Lane-level positioning method, device, system and related equipment

Abstract

The present invention discloses a lane-level positioning method, apparatus, system, and related equipment. The method comprises: locating the current position of a vehicle, randomly generating a predetermined number of position points around the current position; determining the probability that each position point is the actual vehicle position, and determining the probability that each position point is the actual vehicle position based on multiple methods, such as electronic map errors, the lane line shape on both sides of the vehicle's lane as perceived by at least two sensing technologies, and the distance to the lane lines on both sides. The probabilities obtained by the multiple methods are then combined to obtain a final probability, and the vehicle's position is located based on the final probability. The present invention overcomes the problem of low accuracy when using a single positioning method, improves the positioning accuracy and robustness of vehicle positioning in different scenarios, achieves lane-level precise positioning, and can adapt to positioning requirements of different levels in different scenarios, reducing the impact of external factors on positioning accuracy.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and in particular to a lane-level positioning method, device, system and related equipment. Background Art

[0002] In intelligent transportation technology, accurate vehicle positioning is essential for achieving both autonomous driving and intelligent assisted driving. Currently, most vehicle positioning methods rely on global navigation satellite systems (GNSS) and inertial navigation systems (INS). These systems, on the one hand, lack the lane-level positioning accuracy required for intelligent transportation, and on the other hand, they are unable to locate the vehicle's position if they lose signal in certain environments.

[0003] The existing methods that can achieve lane-level positioning of vehicles include: positioning based on carrier phase differential technology, also known as Real-Time Kinematic (RTK) technology, positioning based on visual perception technology such as Mobileye visual hardware, and positioning based on Light Detection and Ranging (LIDAR) technology. However, these positioning methods are based on a single method to locate the lane in which the vehicle is located, and cannot achieve the positioning accuracy required for autonomous driving and intelligent assisted driving. Moreover, due to the use of a single positioning method, they cannot meet the positioning requirements of different levels and cannot adapt to the robustness of different scenarios. The accuracy and recall rate of main and auxiliary road recognition are also relatively low. Among them, the robustness refers to stability in different situations, such as different external conditions and weather conditions, and different scenarios such as tunnels, viaducts, canyons and cities.

[0004] Therefore, the positioning accuracy and adaptability of existing vehicle positioning technologies cannot meet the positioning requirements of autonomous driving and intelligent assisted driving. Therefore, how to accurately locate the vehicle position and achieve lane-level precise positioning has become a technical problem that needs to be solved urgently. Summary of the Invention

[0005] In view of the technical defects and technical drawbacks in the prior art, the embodiments of the present invention provide a lane-level positioning method, apparatus, system and related equipment that overcome the above problems or at least partially solve the above problems.

[0006] As one aspect of an embodiment of the present invention, a lane-level positioning method may include:

[0007] Locating a current position of the vehicle, and randomly generating a predetermined number of position points around the current position;

[0008] determining a first probability that each of the position points is located in each lane based on the position of the position point, an error of the electronic map, and a position range of each lane obtained from the map data;

[0009] Obtaining lane line shapes and distances to lane lines on both sides of the vehicle's lane as perceived by at least two sensing technologies, and matching these with lane line shapes and distances to lane lines on both sides of the lane at each position point obtained from map data, to determine a second probability that each position point is located in each lane;

[0010] A final probability that each of the position points is located in each lane is determined based on the first probability and the second probability, and the lane with the largest final probability is determined to be the lane where the vehicle is located.

[0011] As another aspect of an embodiment of the present invention, a lane-level positioning device may include:

[0012] Current location acquisition module, used to locate the current location of the vehicle;

[0013] a location point generating module, configured to randomly generate a predetermined number of location points around the current location;

[0014] a first probability determination module, configured to determine a first probability that each of the position points is located in each lane based on the position of the position point, an error of the electronic map, and a position range of each lane obtained from map data;

[0015] a second probability determination module configured to obtain lane line shapes and distances to lane lines on both sides of the lane where the vehicle is located, as sensed by at least two sensing technologies, and match the lane line shapes and distances to lane lines on both sides of the lane where each position point is located, as obtained from map data, to determine a second probability that each position point is located in each lane;

[0016] A lane determination module is configured to determine a final probability that each of the position points is located in each lane based on the first probability and the second probability, and determine the lane with the highest final probability as the lane where the vehicle is located.

[0017] As a third aspect of an embodiment of the present invention, a vehicle positioning system is provided, which may include: a server, a vehicle-mounted positioning terminal, and a sensing terminal;

[0018] The sensing terminal is used to sense the lane line shape and distance to the lane lines on both sides of the lane where the vehicle is located;

[0019] The vehicle-mounted positioning terminal is used to locate the current position of the vehicle;

[0020] The server includes the lane-level positioning device according to the second aspect.

[0021] As a fourth aspect of an embodiment of the present invention, it relates to a computer-readable storage medium having a computer program stored thereon, wherein the program implements the above-mentioned lane-level positioning method when executed by a processor.

