A vehicle positioning method, device, electronic equipment and computer storage medium

Abstract

The application discloses a vehicle positioning method and device, electronic equipment and computer storage medium. The vehicle positioning method comprises the following steps: performing road perception processing on a collected target image to obtain a road center point in a vehicle coordinate system; performing projection processing on the road center point according to a vehicle trajectory to obtain current local road information of the vehicle; performing matching processing on the current local road information and each local road information in a positioning area of the vehicle; and determining a road center point in target local road information matched with the current local road information in the positioning area as a target position of the vehicle. In this way, the positioning accuracy of vehicle position information can be improved.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to a vehicle positioning method. Background Technology

[0002] Vehicle positioning technology is an important technology for intelligent vehicles to obtain location information, and obtaining accurate vehicle location information is a prerequisite for realizing autonomous driving.

[0003] Currently, vehicle positioning mainly relies on the Global Navigation Satellite System (GNSS) to obtain real-time vehicle location information. However, GNSS systems suffer from low positioning accuracy and instability due to signal blockage and reflection during the positioning process, resulting in low accuracy of vehicle location information. Summary of the Invention

[0004] The main technical problem addressed by this application is to provide a vehicle positioning method, device, electronic equipment, and computer storage medium that can improve the positioning accuracy of vehicle location information.

[0005] To address the aforementioned technical problems, this application provides a vehicle positioning method, comprising: performing road perception processing on a acquired target image to obtain a road center point in a vehicle coordinate system; projecting the road center point based on the acquired vehicle trajectory to obtain current local road information of the vehicle; matching the current local road information with local road information within the vehicle's positioning area; and determining the road center point in the target local road information within the positioning area that matches the current local road information as the vehicle's target location.

[0006] To solve the above-mentioned technical problems, another technical solution adopted in this application is: to provide a vehicle positioning device, including: a perception module, used to perform road perception processing on the acquired target image to obtain the road center point in the vehicle coordinate system; a projection module, used to project the road center point according to the acquired vehicle trajectory to obtain the current local road information of the vehicle; and a matching module, used to match the current local road information with the local road information in the positioning area of ​​the vehicle.

[0007] To solve the above-mentioned technical problems, another technical solution adopted in this application is: to provide an electronic device, including a memory and a processor, wherein the memory stores program instructions, and the processor retrieves the program instructions from the memory to execute the above-mentioned vehicle positioning method.

[0008] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide a computer-readable storage medium including program data, which is used to implement the above-mentioned vehicle positioning method when executed by a processor.

[0009] The beneficial effects of this application are as follows: Road perception processing is performed on the acquired target image to obtain the road center point in the vehicle coordinate system; the road center point is projected based on the acquired vehicle trajectory to obtain the vehicle's current local road information; the current local road information is matched with the local road information within the vehicle's positioning area to obtain target road information that matches the current local road information; and the road center point in the target road information is determined as the vehicle's target location. Therefore, by combining the vehicle trajectory and the road center point to generate the current local road information, and matching the current local road information with the local road information within the vehicle's positioning area, more accurate vehicle location information is obtained, thus improving the accuracy of vehicle positioning. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein:

[0011] Figure 1 This is a schematic flowchart of an exemplary embodiment of the vehicle positioning method shown in this application;

[0012] Figure 2 yes Figure 1 A flowchart illustrating an exemplary embodiment of step S110 in the vehicle positioning method is shown.

[0013] Figure 3 yes Figure 1 A flowchart illustrating an exemplary embodiment of step S120 in the vehicle positioning method is shown.

[0014] Figure 4 yes Figure 1 A flowchart illustrating an exemplary embodiment of step S130 in the vehicle positioning method is shown.

[0015] Figure 5 yes Figure 1 A flowchart illustrating an exemplary embodiment of the vehicle positioning method prior to step S120 is shown.

[0016] Figure 6 yes Figure 5A flowchart illustrating an exemplary embodiment of the vehicle positioning method after step S530 is shown.

