conditions.

[0052] More specifically, based on the latitude and longitude information of the target vehicle, the projected position of the target vehicle relative to each map road segment can be determined. Then, it is determined whether the projected position is located on the map road segment. If the projected position is outside the map road segment, the current map road segment is considered not to meet the conditions for candidate matching road segments, and the process ends. If the projected position is located on the map road segment, the projected distance between the target vehicle and the projected position is calculated. After calculating the projected distance between the target vehicle and the projected position, it is determined whether the projected distance is less than a preset distance threshold. If the projected distance is greater than or equal to the preset distance threshold, the current map road segment is considered not to meet the conditions for candidate matching road segments, and the process ends. If the projected distance is less than the preset distance threshold, the map road segment is determined as a road segment in the candidate matching road segment set.

[0053] In an optional embodiment, the candidate matching road segments can be further filtered based on the target vehicle's driving direction information. Specifically, after determining that the projected distance is less than a preset distance threshold, the process may further include: determining the angle difference between the vehicle's driving direction and the extension direction of each map road segment based on the target vehicle's driving direction information; and determining whether the angle difference is less than a preset angle threshold. If the angle difference is greater than or equal to the preset angle threshold, the current map road segment is considered not to meet the conditions for a candidate matching road segment, and the process ends; if the angle difference is less than the preset angle threshold, the map road segment is determined as a road segment in the candidate matching road segment set.

[0054] For ease of understanding, Figure 2 The diagram shows a schematic of a branching road application scenario according to an embodiment of this specification.

[0055] exist Figure 2 In the diagram, the rectangle represents the target vehicle, the arrow on the rectangle indicates the vehicle's heading, heading1 represents the extension direction of link 3, and heading2 represents the extension direction of link 4. In step 102, the target vehicle's position and heading are determined. In step 104, the distances and angles between the target vehicle and each road segment (e.g., link 3 and link 4) within its expected forward travel range are calculated. Figure 2 The distance between the target vehicle and link3 is d1, and the distance between the target vehicle and link4 is d2. Furthermore, the angle difference between the target vehicle's direction of travel and the extension direction of link3 is the angle between heading and heading1 (for example, it can be denoted as α1). Figure 2(α1 is not shown in the diagram). The angle difference between the target vehicle's driving direction and the extension direction of link4 is the angle between heading and heading2 (for example, it can be denoted as α2). Figure 2 (α2 is not shown in the text).

[0056] Based on the method in step 104, several road segments can be identified as candidate matching road segments. For example, this may include one, two, or more road segments. The solution in this embodiment mainly addresses the problem of determining the actual road segment where the target vehicle is located when two or more road segments are matched in step 104. In practical application scenarios, this may include road forks, ramp entrances, etc.

[0057] In existing technologies, when encountering complex road scenarios such as intersections, ramps, and upper / lower level roads, it is often impossible to accurately determine the vehicle's current location based solely on the vehicle's position, driving direction, and the location and extension direction of the road segment. Furthermore, considering that vehicles may change lanes in complex road scenarios such as intersections, ramps, and upper / lower level roads during actual driving, relying solely on the aforementioned distance and direction information could lead to significant errors.

[0058] In the embodiments of this specification, after determining the set of candidate matching road segments according to step 104, the actual road segment where the target vehicle is located can be determined from it according to steps 106, 108, and 110. The actual road segment where the target vehicle is located means the road segment where the target vehicle is most likely to actually be traveling, as determined by the method in the embodiments of this specification. Figure 2 If the set of candidate matching road segments includes link3 and link4, then it is necessary to determine whether the target vehicle is actually located on link3 or link4 according to the embodiments of this specification.

[0059] Step 106: For each road segment in the candidate matching road segment set, determine the first matching degree information between the target vehicle and each road segment at the first time.

[0060] Since the vehicle's location points are considered as a sequence, and the Hidden Markov Model (HMM) is well-suited for solving sequence problems, when encountering complex road scenarios such as intersections, ramps, and roads on different levels, in steps 106 to 110, the HMM model and the Viterbi algorithm are used. By calculating the measurement probability of the vehicle's current position and the transition probability from the vehicle's previous position to the current position, the probability of each road segment is calculated using the Viterbi algorithm, and the road segment with the highest probability value is determined as the road segment where the vehicle is located.

[0061] In practical applications, step 106 may specifically include: for each road segment in the candidate matching road segment set, determining the first matching degree information between the target vehicle and each road segment at the first time based on the pose change information of the target vehicle at the first time relative to the previous time and the relative pose information of the target vehicle at the first time relative to each road segment.

[0062] Specifically, step 106 may include: calculating the transfer probability and measurement probability of the target vehicle, and determining the first matching degree information between the target vehicle and the target road segment in the set of alternative matching road segments based on the transfer probability and the measurement probability.