[0022] As a fifth aspect of an embodiment of the present invention, it relates to a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the lane-level positioning method as described above is implemented.

[0023] The embodiments of the present invention achieve at least the following technical effects:

[0024] A predetermined number of points are randomly generated around the current location of the located vehicle. For each point, the probability of its being the actual vehicle location is determined. This probability is determined based on various methods, including electronic map errors, the line shape of the lane lines on both sides of the vehicle's lane as perceived by at least two sensing technologies, and the distance to the lane lines on both sides. The probabilities obtained from these multiple methods are then combined to produce a final probability, which is then used to locate the vehicle's position. This overcomes the low accuracy of positioning using a single method, improves vehicle positioning accuracy, achieves lane-level precision, and can adapt to different levels of positioning requirements in different scenarios. This reduces the impact of external factors on positioning accuracy, improves the robustness of vehicle positioning in different scenarios, and also enhances the accuracy and recall rate of primary and secondary road identification.

[0025] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structure described in the written description, claims, and drawings.

[0026] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0028] Figure 1 A flowchart of the lane-level positioning method provided by an embodiment of the present invention;

[0029] Figure 2 This is a diagram illustrating the principle of DR technology;

[0030] Figure 3 for Figure 1Specific flow chart of step S11 implementation;

[0031] Figure 4 This is a schematic diagram of an example of randomly generating position points in an embodiment of the present invention;

[0032] Figure 5 A specific flow chart for implementing step S13;

[0033] Figure 6 An example diagram of determining a first probability for a lateral position range of each position point;

[0034] Figure 7 This is a specific flow chart for implementing step S14 in the first way;

[0035] Figure 8 Schematic diagram of determining line type probability and distance probability in step S14;

[0036] Figure 9 This is a specific flow chart for implementing the second method of step S14;

[0037] Figure 10 A schematic diagram of the structure of a lane-level positioning device provided by an embodiment of the present invention;

[0038] Figure 11 A schematic diagram of the structure of a lane-level positioning system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0039] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0040] Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, like numbers in different figures represent like or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all possible embodiments consistent with the present invention. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present invention, as detailed in the appended claims.

[0041] This embodiment provides a lane-level positioning method, referring to Figure 1 As shown, the method may include the following steps:

[0042] Step S11: Locate the current position of the vehicle.

[0043] Perform preliminary positioning of the vehicle to obtain the current position of the vehicle. You can select two or more positioning technologies to perform preliminary positioning of the vehicle, and then fuse the preliminary positioning positions to obtain the current position of the vehicle.

[0044] Step S12: randomly generate a predetermined number of position points around the current position.

[0045] Based on a preset estimation algorithm, a preset number of location points are generated around the current location. These location points can be set with a Gaussian distribution. Each location point may be the location of the vehicle. The embodiment of the present invention locates the vehicle position by calculating the probability that the location point is located in each lane and the probability that the vehicle matches each location point, and then obtaining the highest probability value after fusion.

[0046] Step S13: Determine a first probability that each position point is located in each lane based on the position of the position point, the error of the electronic map, and the position range of each lane obtained from the map data.

[0047] This step is because there is a certain error in projecting the spatial coordinates onto the electronic map, so the position point generated by the vehicle's current position will have an error offset in each lane. This step is to calculate the probability of each position point being located in each lane.

[0048] Step S14: Obtain the lane line types and distances to the lane lines on both sides of the lane where the vehicle is located, as perceived by at least two sensing technologies, and match them with the lane line types and distances to the lane lines on both sides of the lane where each position point is located, obtained from the map data, to determine a second probability that each position point is located in each lane.

[0049] Because current perception technologies often have certain errors, the lane markings on both sides of the lane in which the vehicle is located may have some error. While the distance to the lane markings can be identified, it will not be the same for every location point. Matching involves a probability of matching. This step calculates the probability that each location point will match the vehicle in each lane.

[0050] There is no particular order in which step S13 and step S14 are executed. Step S13 may be executed first and then step S14, or step S14 may be executed first and then step S13, or both may be executed simultaneously. This is not specifically limited in the embodiment of the present invention.

[0051] Step S15: Determine the final probability that each position point is located in each lane based on the first probability and the second probability.

[0052] Step S16: Determine the lane with the highest final probability as the lane where the vehicle is located.

[0053] The embodiment of the present invention randomly generates a predetermined number of location points around the current location of the located vehicle, and determines the probability of each location point being the actual vehicle location. When determining the probability, the probability of each location point being the actual vehicle location is determined based on multiple methods, such as electronic map errors, the lane line shape on both sides of the vehicle's lane perceived by at least two sensing technologies, and the distance to the lane lines on both sides. The probabilities obtained by the multiple methods are then combined to obtain a final probability, and the vehicle's position is located based on the final probability. This overcomes the problem of low accuracy when using a single positioning method, improves the positioning accuracy of vehicle positioning, achieves lane-level precise positioning, and can adapt to different levels of positioning requirements in different scenarios. It reduces the impact of external factors on positioning accuracy, improves the robustness of vehicle positioning in different scenarios, and also improves the accuracy and recall rate of main and auxiliary road identification.