[0017] Figure 7 This is a flowchart illustrating another exemplary embodiment of the vehicle positioning method shown in this application;

[0018] Figure 8 This is a schematic diagram of an embodiment of the vehicle positioning device shown in this application;

[0019] Figure 9 This is a schematic diagram of the structure of an embodiment of the vehicle positioning method electronic device provided in this application;

[0020] Figure 10 This is a schematic diagram of the structure of an embodiment of the computer-readable storage medium of this application. Detailed Implementation

[0021] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It is understood that the specific embodiments described herein are only for explaining this application and not for limiting it. Furthermore, it should be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all structures. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0022] See Figure 1 , Figure 1 This is a flowchart illustrating an exemplary embodiment of the vehicle positioning method shown in this application. Specifically, it may include the following steps:

[0023] S110: Perform road perception processing on the acquired target image to obtain the road center point in the vehicle coordinate system.

[0024] The target image can be an image of the vehicle's surroundings acquired while the vehicle is in motion. Alternatively, it can be a high-resolution image from among multiple images acquired during the vehicle's journey. For example, the target image can be acquired using a sensor, such as a camera.

[0025] A vehicle coordinate system can be a coordinate system established with the vehicle as a reference point. For example, a vehicle coordinate system can be a three-dimensional coordinate system, consisting of three mutually perpendicular coordinate axes: the x-axis, y-axis, and z-axis, used to describe the position of points in space. As an example, a vehicle coordinate system can be established with the center of the vehicle's rear axle as the origin, the x-axis pointing forward of the vehicle, the y-axis pointing to the left of the vehicle, and the z-axis pointing upwards perpendicular to the vehicle's driving plane.

[0026] The road center point refers to the geometric center of a road. For example, the road center point can be defined as the distance between two adjacent lane lines, which is half the width of the lanes. For example, the number of road center points in a single target image is determined by the number of lane lines and the length of the road in the target image.

[0027] The vehicle positioning device collects target images during vehicle movement, performs road perception processing on the target images, and obtains the road center point in the vehicle coordinate system.

[0028] S120: Project the center point of the road based on the acquired vehicle trajectory to obtain the current local road information of the vehicle.

[0029] A vehicle trajectory, also known as a dead reckoning (DR) trajectory, refers to the path formed by a vehicle while traveling on a road. For example, the vehicle trajectory can be calculated using real-time vehicle speed information sensed by vehicle sensors. These sensors can be IMU sensors or wheel speed sensors to acquire vehicle speed information. Vehicle speed information can include vehicle acceleration, angular velocity, and body speed. For example, the vehicle's real-time motion state can be calculated using a filtering algorithm to obtain the vehicle trajectory. The filtering algorithm could be an error state Kalman filter.

[0030] Current local road information can be road-related data for a specific area within which a vehicle is traveling at the current moment. For example, it could be road information contained in the road traveled by the vehicle from a moment before the current moment up to the current moment. Current local road information can be obtained based on the vehicle's trajectory and the road center point. For example, the current local road information at time k can be represented as R. k .

[0031] The vehicle positioning device calculates the vehicle's trajectory based on the vehicle speed information obtained by sensors, and projects the road center point according to the vehicle's trajectory to obtain the vehicle's current local road information.

[0032] S130: Match the current local road information with the local road information within the vehicle's location area.

[0033] The location area refers to the approximate area where a vehicle is located, obtained using satellite positioning.

[0034] Local road information within a positioning area refers to road information for a segment of road within the approximate location area of ​​the vehicle. For example, the positioning area determined by satellite positioning can be determined on a map, and the road information within that positioning area can be used as local road information. The map can be a standard navigation map. For example, the j-th local road information within the positioning area can be represented as...

[0035] The matching process can involve calculating the similarity between road information contained in the current local road information and road information contained in other local road information. For example, the calculation method can include HMM-based algorithms, particle filter algorithms, and optimization algorithms.