[0063] More specifically, step 106 may include: on the one hand, for the target road segment in the candidate matching road segment set, determining the transition probability based on the vehicle driving information of the target vehicle at the first moment and the vehicle driving information of the target vehicle at the previous moment; the transition probability is used to reflect the probability that the target vehicle changes from the pose corresponding to the previous moment to the pose corresponding to the first moment; on the other hand, for the target road segment, determining the measurement probability based on the vehicle driving information of the target vehicle at the first moment and the road segment attribute information of the target road segment in the map data; the measurement probability is used to reflect the correlation between the relative pose information of the target vehicle relative to the target road segment and the actual location of the target vehicle in the target road segment; then, based on the transition probability and the measurement probability, determining the first matching degree information between the target vehicle and the target road segment, for example, the transition probability can be multiplied by the measurement probability to obtain the first matching probability between the target vehicle and the target road segment.

[0064] Transition probability is an important concept in Markov chains. If a Markov chain consists of m states, historical data can be transformed into a sequence of these m states. Starting from any state, after any transition, one of the states 1, 2, ..., m will inevitably appear. This transition between states is called the transition probability.

[0065] In the embodiments of this specification, the transition probability can refer to the probability of moving from the vehicle's previous position to its current position. A method for determining the transition probability may specifically include: calculating the position change of the target vehicle from the previous moment to the first moment based on the latitude and longitude information of the target vehicle at the first moment and the latitude and longitude information of the target vehicle at the previous moment; and determining the transition probability of the target vehicle from the previous moment to the first moment based on the position change. The transition probability may be negatively correlated with the position change. In the embodiments of this specification, a negative correlation between the first physical quantity and the second physical quantity can mean that the first physical quantity decreases as the second physical quantity increases, and vice versa.

[0066] In practical applications, the distance from the vehicle's previous position to its current position can be used as input to calculate the transition probability. Alternatively, the transition probability can be calculated using a preset function (e.g., the exp function).

[0067] Measurement probability is the probability of transitioning from a hidden state to a visible state, also known as emission probability.

[0068] In the embodiments of this specification, the implicit state represents the vehicle's driving information (including latitude and longitude positioning information and driving direction information), and the visible state represents the road segment where the vehicle is located. The method for determining the measurement probability may specifically include: calculating the distance between the target vehicle and the target road segment at the first moment based on the latitude and longitude information of the target vehicle at the first moment; determining a first probability related to the distance based on the distance, wherein the first probability is negatively correlated with the distance; calculating the angle difference between the vehicle's driving direction at the first moment and the extension direction of the target road segment; determining a second probability related to the angle difference based on the angle difference, wherein the second probability is negatively correlated with the angle difference; and calculating a measurement probability based on the first probability and the second probability, wherein the measurement probability is positively correlated with both the first and second probabilities. For example, the measurement probability can be obtained by summing the first probability and the second probability. In the embodiments of this specification, a positive correlation between the first physical quantity and the second physical quantity may mean that the first physical quantity increases as the second physical quantity increases, and the first physical quantity decreases as the second physical quantity decreases.

[0069] In practical applications, the above text will be followed. Figure 2 The example in the example can take d1, d2, α1, and α2 as inputs and calculate the measurement probability using a preset function (e.g., the exp function).

[0070] Specifically, the first probability is calculated based on the distance between the target vehicle and the target road segment at the first moment. For example, the first probability corresponding to link3 can be calculated based on d1 and the first probability is negatively correlated with d1, and the first probability corresponding to link4 can be calculated based on d2 and the first probability is negatively correlated with d2.

[0071] The second probability is calculated based on the angle difference between the vehicle's direction of travel at the first moment and the extension direction of the target road segment. For example, the second probability corresponding to link3 can be calculated based on α1 and the second probability is negatively correlated with α1, and the second probability corresponding to link4 can be calculated based on α2 and the second probability is negatively correlated with α2.

[0072] The measurement probability of a road segment at the first time step is calculated based on the first probability and the second probability. For example, the measurement probability of link 3 at the first time step can be obtained by adding the first probability corresponding to link 3 to the second probability corresponding to link 3 (for example, denoted as k). 测量-link3 Add the first probability corresponding to link4 to the second probability corresponding to link4 to obtain the measurement probability corresponding to link4 at the first time point (e.g., denoted as k). 测量-link4 ).