[0054] The detailed description of the above steps in the embodiment of the present invention is as follows:

[0055] In step S11, the vehicle's current position can be determined using a variety of methods, such as GPS navigation, Beidou navigation, or Dead Reckoning (DR) positioning, which is based on inertial motion measurement. To address the problem of these positioning methods failing to properly receive signals under certain circumstances, embodiments of the present invention utilize at least two positioning technologies to determine the vehicle's current position.

[0056] The specific implementation of this step is as follows: based on the at least two selected positioning technologies, at least two preliminary positioning positions of the vehicle are determined, and a preset position fusion method is used to fuse the at least two preliminary positioning positions to obtain the current position of the vehicle.

[0057] The embodiment of the present invention determines a preliminary positioning position of the vehicle through carrier phase differential technology (also known as real-time kinematic differential technology, Real-Time Kinematic, abbreviated as RTK), where RTK technology refers to a new and commonly used GPS measurement method. Before RTK technology, static, fast static, and dynamic measurements all required post-processing to obtain centimeter-level accuracy. RTK technology is a measurement method that can obtain centimeter-level positioning accuracy in real time in the field. It can provide the three-dimensional coordinates of the observation point in real time and achieve high centimeter-level accuracy. Similar to the pseudo-range differential principle, the base station transmits its carrier observations and station coordinate information to the user station in real time through the data link. The user station receives the carrier phase of the GPS satellite and the carrier phase from the base station, and composes the phase differential observation value for real-time processing, which can provide centimeter-level positioning results in real time.

[0058] The embodiments of the present invention use DR positioning technology to determine another preliminary vehicle location. Unlike the traditional positioning method of receiving navigation satellite data using a GPS module, DR positioning relies on measuring actual motion to calculate the target's position and trajectory. Therefore, it is less susceptible to environmental factors and can effectively compensate for positioning conditions in situations where GPS signal reception is poor or lost. However, as with most algorithms, due to the inherent errors of the positioning sensor, the positioning results will gradually diverge over time, ultimately resulting in inaccurate positioning.

[0059] Specifically, the vehicle's previous position can be determined using RTK technology or a GPS system as the vehicle's initial position. For example, when a user is driving a vehicle and the satellite signal is unstable, the satellite positioning signal measured just before the satellite signal becomes unstable can be used as the vehicle's initial position, recorded as (x0, y0).

[0060] The vehicle's direction and distance of travel from the initial position are then determined based on sensing data from an inertial sensor on the vehicle. The inertial sensor is a sensor that primarily detects and measures acceleration, tilt, shock, vibration, rotation, and multi-degree-of-freedom (DoF) motion, and is a crucial component for navigation, orientation, and motion control. For example, commonly used inertial sensors such as gyroscopes and accelerometers obtain data such as the vehicle's angular velocity and linear velocity, and then integrate the data to obtain the vehicle's position information.

[0061] Finally, the first current position is determined based on the initial position and the driving direction and distance. Figure 2 As shown, the above-mentioned inertial sensor can obtain the vehicle's travel distance S0, S1, S2, S3... and the angle value θ of the travel direction, the angle change value △θ0, △θ1, △θ2, △θ3...; because the vehicle's initial position (x0, y0) is known, the coordinates of the vehicle at each point (x1, y1), (x2, y2), etc. can be calculated by the following formula, and then the coordinates of the first current position are calculated as (x D ,y D ).

[0062] x1=x0+S0*sinθ0; x2=x1+S1*sinθ1

[0063] y1=y0+S0*cosθ0; y2=y1+S1*cosθ1

[0064] θ1=θ0+△θ0; θ2=θ1+△θ1

[0065]

[0066] θ k =θ k-1 +△θ k-1 k=1,2,3……

[0067] In the embodiment of the present invention, the preliminary positioning position of the absolute positioning obtained by the RTK technology and the DR technology is fused using a preselected position fusion method, for example, using the Kalman filter technology for fusion processing, referring to Figure 3 As shown, the specific steps may include:

[0068] Step S111: Determine the Kalman gain coefficient according to the variance of each preliminary positioning position.

[0069] Kalman filtering is an algorithm that uses linear system state equations and observation data from the system input and output to optimally estimate the system state. Because the observation data includes the effects of noise and interference in the system, the optimal estimate can also be considered a filtering process. Fusion refers to the use of Kalman filtering technology to combine two different data to obtain the most appropriate data. Kalman filtering uses the dynamic information of the target to remove the influence of noise and obtain an optimal estimate of the target position. This estimate can be an estimate of the current target position (filtering), an estimate of the future position (prediction), or an estimate of the past position (interpolation or smoothing).

[0070] For example, the coordinates of a preliminary positioning position of a vehicle determined by the above RTK positioning technology are (x R ,y R ), the coordinates of another preliminary positioning position of the vehicle determined by DR positioning technology are (x D ,y D ), and (x D ,y D ) and (x R ,y R ) The fused position coordinates are (x K ,y K Specifically, the Kalman gain coefficient can be determined based on the variance of the first preliminary positioning position and the variance of the second preliminary positioning position. The Kalman gain coefficient (also known as "empirical Kalman filtering") refers to the conditions under which the variance of sensor data is large and the conditions under which the variance is small, based on experience, so as to change the gain coefficient.