[0036] The vehicle positioning device obtains the vehicle's positioning area, divides the roads in the positioning area into multiple local roads, obtains the local road information of each local road through a map, and calculates the road similarity between the current local road information and the local road information in the vehicle's positioning area.

[0037] S140: The center point of the target local road information that matches the local road information within the positioning area with the current local road information is determined as the target location of the vehicle.

[0038] Target local road information refers to the road where the vehicle is currently located on the map. For example, this can be achieved by matching the local road information within the positioning area with the current local road information, and then using the local road information within the positioning area that matches the current local road information as the target local road information. The target local road information can be the one with the smallest difference between the local road information within the positioning area and the current local road information.

[0039] The target location refers to the vehicle's specific position on the map at the current moment. For example, the center point of the road in the target local road information can be determined as the vehicle's target location at the current moment.

[0040] The vehicle positioning device matches the local road information in the positioning area with the current local road information, calculates the road similarity between the local road information in the positioning area and the current local road information, takes the local road information in the positioning area with a road similarity greater than the similarity threshold as the target local road information, and takes the center point of the road in the target local road information as the target position of the vehicle.

[0041] As can be seen, the vehicle positioning method in this embodiment performs road perception processing on the acquired target image to obtain the road center point in the vehicle coordinate system; projects the road center point based on the acquired vehicle trajectory to obtain the vehicle's current local road information; matches the current local road information with the local road information within the vehicle's positioning area to obtain target road information that matches the current local road information; and determines the road center point in the target road information as the vehicle's target location. Thus, by combining the vehicle trajectory and the road center point to generate the current local road information, and matching the current local road information with the local road information within the vehicle's positioning area, more accurate vehicle location information is obtained, improving the accuracy of vehicle positioning.

[0042] Based on the above embodiments, the embodiments of this application adopt... Figure 2 The flowchart details how to obtain the road center point in the vehicle coordinate system. Please refer to [link / reference]. Figure 2 , Figure 2 yes Figure 1 The illustrated flowchart shows an exemplary embodiment of step S110 in the vehicle positioning method. Specifically, step S110, which involves performing road perception processing on the acquired target image to obtain the road center point in the vehicle coordinate system, includes the following steps:

[0043] S210: Determine the road center point in the target image based on the feature similarity between the target features of each detected target in the target image and the preset road center point features.

[0044] Detected targets refer to specific objects identified and located in a target image. For example, a specific target could be a pedestrian, vehicle, lane line, road center point, etc. Exemplarily, a target detection model can be used to detect targets in a target image, obtaining the various detected targets. This target detection model can be a convolutional neural network.

[0045] Target features refer to the identifiable and distinguishable attributes or characteristics used to describe the detected targets. For example, target features of each detected target can be obtained during the target image detection process. For instance, when performing target detection using a convolutional neural network, the convolutional layers in the convolutional neural network extract features from the target image by performing convolution operations on the target image using a set of convolutional kernels, thereby obtaining the target features of each detected target.

[0046] Predefined road center point features refer to a set of features defined in advance to describe the center point of a road. For example, the preset road center point features can be obtained by extracting the features of the center points of a portion of the roads in the current map. In other embodiments, the preset road center point features can be obtained by extracting the center point features of roads from images containing roads.

[0047] Feature similarity measures the degree of similarity between a target feature and a preset road center point feature. For example, the feature similarity between a target feature and a preset road center point feature can be obtained using a feature similarity measurement method. For example, when the feature similarity is greater than a preset threshold, the target feature can be considered to correspond to the road center point.

[0048] The vehicle positioning device performs target detection on the target image, obtains the target features of each detected target in the target image, calculates the feature similarity between the target features of each detected target and the features of the preset road center point, determines the degree of similarity between each detected target and the road center point based on the feature similarity, and thus determines the road center point in the target image.

[0049] S220: Determine the coordinate information of the lane line in the vehicle coordinate system based on the pixel coordinates of the road center point in the target image.