[0073] Combining the aforementioned calculation of the road segment's transition probability at the first time based on the target vehicle's position change from the second time point to the first time point (e.g., denoted as D), for example, we obtain the transition probability corresponding to link3 at the first time point (e.g., denoted as k). 转移-link3 ), and the transition probability corresponding to link4 at the first time step (e.g., denoted as k). 转移-link4 Then, the first matching probability k corresponding to link3 at the first time step can be calculated. 第一匹配-link3 =k 测量-link3 +k 转移-link3 The matching probability k corresponding to link4 at the first moment 第一匹配-link4 =k 测量-link4 +k 转移-link4 .

[0074] It should be noted that the k given above... 第一匹配-link3 and k 第一匹配-link4 The calculation method described is merely an example for illustrative purposes. In practical applications, other calculation methods can be used to calculate the first matching degree information between the target vehicle and each road segment at the first moment.

[0075] Step 108: Based on the first matching degree information between the target vehicle and each road segment at the first time, determine the matching probability between the target vehicle and each road segment.

[0076] In an optional embodiment, determining the matching probability of the target vehicle and each road segment based on the first matching degree information of the target vehicle and each road segment at the first time point may specifically include: obtaining the second matching degree information of the target vehicle and each road segment at the previous time point; and determining the matching probability of the target vehicle and each road segment based on the first matching degree information and the second matching degree information. Hereinafter, the previous time point refers to the time preceding the first time point, and may also be referred to as the second time point.

[0077] The second matching degree information can be determined based on the pose change information of the target vehicle at the second time point relative to the third time point and the relative pose information of the target vehicle at the second time point relative to each road segment. The third time point is earlier than the second time point, or in other words, the third time point is a time preceding the second time point. The specific method for determining the second matching degree information can be similar to the specific method for determining the first matching degree information. For example, the transition probability and measurement probability of the target road segment at the second time point can be calculated first, and then the transition probability and measurement probability of the target road segment at the second time point can be multiplied to obtain the second matching probability between the target vehicle and the target road segment.

[0078] In step 108, specifically, based on the first matching probability of the target vehicle and the target road segment at the first moment, a second matching probability at a second moment before the first moment can be further considered to obtain the overall matching probability of the target vehicle and the target road segment. Optionally, the matching probability of the target vehicle and the target road segment can be obtained by calculating the product of the first matching probability and the second matching probability.

[0079] Using the example above, for link3, we can calculate the first matching probability k between the target vehicle and link3 at the first time step. 第一匹配-link3 And the second matching probability k between the target vehicle and link3 at the second time step. 第二匹配-link3 And then based on k 第一匹配-link3 and k 第二匹配-link3 Obtain the matching probability between the target vehicle and link3, optionally, k 匹配-link3 =k 第一匹配-link3 ×k 第二匹配-link3 Similarly, for link4, the first matching probability k between the target vehicle and link4 at the first moment can be calculated. 第一匹配-link4 And the second matching probability k between the target vehicle and Link4 at the second time step. 第二匹配-link4 And then based on k 第一匹配-link4 and k 第二匹配-link4 Obtain the matching probability between the target vehicle and link4, optionally, k 匹配-link4 =k 第二匹配-link4×k 第二匹配-link4 .

[0080] In practical applications, after determining the set of candidate matching road segments in step 104, in steps 106 and 108, each road segment in the set of candidate matching road segments can be traversed to calculate the matching probability value between each road segment and the target vehicle.

[0081] Step 110: Determine the road segment corresponding to the maximum matching probability as the road segment where the target vehicle is located.

[0082] In practical applications, on the one hand, based on the matching probabilities of each road segment determined in step 108, the road segment with the highest matching probability can be identified as the optimal matching road segment for the target vehicle at the current moment (i.e., the first moment), serving as the current road segment of the target vehicle for navigation needs. Continuing with the example above, we can compare k... 匹配-link3 With k 匹配-link3 The value of the probability value is used to determine the optimal matching road segment for the target vehicle, which is considered to be the road segment where the target vehicle is currently located.

[0083] On the other hand, the matching probability of the target vehicle with each road segment at the current moment, as determined in step 108, can be stored for use when calculating the matching probability of the target vehicle with each road segment at the next moment.

[0084] It should be understood that in the methods described in one or more embodiments of this specification, the order of some steps may be adjusted according to actual needs, or some steps may be omitted.

[0085] Figure 1 The proposed method treats the vehicle's states during its journey as a sequence. When calculating the matching probability between the vehicle and each road segment at the current moment, it references the matching probability of the vehicle with each road segment at the previous moment. This improves the accuracy of calculating the matching probability between the vehicle and each road segment, thereby enhancing the accuracy of determining the vehicle's location on the road. In particular, for complex road scenarios such as intersections, ramps, and roads between different levels, it can significantly improve the accuracy of vehicle positioning in high-precision maps, providing strong support for driving navigation applications.

[0086] based on Figure 1 In addition to the method described herein, this specification also provides some specific implementation methods of this method, which will be described below.