[0071] For example, the Kalman gain coefficient is recorded as K i , the Kalman gain coefficient can be obtained by the following formula: K i =(Pi-1 +Q) / (P i-1 +Q+R). Among them, P i-1 is the variance of the previous estimate, Q is the variance of the Gaussian noise, and R is the variance of the measured value, with both Q and R being constants. As can be seen from the above formula, K is determined by the ratio of the two variances. Since Q and R are both constants, the variance of the previous estimate plays a decisive role. If the variance is large after the previous estimate, it means the estimate is not very reliable. In this formula, it is easy to see that K will increase and become closer to 1. Combined with the previous formula, the proportion of the measured value will be larger, which means that the current estimate is more trustworthy than the measured value.

[0072] Step S112: Use the Kalman gain coefficient to assign weights to each preliminary positioning position.

[0073] Use the Kalman gain coefficient to assign weights to the first preliminary positioning position and the second preliminary positioning position to obtain the fusion: X i =(1-K i )X i-1 +K i *Z i , the variance of the fused position is P i =(1-K i )P i-1 Among them, the first preliminary positioning position is recorded as X i-1 , taking the first preliminary positioning position as the estimated value at the last moment, the second preliminary positioning position Z i , taking the second preliminary positioning position as the measurement value. From this formula, we can see that the current estimated value is determined by the last estimated value and this measurement value, and the weights of the two are determined by K i Decision, K i It is the Kalman gain coefficient.

[0074] Step S113: Perform weighted calculation on the preliminary positioning position based on the assigned weights to obtain the current position of the vehicle. In this step, the position coordinates of the vehicle are finally obtained after fusion as (x K ,y K ).

[0075] In the above-mentioned step S11 of the embodiment of the present invention, preliminary positioning is performed through at least two technologies, and then the at least two preliminary positioning positions are fused through Kalman filtering technology to obtain the absolute position of the vehicle, thereby avoiding the influence of the inability to receive the positioning signal due to weak signal, and further fusing multiple positioning methods to make the positioning result more in line with the actual situation and more accurate.

[0076] In the above step S12, a predetermined number of position points are randomly generated around the current position.

[0077] The location points are multiple noise points distributed with a Gaussian distribution within the target area when estimating using a Kalman filter based on the Bayesian estimation principle. These noise points are matched according to the target features. The preset number can be set based on actual needs; generally, 50 to 100 location points are sufficient for high-confidence positioning.

[0078] Specifically, this step may include: determining the variance of the current position according to the variance of the preliminary positioning position and the Kalman gain coefficient; and randomly generating a predetermined number of position points around the current position according to the variance of the current position.

[0079] The variance P of the fused position calculated in the above steps is i Then, based on the variance of the fused position, a predetermined number of position points are randomly generated around the fused position, thereby randomly generating these position points in a manner that conforms to the Gaussian distribution. For example, referring to Figure 4 As shown, a predetermined number of position points are randomly generated around the fused position.

[0080] In the above step S13, a first probability that each of the position points is located in each lane is determined based on the position of the position point, the error of the electronic map, and the position range of each lane obtained from the map data.

[0081] By randomly generating a preset number of location points, the coordinates of each location point can be obtained, and the width and position of each lane can be obtained based on the map data. Each time the vehicle is positioned, since there is a certain error in the electronic map, the probability of each location point being located in each lane can be determined based on the above data, which is the first probability.

[0082] Specifically, refer to Figure 5 As shown, the following steps may be included:

[0083] Step S131 : determining the horizontal position range of the location point on the electronic map according to the error of the electronic map and the location of the location point.

[0084] The lateral position range refers to the position range in the direction perpendicular to the lane's travel direction. The electronic map error refers to the error generated when each position point is projected from its actual spatial position onto the electronic map. The specific error can be determined based on actual conditions and historical experience. For example, on a common highway with a road width of 3.5 meters, the electronic map error can be set to 1 meter on each side.

[0085] Step S132: Determine a first probability that each position point is located in each lane in the electronic map based on the lateral position range and the position range of each lane.

[0086] For example, refer to Figure 6 As shown, the Gaussian distribution can be used to obtain the coordinates of a given point. To calculate the probability of a particular lane, only the horizontal projection coordinates of the point need to be determined, without considering coordinates in other directions. For example, if the point is 0.8m from the right edge of lane 2, and based on the road width of 3.5m and a preset error of 1m to the left and right, the coverage of lane 2 is 1.8m, and that of lane 3 is 0.2m. Based on the calculation, the coverage of lanes 1, 4, and 5 is 0m. Therefore, the probability of the point being in lanes 1, 4, and 5 is 0%. The probability of the point being in lane 2 is 1.8 / (1.8 + 0.2), meaning it has a 90% probability of being in lane 2. The probability of the point being in lane 3 is 0.2 / (1.8 + 0.2), meaning it has a 10% probability of being in lane 3. Based on this probability calculation principle, the probability of all points being in each lane can be determined.