[0050] Pixel coordinates refer to the coordinates of a location in a target image, expressed in pixels. For example, the center point of a road in the target image can be represented by pixel coordinates. For example, the pixel coordinates of the i-th road center point in the target image can be represented as... Among them, u i v represents the horizontal coordinates of the i-th road center point on the target image. i This represents the vertical coordinates of the i-th road center point on the target image, where 1 represents the constant term in the homogeneous coordinates.

[0051] Lane lines are markings on a road used to indicate the direction of vehicle travel and to define lanes. For example, lane lines can be markings symmetrically placed on either side of the center point of the road in a target image.

[0052] Coordinate information refers to the numerical values ​​on the coordinate axes used to represent the position of an object or point in a vehicle coordinate system. For example, the coordinate information of a lane line in the vehicle coordinate system can be obtained from the pixel coordinates of the road center point. Specifically, coordinate information includes the three-dimensional point coordinates of the lane line in the vehicle coordinate system. For example, the lane line coordinate information corresponding to the i-th road center point in the target image can be represented as... in, This represents the x-axis coordinate of the lane line in the vehicle coordinate system. This represents the lane line's coordinates on the y-axis in the vehicle coordinate system, -hb This indicates the coordinates of the lane line on the z-axis in the vehicle coordinate system.

[0053] The vehicle positioning device acquires the pixel coordinates of the road center point in the target image, and calculates the coordinate information of the lane line corresponding to the road center point in the vehicle coordinate system based on the acquired pixel coordinates of the road center point in the target image. Specifically, it calculates the product between the transpose of the pixel coordinates of the road center point in the target image and preset camera intrinsic parameters, camera pixel scale, and a first preset camera extrinsic parameter; the sum of the product and a second preset camera extrinsic parameter is used as the three-dimensional point coordinates of the lane line in the vehicle coordinate system. The pixel scale refers to the physical space occupied by one pixel in the target image.

[0054] For example, the coordinate information of the lane line in the vehicle coordinate system calculated based on the pixel coordinates of the road center point satisfies the following formula:

[0055]

[0056] in, λ represents the lane coordinate information corresponding to the i-th road center point in the target image. i R represents the pixel scale. bc K represents the first preset camera extrinsic parameters, and K represents the camera's intrinsic parameters. Let t represent the pixel coordinates of the i-th road center point in the target image. bc This indicates the second preset camera extrinsic parameters.

[0057] S230: Determine the center point of the road in the vehicle coordinate system based on the coordinate information of the lane lines in the vehicle coordinate system.

[0058] The road center point in the vehicle coordinate system can be determined by the vehicle positioning device based on the coordinate information of the lane lines in the vehicle coordinate system. For example, the road center point in the vehicle coordinate system can be calculated using the IPM algorithm. For example, for a single target image, road perception processing is performed on the target image to obtain all road center points in the target image, and the road center point closest to the vehicle's location on each road is selected as the road center point of that road. It should be noted that the single target image can be any of the acquired target images. For example, the road center points include all road center points obtained from the road perception processing of the target image. All roads include the road the vehicle is currently traveling on and its branch roads.

[0059] As can be seen, the vehicle localization method in this application determines the road center point in the target image based on the feature similarity between the target features of each detected target in the target image and the features of the preset road center point; it determines the coordinate information of the lane lines in the vehicle coordinate system based on the pixel coordinates of the road center point in the target image; and it determines the road center point in the vehicle coordinate system based on the coordinate information of the lane lines in the vehicle coordinate system. This allows for the extraction of the road center points of all roads in the target image, ensuring the integrity of the current local road information.