[0087] In some complex road scenarios, such as elevated highways or other multi-level roads, the continuity between road segments is of great reference value for determining the current road segment of a vehicle. Therefore, in the embodiments of this specification, in the aforementioned... Figure 1Based on the proposed scheme, when calculating the road segment where the target vehicle is located at the current moment, the situation of the road segment where the target vehicle was located at the previous moment can be further referenced.

[0088] Specifically, before determining the matching probability between the target vehicle and each road segment based on the first matching degree information of the target vehicle and each road segment at the first time, the method may further include: for the target road segment in the set of candidate matching road segments, determining whether the target road segment is an associated road segment of the road segment where the target vehicle was at the previous time, and obtaining an associated road segment determination result.

[0089] The determination result of the associated road segment includes a first result and a second result. The first result indicates that the target road segment is an associated road segment of the road segment where the target vehicle was located at the previous moment; the second result indicates that the target road segment is not an associated road segment of the road segment where the target vehicle was located at the previous moment. Specifically, the target road segment being an associated road segment of the road segment where the target vehicle was located at the previous moment includes either the target road segment being the same road segment as the road segment where the target vehicle was located at the previous moment, or the target road segment being a connecting road segment to the road segment where the target vehicle was located at the previous moment.

[0090] Based on the aforementioned judgment, the step of determining the matching probability of the target vehicle and each road segment based on the first matching degree information of the target vehicle and each road segment at the first time moment may specifically include: if the associated road segment judgment result indicates that the target road segment is an associated road segment of the road segment where the target vehicle was located at the previous time moment, then a preset associated road segment weight coefficient is obtained; and the matching probability of the target vehicle and the target road segment is determined based on the first matching degree information of the target road segment and the associated road segment weight coefficient.

[0091] For example, the matching probability of a target road segment can be obtained by multiplying the probability obtained based on the first matching degree information (or first matching probability) and the second matching degree information (or second matching probability) by the associated road segment weight coefficient. The associated road segment weight coefficient can be a number greater than 1.

[0092] Figure 3 A schematic diagram of an application scenario of upper and lower level roads according to an embodiment of this specification is shown.

[0093] like Figure 3 As shown, the target vehicle (rectangle) is displayed at the current moment as its "current vehicle position" and heading, along with the two candidate matching road segments link1 and link2. Furthermore, the target vehicle's "previous state vehicle position" at the previous moment is also shown. Link1 is the connecting road segment to link4, and link2 is the connecting road segment to link3. Figure 3In the scenario shown, the vehicle is currently close to both link1 and link2, and the angles between the vehicle and link1 and link2 are also close. It is not possible to accurately determine the road segment where the vehicle is located based solely on distance and angle.

[0094] In practical applications, if link3 is determined to be the road segment where the target vehicle was located in the previous moment, then in the current moment, when determining the road segment where the vehicle is located from the alternative matching road segments link1 and link2, a preset coefficient can be multiplied when calculating the matching probability of link2, the connecting road segment of link3, to increase the calculated matching probability of the target vehicle with link2, thereby improving the accuracy of vehicle-road segment matching.

[0095] like Figure 3 If the matching probability k between the target vehicle and link1 at the current time has been calculated... 匹配-link1 And the matching probability k with link2 匹配-link2 Based on this, if link3 is the road segment where the target vehicle was located in the previous moment and link3 is the road segment that link2 continues, then the matching probability k of link2 can be calculated. 匹配-link2 To update, assuming the preset coefficient is a, then k 匹配-link2 The value is updated to a*k 匹配-link2 .

[0096] In the embodiments described above, considering the prior nature of high-precision maps, the connectivity attributes of lanes in the map data are used to add weight values ​​to the probabilities calculated by the HMM model, thereby improving the accuracy of vehicle location on the road segment.

[0097] In reality, due to limitations or defects in mapping technology, the direction of road segments in map data may deviate significantly from the actual direction of traffic, while the actual direction of vehicle travel is usually not much different. In such cases, traditional distance- and direction-based methods may have significant errors, potentially leading to misjudgments and the vehicle repeatedly veering across map road segments during navigation. In the embodiments described in this specification, a Hidden Markov Model (HMM) is used to consider historical state sequences, greatly improving accuracy. Furthermore, even if a misjudgment occurs, it is promptly corrected with updates to the positioning points, preventing repeated veering.

[0098] Based on this, in the embodiments of this specification, for specific road scenarios, a large amount of real-world vehicle driving data for these scenarios can be used to train a deep learning model corresponding to that specific road scenario. Then, when a vehicle travels to these specific road scenarios, it can... Figure 1Based on the proposed solution, when calculating the road segment where the target vehicle is located in the current road scenario, the prediction results of the pre-trained deep learning model corresponding to the current road scenario can be combined to determine the actual road segment where the target vehicle is located.