[0087] In the above step S14, the lane line types and distances to the lane lines on both sides of the lane where the vehicle is located, as perceived by at least two sensing technologies, are obtained and matched with the lane line types and distances to the lane lines on both sides of the lane where each position point is located, obtained from the map data, to determine the second probability that each position point is located in each lane.

[0088] This step can obtain the second probability that the vehicle is matched to each position point in each lane in two ways. The lane line shape and distance to the lane lines on both sides of the lane where the vehicle is located can be obtained through different perception technologies. The line shape and distance perceived by different technical methods can be matched with the line shape and distance corresponding to the respective position points, and then the matching results can be fused. Alternatively, the line shape and distance perceived by different technologies can be fused first, and then matched with the line shape and distance corresponding to the position point.

[0089] <Method 1>, this method is to match first and then merge, refer to Figure 7 As shown, the following steps may be included:

[0090] Step S1411: Determine the lane line type and distance to the lane lines on both sides of the lane where the vehicle is located based on the captured road image and the collected point cloud data.

[0091] The lane marking type refers to the different types of lane center lines that divide lanes on a road. For example, lane marking types may include single solid line, single dashed line, double solid line, double dashed line, etc. The distance from the vehicle to the lane markings on both sides refers to the horizontal projection length of the vehicle from the left and right lane edges on both sides.

[0092] In an embodiment of the present invention, images of the road can be captured using visual perception technologies such as Mobileye visual hardware, and point cloud images of the road can also be captured using Lidar laser perception technology. The obtained images can be used to identify the lane line type and the distance to the lane lines on both sides of the lane where the vehicle is located. In an embodiment of the present invention, when Mobileye technology is used to identify the line type, the color of the lane line can also be identified, and the color of the lane line can be used as a matching parameter for registration to obtain the probability of matching the vehicle with the location point. For example, the color of the existing lane line can be white, yellow, etc.

[0093] Step S1412: Match the lane line type and distance to the lane lines on both sides of the vehicle's lane determined based on the road image with the lane line type and distance to the lane lines on both sides of the lane of each position point obtained from the map data to obtain a third probability that each position point is located in each lane.

[0094] Since the line type accuracy of the lane lines identified by Mobileye visual technology does not completely match the lane lines on the actual high-precision map, the accuracy of the identified lane lines is not certain. For example, the accuracy of the Mobileye visual output provided by the embodiments of the present invention is 95%. The distance to the lane lines identified by Mobileye visual technology is certain, but the distances between the lane lines on both sides of different locations are not the same. Therefore, the third probability mentioned above is a fusion of the two probabilities of line type probability and distance probability. Suppose that at this time, the line type of the lane where the vehicle is located is identified by Mobileye visual technology as a solid line on the left and a dotted line on the right, and the distance μ between the vehicle and the left lane line is identified as 1m.

[0095] The above line type probability can be calculated based on the line type of the lane lines on both sides of the lane where each position point is located and the distance to the two lane lines obtained from the map data. The line type of the lane lines on both sides of the lane where the position point is located is matched with the line type of the lane line where the vehicle is located using Mobileye visual technology to obtain the probability of whether the line type of the lane lines on both sides of the lane where the position point is located is the identified line type, and finally the line type probability of the position point in each lane is obtained. For example, single solid line and double solid line are not considered at this time, refer to Figure 8 As shown, the line type probability of the lane line type matching for position point A is 95%*95%, while the line type probability of the lane line type matching for position point B is 95%*5%.

[0096] The above distance probability can be calculated based on the difference between the distance from the position point obtained from the map data to the lane lines on both sides and the distance from the vehicle to the corresponding lane line identified by Mobileye visual technology, and the selected probability estimation calculation method is used to obtain the distance probability of each position point in each lane. Figure 8As shown in Figure 1, for example, based on the coordinates of the lane where the position points are located, it can be obtained that the distance between position point A and the left lane line is 2.8m, and the distance between position point B and the left lane line is 1.7m. The probability of the vehicle coinciding with position point A or position point B can be calculated according to formula (1).

[0097]

[0098] Where μ is the perceived distance between the vehicle and the left (or right) lane line; x is the distance from the position point to the left (or right) lane line; and σ is a constant related to the height and angle of the camera on the vehicle, which is generally an empirical value or evaluation value.

[0099] Step S1413: Match the lane line type and the distance to the lane lines on both sides of the lane where the vehicle is located, determined based on the point cloud data, with the lane line type and the distance to the lane lines on both sides of the lane where each position point is located, obtained from the map data, to obtain a fourth probability that each position point is located in each lane.

[0100] Calculating the fourth probability of each location point being located in each lane in this step is similar to step S1412 described above. However, unlike step S1412, the lane line accuracy and vehicle-to-lane distance identified using point cloud data differ from those obtained using road images. Therefore, the probability of each location point being located in each lane obtained from the collected point cloud data is not necessarily the same as the probability of each location point being located in each lane obtained from the captured road image.