[0099] Specifically, before determining the matching probability of the target vehicle and each road segment based on the first matching degree information of the target vehicle and each road segment at the first time, the method may further include: determining whether the current scenario of the target vehicle is a preset specific scenario; if the current scenario of the target vehicle is a preset specific scenario, then using a pre-trained deep learning model corresponding to the preset specific scenario to calculate the model probability corresponding to each road segment in the candidate matching road segment set; the deep learning model is trained based on the vehicle's actual driving data under the preset specific scenario, and the vehicle's actual driving data includes vehicle driving information and the corresponding vehicle actual driving segment information in the data map; based on the model probability corresponding to each road segment, the model weight coefficient corresponding to each road segment is determined, and in practical applications, the model weight coefficient can be positively correlated with the model probability. Accordingly, determining the matching probability of the target vehicle and each road segment based on the first matching degree information of the target vehicle and each road segment at the first time may specifically include: determining the matching probability of the target vehicle and each road segment based on the first matching degree information of the target road segment and the model weight coefficient.

[0100] The method of calculating the model probability of each road segment in the candidate matching road segment set by using a pre-trained deep learning model corresponding to the preset specific scenario may specifically include: inputting the vehicle driving information of the current state and the road segment attribute information of each road segment in the candidate road segment set, and outputting the model probability of each road segment.

[0101] The step of determining the matching probability between the target vehicle and each road segment based on the first matching degree information of the target road segment and the model weight coefficient may specifically include multiplying the probability obtained based on the first matching degree information (or first matching probability) and the second matching degree information (or second matching probability) by the model weight coefficient. The model weight coefficient can be a number greater than 1.

[0102] In the above embodiments, based on the Hidden Markov Model (HMM), for specific road scenarios, a large amount of real-world driving data for that specific road scenario can be used as training data (e.g., distances from vehicles to the road boundaries on both sides, distances from vehicles to the lane boundaries on both sides, slopes of road segments or lanes, curvatures of road segments or lanes, lane connectivity, etc.). A deep learning model is trained for this scenario using logistic regression, resulting in a deep learning model corresponding to that specific road scenario. In application, the output of the deep learning model is the model probability values ​​corresponding to multiple candidate road segments. These model probability values ​​can add a weight value to the probability values ​​calculated based on the HMM model. Combined with the prior knowledge of high-precision maps, this improves the accuracy of determining the vehicle's location in that scenario.

[0103] Figure 4 A schematic diagram of another branching application scenario provided in the embodiments of this specification is shown.

[0104] exist Figure 4 In the diagram, the rectangle represents the target vehicle's position, the arrow on the rectangle indicates the vehicle's heading direction, heading1 represents the extension direction of link 3, and heading2 represents the extension direction of link 4. d1 represents the distance between the target vehicle and link 3, and d2 represents the distance between the target vehicle and link 4. Figure 4 In the example, the actual driving direction of the vehicle (heading, reflecting the actual direction of traffic on the road) differs significantly from the road segment extension direction (heading2) recorded in the map data, which typically leads to matching errors. However, the solution according to the embodiments of this specification, through pre-training such as... Figure 4 The deep learning model corresponding to the scenario shown can call the pre-trained model to calculate the corresponding model probability when the target vehicle is in this specific scenario. This provides the corresponding weight coefficients for calculating the matching probability of each road segment, thereby improving the accuracy of vehicle-to-road segment matching.

[0105] like Figure 3 If the matching probability k between the target vehicle and link3 at the current time has been calculated... 匹配-link3 And the matching probability k with link4 匹配-link4 Based on this, we can further combine the probability coefficient b3 of the target vehicle currently being in road segment link3 and the probability coefficient b4 of the target vehicle currently being in road segment link4 given by the pre-trained deep learning model corresponding to this scenario. Then, we can further determine the probability coefficient b4 of the target vehicle. 匹配-link3 Updated to b3*k 匹配-link3 , will k 匹配-link4 Updated to b4*k 匹配-link4 .

[0106] The training methods for deep learning models will be explained in detail below.

[0107] Figure 5 A flowchart illustrating a training method for a deep learning model used to predict the road where a vehicle is located, according to an embodiment of this specification, is shown.

[0108] like Figure 5 Step 501: Preprocess the data. In the original training data, due to the different sources and units of measurement for each feature dimension, the distribution range of feature values ​​varies greatly. When calculating the Euclidean distance between different samples, features with a large value range will play a dominant role. Therefore, it is necessary to preprocess the samples to normalize the features of each dimension to the same value range and eliminate the correlation between different features.

[0109] Step 502, Create a model. In the embodiments of this specification, since the method for determining the road segment where the vehicle is located is specifically a classification problem, a model for classification can be created in step 502. Specifically, a model based on logistic regression can be created for classification.

[0110] Step 503, Add layers to the model. Build the neural network by adding layers to the model.

[0111] Step 504: Select the optimizer and loss function. The optimizer is the loss function that updates the network weights during network training to optimize the model. Specifically, gradient-based deep learning optimizers can be used, such as SGD, Momentum, AdaGrad, Adam, Nesterov, or RMSprop. After the model design is complete, the optimal value of the model needs to be found through training configuration, i.e., the model's performance is measured by the loss function.

[0112] Step 505, Train the model. Input the training dataset for training data, and use the validation set data to validate the data. The training data input can include parameters such as slope, curvature, direction, and distance.

[0113] Step 506, Save the model. Save the trained model as a model file. When applying the model, simply load the model file and run it.

[0114] In the embodiments of this specification, map matching is performed based on high-precision maps. High-precision lane-level positioning information can be used as the initial input vehicle driving information, and lane connection information is also provided, fully reflecting the prior nature of the high-precision map. Based on this, an Hidden Markov Model (HMM) and the Viterbi algorithm are employed to more comprehensively consider the changing states of vehicle positioning points, making the determined vehicle positioning road segments closer to the actual driving conditions of the vehicles. The lane connection information and road segment connection information based on the high-precision map can further improve the calculation accuracy of the HMM model.

[0115] Furthermore, to address the anomaly of significant deviations between map road segment directions and actual road directions caused by mapping defects, a logistic regression method was specifically employed for learning. This method avoids mapping defects in road segment directions, thereby improving the accuracy of vehicle location determination in this scenario. Additionally, in practical applications, deep learning models are generally not required for weight calculations, resulting in low computational resource consumption. High-precision maps are a crucial component of autonomous driving; accurate map information and high-precision positioning are indispensable. Therefore, map matching provides correct map information, and providing accurate matching road segment information enables autonomous driving decisions to make more precise judgments, fully demonstrating the prior knowledge and accuracy of high-precision maps.

[0116] Based on the same idea, embodiments of this specification also provide apparatus corresponding to the above methods.

[0117] Figure 6 The embodiments provided in this specification correspond to Figure 1 A schematic diagram of a device for determining the road segment where a vehicle is located. (For example...) Figure 6 As shown, the device may include:

[0118] The information acquisition module 602 is used to acquire the vehicle driving information of the target vehicle at a first moment; the vehicle driving information includes at least one of latitude and longitude information and driving direction information.

[0119] The alternative matching road segment determination module 604 is used to determine the set of alternative matching road segments of the target vehicle in the map data at the first moment based on the vehicle driving information.

[0120] The matching degree information determination module 606 is used to determine the first matching degree information between the target vehicle and each road segment in the candidate matching road segment set at the first time.

[0121] The matching probability determination module 608 is used to determine the matching probability between the target vehicle and each road segment based on the first matching degree information between the target vehicle and each road segment at the first time.

[0122] The road segment determination module 610 is used to determine the road segment corresponding to the maximum matching probability as the road segment where the target vehicle is located.

[0123] It is understood that the modules mentioned above refer to computer programs or program segments used to perform one or more specific functions. Furthermore, the distinction between these modules does not imply that the actual program code must also be separate.

[0124] Based on the same idea, this specification also provides devices corresponding to the above methods in its embodiments.

[0125] In an optional embodiment of this specification, a fusion positioning module is provided, which may include, for example: Figure 6 The device shown is used to determine the location of a vehicle on a road segment. The device is used to determine the location of the vehicle on a road segment in a high-precision map, and to assist in cross-validation with data from other vehicle sensors to achieve high-precision fusion positioning and obtain the precise location of the vehicle. The other vehicle sensors include at least one of inertial navigation, GNSS / RTK, vision, and lidar.

[0126] Based on the same idea, embodiments of this specification also provide high-precision map engines corresponding to the above-mentioned methods, apparatus and devices.

[0127] Figure 7 This is a schematic diagram of the structure of a high-precision map engine provided in the embodiments of this specification.

[0128] like Figure 7 As shown, the high-precision map engine 700 may include:

[0129] Includes such as Figure 6 The device shown is a fusion positioning module 701 for determining the road segment where a vehicle is located;

[0130] The electronic horizon module 702 is used to receive external high-precision vehicle location information and match it to a map, providing a functional interface for autonomous driving applications to make control and judgments.

[0131] In addition, at least one of the following: autonomous driving design and operation domain judgment module 703, map update module 704, crowdsourcing preprocessing and feedback module 705, path intersection association module 706, and lane-level path planning module 707.