[0101] Step S1414: Fusing the third probability and the fourth probability of each position point to obtain a second probability that each position point is located in each lane.

[0102] The third probability of matching the vehicle and the position point obtained in step S1412 and the fourth probability of matching the vehicle and the position point obtained in step S1413 are fused, for example, by multiplying the two probability values ​​obtained above to obtain a second probability that the vehicle is matched to each position point in each lane.

[0103] <Method 2> This method is to merge first and then match, refer to Figure 9 As shown, the following steps may be included:

[0104] Step S1421: Determine the lane line type and distance to the lane lines on both sides of the lane where the vehicle is located based on the captured road image and the collected point cloud data, and fuse the determined line types and distances.

[0105] This step is to fuse the lane line types and distances to the lane lines on both sides of the road where the vehicle is located, identified by different technologies, and determine the lane line types and distances to the lane lines on both sides of the road where the vehicle is located after the fusion of different technologies. For example, the probability that the lane line type of the vehicle in the road image captured by Mobileye vision technology is a solid line on the left and a dashed line on the right is 95%, while the probability that the lane line type of the vehicle is a solid line on the left and a dashed line on the right is determined by the point cloud data collected by Lidar technology is 90%. The above-determined probabilities can be fused, for example, after multiplication, the probability that the lane line type of the vehicle is a solid line on the left and a dashed line on the right is 85.5%. Similarly, the distance value after the fusion of the distances from the vehicle to the lane lines on both sides determined by the two different technologies can be calculated by taking the average of the distances or by a preset method.

[0106] This step integrates the lane line shapes and distances to the lane lines on both sides identified by different technologies, avoiding errors caused by a single algorithm and further improving the accuracy of positioning and recognition.

[0107] Step S1422: Match the lane line type and distance to the lane lines on both sides of the lane where the vehicle is located after fusion processing with the lane line type and distance to the lane lines on both sides of the lane where each position point is located obtained from the map data to obtain a second probability that each position point is located in each lane.

[0108] Specifically, this step can refer to Figure 9 As shown, the implementation steps are as follows:

[0109] Step S14221: Obtain the line type probability of the position point being located in each lane based on the line type of the lane lines on both sides of the lane where the position point is located obtained from the map data and the probability that the lane lines on both sides of the lane where the vehicle is located after fusion processing are of this line type.

[0110] Step S14222: Based on the difference between the distance from the position point to the lane lines on both sides obtained from the map data and the distance to the corresponding lane line after fusion processing, a selected probability estimation algorithm is used to obtain the distance probability of the position point being located in each lane.

[0111] Step S14223: The line type probability and the distance probability of each position point are respectively integrated to obtain a second probability that each position point is located in each lane.

[0112] In this step, the method for matching the lane line shape and the distance to the lane lines on both sides of the fused lane in the vehicle's lane with the lane line shape corresponding to each position point can be referred to the detailed description of step S1412 above, and will not be repeated here. It should be noted that the order of execution of steps S14221 and S14222 above is not specific, and step S14221 or step S14222 can be executed first, or they can be executed simultaneously. This embodiment of the present invention does not specifically limit this.

[0113] The embodiment of the present invention first fuses positioning obtained by dead reckoning technology and carrier phase difference technology based on Kalman filtering to obtain the vehicle's current position. Then, a predetermined number of position points are randomly generated based on the vehicle's current position. For each position point, the probability that it is the actual vehicle position is determined. When determining the probability, the probability of each position point being the actual vehicle position is determined based on multiple methods, such as electronic map error, the lane line shape of the lane in the vehicle's lane as perceived by at least two sensing technologies, and the distance to the lane line. The probabilities obtained by the multiple methods are then combined to obtain a final probability, and the vehicle's position is located based on the final probability. The above method not only overcomes the low accuracy problem of positioning using a single method, improves the positioning accuracy of the vehicle, achieves lane-level precision, and can adapt to different scenarios and different levels of positioning requirements, reducing the impact of external factors on positioning accuracy, while meeting different levels of positioning requirements and adapting to different scenarios. It also improves the accuracy and recall rate of main and auxiliary road identification.

[0114] Based on the same inventive concept, the embodiment of the present invention provides a vehicle-to-level positioning device, referring to Figure 10 As shown, the device may include: a current position acquisition module 11, a position point generation module 12, a first probability determination module 13, a second probability determination module 14 and a lane determination module 15, and its working principle is as follows:

[0115] The current position acquisition module 11 locates the current position of the vehicle. Specifically, the current position acquisition module 11 determines at least two preliminary positioning positions of the vehicle based on at least two selected positioning technologies, and fuses the at least two preliminary positioning positions using a preselected position fusion method to obtain the current position of the vehicle. More specifically, the current position acquisition module 11 fuses the at least two preliminary positioning positions using a preselected position fusion method to obtain the current position of the vehicle, including: the current position acquisition module 11 determines the Kalman gain coefficient based on the variance of each of the preliminary positioning positions; the current position acquisition module 11 uses the Kalman gain coefficient to assign weights to each of the preliminary positioning positions; the current position acquisition module 11 performs weighted calculation on the preliminary positioning positions based on the assigned weights to obtain the current position of the vehicle.