[0132] The autonomous driving design operation domain judgment module 703 is used to configure the autonomous driving area and judgment requirements.

[0133] The map update module 704 is used to obtain map data update information of a high-precision map based on the vehicle location and the planned route.

[0134] The crowdsourcing preprocessing and feedback module 705 is used to preprocess UGC visual vector data by filtering and fusion, and then feed it back to the cloud and update the map data center.

[0135] The path cross-association module 706 is used to synchronize the global path planning results initiated by the user to the autonomous driving system, and obtain the matching path of the navigation path on the high-precision map by cross-associating with the high-precision map.

[0136] The lane-level path planning module 707 is used to output lane-level local path planning within a certain length range in front of the vehicle based on the results of navigation path matching and route correction.

[0137] The foregoing has described specific embodiments of this specification. In some cases, the actions or steps described in the claims may be performed in a different order than those shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0138] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other.

[0139] The apparatus, devices and methods provided in the embodiments of this specification are corresponding. Therefore, the apparatus and devices also have similar beneficial technical effects as the corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the corresponding apparatus and devices will not be repeated here.

[0140] For ease of description, the above devices are described separately by function as various units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0141] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0142] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0143] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0144] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 Figure 1 The steps of the function specified in one or more boxes.

[0145] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0146] This application can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0147] The above description is merely an embodiment of this application and is not intended to limit the scope 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 the claims of this application.

Claims

1. A method of determining a road segment on which a vehicle is located, characterized by, The method comprises: obtaining vehicle travel information of a target vehicle at a first time; the vehicle travel information comprises at least one of latitude and longitude information and travel direction information; based on the vehicle travel information, determining a set of candidate matching road segments of the target vehicle in the map data at the first time; for each road segment in the set of candidate matching road segments, determining first matching degree information of the target vehicle and the each road segment at the first time; based on the first matching degree information of the target vehicle and the each road segment at the first time, determining a matching probability of the target vehicle and the each road segment; determining the road segment corresponding to the maximum value of the matching probability as the road segment where the target vehicle is located.

2. The method of claim 1, wherein the determining, for each road segment in the set of candidate matching road segments, the first matching degree information of the target vehicle and the each road segment at the first time comprises: calculating a transition probability and a measurement probability of the target vehicle, and determining the first matching degree information of the target vehicle and a target road segment in the set of candidate matching road segments based on the transition probability and the measurement probability.

3. The method of claim 2, wherein the calculating the transition probability and the measurement probability of the target vehicle, and determining the first matching degree information of the target vehicle and a target road segment in the set of candidate matching road segments based on the transition probability and the measurement probability comprises: for the target road segment in the set of candidate matching road segments, determining the transition probability based on the vehicle travel information of the target vehicle at the first time and the vehicle travel information of the target vehicle at a previous time; for the target road segment, determining the measurement probability based on the vehicle travel information of the target vehicle at the first time and road attribute information of the target road segment in the map data; determining the first matching degree information of the target vehicle and the target road segment based on the transition probability and the measurement probability.

4. The method of claim 2, wherein the vehicle travel information specifically comprises latitude and longitude information, and the calculating the transition probability of the target vehicle comprises: based on the latitude and longitude information of the target vehicle at the first time and the latitude and longitude information of the target vehicle at a previous time, calculating a position change amount of the target vehicle from the previous time to the first time; based on the position change amount, determining the transition probability of the target vehicle from the previous time to the first time; the transition probability is negatively correlated with the position change amount.

5. The method of claim 2, wherein the vehicle travel information specifically comprises latitude and longitude information and travel direction information, and the calculating the measurement probability of the target vehicle comprises: based on the latitude and longitude information of the target vehicle at the first time, calculating a distance between the target vehicle at the first time and the target road segment; based on the distance, determining a first probability related to the distance; the first probability is negatively correlated with the distance; calculating an angle difference value between a vehicle travel direction of the target vehicle at the first time and an extension direction of the target road segment; determine a second probability related to the angle difference based on the angle difference; the second probability is negatively related to the angle difference; calculate a measurement probability based on the first probability and the second probability; the measurement probability is positively related to the first probability, and the measurement probability is positively related to the second probability. 6.The method of claim 1, before the determining the matching probability of the target vehicle and each road segment based on the first matching degree information of the target vehicle and each road segment at the first time point, further comprising: determining whether a target road segment in the set of candidate matching road segments is an associated road segment of a road segment where the target vehicle was located at a previous time point, to obtain an associated road segment determination result; and the determining the matching probability of the target vehicle and each road segment based on the first matching degree information of the target vehicle and each road segment at the first time point comprises: if the associated road segment determination result indicates that the target road segment is the associated road segment of the road segment where the target vehicle was located at the previous time point, obtaining a preset associated road segment weight coefficient; and determining the matching probability of the target vehicle and the target road segment based on the first matching degree information of the target road segment and the associated road segment weight coefficient. 7.The method of claim 1, before the determining the matching probability of the target vehicle and each road segment based on the first matching degree information of the target vehicle and each road segment at the first time point, further comprising: determining whether a current scene where the target vehicle is located is a preset specific scene; if the current scene where the target vehicle is located is the preset specific scene, using a pre-trained deep learning model corresponding to the preset specific scene to calculate a model probability corresponding to each road segment in the set of candidate matching road segments; the deep learning model is trained based on vehicle real driving data in the preset specific scene, and the vehicle real driving data includes vehicle driving information and vehicle real driving road segment information corresponding to the data map; determining a model weight coefficient corresponding to each road segment according to the model probability corresponding to each road segment; and determining the matching probability of the target vehicle and each road segment based on the first matching degree information of a target road segment and the model weight coefficient. 8.The method of claim 1, the determining the matching probability of the target vehicle and each road segment based on the first matching degree information of the target vehicle and each road segment at the first time point comprises: obtaining second matching degree information of the target vehicle and each road segment at a previous time point; and determining the matching probability of the target vehicle and each road segment based on the first matching degree information and the second matching degree information. 9.The method of claim 1, the vehicle driving information includes latitude and longitude information; and the determining the set of candidate matching road segments of the target vehicle in the map data at the first time point based on the vehicle driving information comprises: determining a projection position of the target vehicle relative to each map road segment based on the latitude and longitude information of the target vehicle. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ If the projection position is located on the map road segment, a projection distance between the target vehicle and the projection position is calculated; If the projection distance is less than a preset distance threshold, the map road segment is determined as a road segment in the set of candidate matching road segments. 10.The method of claim 9, wherein the vehicle travel information further comprises travel direction information; and before the determining the map road segment as a road segment in the set of candidate matching road segments, the method further comprises: determining an angle difference between a travel direction of the target vehicle and an extension direction of each map road segment based on the travel direction information of the target vehicle; and wherein the determining the map road segment as a road segment in the set of candidate matching road segments comprises: if the angle difference is less than a preset angle threshold, determining the map road segment as a road segment in the set of candidate matching road segments. The apparatus comprises: an information obtaining module configured to obtain vehicle travel information of a target vehicle at a first time; the vehicle travel information comprises at least one of latitude and longitude information and travel direction information; a candidate matching road segment determining module configured to determine, based on the vehicle travel information, a set of candidate matching road segments of the target vehicle in map data at the first time; 11. A device for determining the road segment where a vehicle is located, characterized in that, a matching degree information determining module configured to determine, for each road segment in the set of candidate matching road segments, first matching degree information of the target vehicle and the each road segment at the first time; a matching probability determining module configured to determine, based on the first matching degree information of the target vehicle and the each road segment at the first time, a matching probability of the target vehicle and the each road segment; a road segment determining module configured to determine, as a road segment where the target vehicle is located, a road segment corresponding to the maximum value of the matching probability. 12.A fusion positioning module comprising the apparatus for determining a road segment where a vehicle is located according to claim 11, wherein the apparatus for determining a road segment where a vehicle is located is configured to determine a road segment position of the vehicle in a high-precision map, to assist in cross-verification with other sensor data of the vehicle, to achieve high-precision fusion positioning, and to obtain an accurate position of the vehicle; the other sensor data of the vehicle comprises at least one of inertial navigation, GNSS / RTK, vision, and laser radar. comprises: the fusion positioning module according to claim 12; an electronic horizon module configured to receive external high-precision vehicle position information and match the external high-precision vehicle position information to a map, and to provide a functional interface for an automatic driving application to make a control judgment; 13. A high-precision map engine, characterized by, and at least one of an automatic driving design and operation domain judgment module, a map updating module, a crowdsourcing preprocessing and feedback module, a path intersection correlation module, and a lane-level path planning module; wherein the automatic driving design and operation domain judgment module is configured to configure an automatic driving area and to make a judgment requirement; the map updating module is configured to obtain map data updating information of a high-precision map based on a vehicle position and a planned path; the crowdsourcing preprocessing and feedback module is configured to filter and fuse UGC visual vector data for preprocessing, to feed back to a cloud and to update a map data center; ​ ​ ​ The path intersection association module is configured to synchronize a global path planning result initiated by a user to an automatic driving system, cross-associate the global path planning result with a high-precision map, and obtain a matching path of the navigation path on the high-precision map. The lane-level path planning module is configured to output a lane level and a local path planning in a certain length range in front of the vehicle according to a result of navigation path matching and route correction.

Images