[0116] The location point generation module 12 randomly generates a predetermined number of location points around the current location. Specifically, the location point generation module 12 determines the variance of the current location based on the variance of the preliminary positioning location and the Kalman gain coefficient; and the location point generation module 12 randomly generates a predetermined number of location points around the current location based on the variance of the current location.

[0117] The first probability determination module 13 determines a first probability that each location point is located in each lane based on the location of the location point, the error in the electronic map, and the location range of each lane obtained from the map data. Specifically, the first probability determination module 13 determines a lateral location range of the location point on the electronic map based on the error in the electronic map and the location of the location point. The first probability determination module 13 determines a first probability that each location point is located in each lane on the electronic map based on the lateral location range and the location range of each lane.

[0118] The second probability determination module 14 obtains the lane line types and distances to the lane lines on both sides of the lane where the vehicle is located as perceived by at least two perception technologies, and matches them with the lane line types and distances to the lane lines on both sides of the lane where each position point is located obtained from the map data to determine the second probability that each position point is located in each lane.

[0119] Specifically, the second probability determination module 14 determines the lane line type and the distance to the lane lines on both sides of the lane where the vehicle is located based on the captured road image and the collected point cloud data; the second probability determination module 14 matches the lane line type and the distance to the lane lines on both sides of the lane where the vehicle is located determined based on the road image with the lane line type and the distance to the lane lines on both sides of the lane where each position point is located obtained from the map data to obtain a third probability that each position point is located in each lane; the second probability determination module 14 matches the lane line type and the distance to the lane lines on both sides of the lane where the vehicle is located determined based on the point cloud data with the lane line type and the distance to the lane lines on both sides of the lane where each position point is located obtained from the map data to obtain a fourth probability that each position point is located in each lane; the second probability determination module 14 fuses the third probability and the fourth probability of each position point to obtain a second probability that each position point is located in each lane.

[0120] Alternatively, specifically, the second probability determination module 14 determines the line type of the lane lines on both sides of the lane where the vehicle is located and the distance to the lane lines on both sides according to the captured road image and the collected point cloud data, and fuses the determined line types and distances; the second probability determination module 14 matches the line types of the lane lines on both sides of the lane where the vehicle is located and the distance to the lane lines on both sides after the fusion processing with the line types of the lane lines on both sides and the distance to the lane lines on both sides of the lane where each of the position points is located obtained from the map data, to obtain a second probability that each of the position points is located in each lane.

[0121] More specifically, the second probability determination module 14 obtains the line type probability of the position point being in each lane based on the line type of the lane lines on both sides of the lane where the position point is located obtained from the map data and the probability that the lane lines on both sides of the lane where the vehicle is located are of the line type after fusion processing; the second probability determination module 14 obtains the distance probability of the position point being in each lane based on the difference between the distance from the position point to the lane lines on both sides obtained from the map data and the distance to the corresponding side lane lines after fusion processing, using the selected probability estimation algorithm; the second probability determination module 14 respectively fuses the line type probability and the distance probability of each position point to obtain a second probability that each position point is located in each lane.

[0122] The lane determination module 15 determines a final probability that each position point is located in each lane according to the first probability and the second probability, and determines the lane with the largest final probability as the lane where the vehicle is located.

[0123] The technical effects and related examples of the device described in this embodiment can be found in the relevant content of the above method, which will not be repeated here.

[0124] Based on the same inventive concept, the embodiment of the present invention provides a lane-level positioning system, referring to Figure 11 As shown, the system may include: a server 1, a vehicle positioning terminal 2 and a sensing terminal 3;

[0125] The sensing terminal 3 is used to sense the lane line shape and distance to the lane lines on both sides of the lane where the vehicle is located, and may include a camera device 31 and a point cloud acquisition device 32;

[0126] The vehicle-mounted positioning terminal 2 is used to locate the current position of the vehicle;

[0127] The server 1 includes the lane-level positioning device as described above.

[0128] The technical effects and related examples of the system of this embodiment can be found in the relevant content of the above method embodiment, which will not be repeated here.

[0129] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage and optical storage, etc.) containing computer-usable program code.

[0130] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts 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, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0131] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0132] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0133] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A lane-level positioning method, comprising: Locating a current position of the vehicle, and randomly generating a predetermined number of position points around the current position; determining a lateral position range of the location point on the electronic map based on an error of the electronic map and the location of the location point; determining a first probability that each of the position points is located in each lane in the electronic map based on the lateral position range and the position range of each lane; Obtaining lane line shapes and distances to lane lines on both sides of the vehicle's lane as perceived by at least two sensing technologies, and matching these with lane line shapes and distances to lane lines on both sides of the lane at each position point obtained from map data, to determine a second probability that each position point is located in each lane; A final probability that each of the position points is located in each lane is determined based on the first probability and the second probability, and the lane with the largest final probability is determined to be the lane where the vehicle is located.

2. The method according to claim 1, wherein locating the current position of the vehicle comprises: At least two preliminary positioning positions of the vehicle are determined based on the at least two selected positioning technologies, and the at least two preliminary positioning positions are fused using a preselected position fusion method to obtain the current position of the vehicle.

3. The method according to claim 2, wherein fusing the at least two preliminary positioning positions using a preselected position fusion method to obtain the current position of the vehicle comprises: Determining a Kalman gain coefficient based on the variance of each of the preliminary positioning positions; The Kalman gain coefficient is used to assign a weight to each of the preliminary positioning positions, and a weighted calculation is performed on the preliminary positioning positions based on the assigned weight to obtain the current position of the vehicle.

4. The method according to claim 3, wherein randomly generating a predetermined number of location points around the current location comprises: Determining the variance of the current position according to the variance of the preliminary positioning position and the Kalman gain coefficient; A predetermined number of position points are randomly generated around the current position according to the variance of the current position.

5. The method according to claim 1, wherein obtaining the lane line shapes and distances to the lane lines on both sides of the vehicle's lane as perceived by at least two sensing technologies and matching them with the lane line shapes and distances to the lane lines on both sides of the lane of each position point obtained from map data to determine a second probability that each position point is located in each lane comprises: Determine the lane line shape and distance to the lane lines on both sides of the vehicle's lane based on the captured road image and the collected point cloud data; Matching the lane line types and distances to the lane lines on both sides of the vehicle's lane determined based on the road image with the lane line types and distances to the lane lines on both sides of the lane of each position point obtained from the map data to obtain a third probability that each position point is located in each lane; Matching the lane line type and distance to the lane lines on both sides of the lane where the vehicle is located, as determined based on the point cloud data, with the lane line type and distance to the lane lines on both sides of the lane where each position point is located, as obtained from the map data, to obtain a fourth probability that each position point is located in each lane; The third probability and the fourth probability of each position point are fused to obtain a second probability that each position point is located in each lane.

6. The method according to claim 1, wherein obtaining the lane line shapes and distances to the lane lines on both sides of the vehicle's lane as perceived by at least two sensing technologies and matching them with the lane line shapes and distances to the lane lines on both sides of the lane of each position point obtained from map data to determine a second probability that each position point is located in each lane comprises: Determine the lane line shape and distance to the lane lines on both sides of the vehicle's lane based on the captured road image and collected point cloud data, and fuse the determined line shape and distance; The lane line shape and distance to the lane lines on both sides of the lane where the vehicle is located after fusion processing are matched with the lane line shape and distance to the lane lines on both sides of the lane where each position point is located obtained from the map data to obtain a second probability that each position point is located in each lane.

7. The method according to claim 6, wherein the step of matching the lane line shape and distance to the lane lines on both sides of the lane where the vehicle is located after fusion processing with the lane line shape and distance to the lane lines on both sides of the lane where each position point is located obtained from map data to obtain a second probability that each position point is located in each lane comprises: Obtaining the line type probability of the position point being located in each lane based on the line types of the lane lines on both sides of the lane where the position point is located obtained from the map data and the probability that the lane lines on both sides of the lane where the vehicle is located are of the line type after fusion processing; Using a selected probability estimation algorithm, the distance probability of the position point being located in each lane is obtained based on the difference between the distance from the position point to the lane lines on both sides obtained from the map data and the distance to the corresponding lane line after fusion processing; The line type probability and the distance probability of each position point are fused separately to obtain a second probability that each position point is located in each lane.

8. A lane-level positioning device, comprising: Current location acquisition module, used to locate the current location of the vehicle; a location point generating module, configured to randomly generate a predetermined number of location points around the current location; a first probability determination module configured to determine a lateral position range of the position point on the electronic map based on an error in the electronic map and the position of the position point; and determine a first probability that each of the position points is located in each lane on the electronic map based on the lateral position range and the position range of each lane; a second probability determination module configured to obtain lane line shapes and distances to lane lines on both sides of the lane where the vehicle is located, as sensed by at least two sensing technologies, and match the lane line shapes and distances to lane lines on both sides of the lane where each position point is located, as obtained from map data, to determine a second probability that each position point is located in each lane; A lane determination module is configured to determine a final probability that each of the position points is located in each lane based on the first probability and the second probability, and determine the lane with the highest final probability as the lane where the vehicle is located.

9. A vehicle positioning system comprising: Server, vehicle positioning terminal and perception terminal; The sensing terminal is used to sense the lane line shape and distance to the lane lines on both sides of the lane where the vehicle is located; The vehicle-mounted positioning terminal is used to locate the current position of the vehicle; The server includes the lane-level positioning device according to claim 8.

10. A computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the lane-level positioning method according to any one of claims 1 to 7 is implemented.

11. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the lane-level positioning method according to any one of claims 1 to 7 when executing the program.

Images