[0060] Based on the above embodiments, the embodiments of this application adopt... Figure 3 The flowchart details how the vehicle's current local road information is obtained; please refer to [link / reference]. Figure 3 , Figure 3 yes Figure 1 The illustrated flowchart shows an exemplary embodiment of step S120 in the vehicle positioning method. Specifically, step S120, which projects the road center point based on the acquired vehicle trajectory to obtain the vehicle's current local road information, includes the following steps:

[0061] S310: Project the road center point in the vehicle coordinate system at each time moment onto the vehicle trajectory at the corresponding time moment to obtain the projected vehicle trajectory. The vehicle coordinate system at each time moment is determined based on the actual position of the vehicle.

[0062] The vehicle coordinate system at different times refers to the vehicle's coordinate system at different moments during its journey. For example, when the vehicle turns, accelerates, or decelerates, the vehicle coordinate system also changes. For instance, when the vehicle turns, the vehicle coordinate system rotates accordingly.

[0063] The actual position of a vehicle refers to its location at a given moment. For example, the actual position of a vehicle may change depending on the road it is traveling on.

[0064] Projection refers to transforming a point in one coordinate system to obtain the position of that point in another coordinate system. For example, the road center point in the vehicle coordinate system at each moment can be transformed to obtain the vehicle's trajectory at the corresponding moment.

[0065] The vehicle positioning device determines the vehicle coordinate system at each time based on the actual position of the vehicle at different times, determines the road center point under the vehicle coordinate system at each time, and projects the road center point under the vehicle coordinate system at each time onto the vehicle trajectory at the corresponding time to obtain the projected vehicle trajectory.

[0066] S320: Project the road center point at other times in the projected vehicle trajectory onto the vehicle coordinate system where the road center point is located at the target time to obtain the road centerline.

[0067] The road center point at other times refers to the road center point at historical times prior to the current time. For example, historical times can be all times within a preset period. This preset period can be the time a vehicle takes to travel a certain distance. This certain distance could be 1 kilometer, 5 kilometers, 10 kilometers, etc.

[0068] The road center point at the target time refers to the road center point where the vehicle is located at the current time. For example, the road center point at the target time can be obtained based on the projected vehicle trajectory.

[0069] The road centerline is a trajectory formed by connecting the center points of the road. For example, to ensure that the road centerline is on the same reference plane, the road center points at other times are projected onto the vehicle coordinate system at the target time. The road centerline is obtained by connecting the road center points at other times and the road center points at the target time in the vehicle coordinate system at the target time.

[0070] For example, the calculation of projecting the road center point at other times into the vehicle coordinate system at the target time satisfies the following formula:

[0071]

[0072] in, T represents the position of the i-th road center point at time j in the vehicle coordinate system at time k. k T represents the position of the vehicle in the trajectory of the vehicle at time k. j This represents the position of the vehicle in the trajectory at time j. This represents the position of the i-th road center point in the vehicle coordinate system at time j.

[0073] The vehicle positioning device projects the road center point in the vehicle coordinate system at each time onto the vehicle trajectory at the corresponding time to obtain the projected vehicle trajectory. Then, it projects the road center points at other times onto the vehicle coordinate system of the road center point at the target time and connects the road center points at other times with the road center point at the target time to obtain the road centerline.

[0074] S330: Construct local road information based on the road centerline and the lane lines corresponding to the road center point at each time.

[0075] The vehicle positioning device obtains the road centerline and the corresponding lane lines based on the road center point at each time, and combines the road centerline and the corresponding lane lines to obtain the current local road information at the target time.

[0076] As can be seen, the vehicle positioning method in this application projects the road center point in the vehicle coordinate system at each time moment onto the vehicle trajectory at the corresponding time moment to obtain the projected vehicle trajectory. The vehicle coordinate system at each time moment is determined based on the actual position of the vehicle. The road center point in the projected vehicle trajectory at other times moment is projected onto the vehicle coordinate system where the road center point at the target time moment is located to obtain the road centerline. The current local road information is constructed based on the road centerline and the lane lines corresponding to the road center points at each time moment. Thus, combining the road center point and the lane lines can obtain more comprehensive road information, thereby generating more accurate current local road information.

[0077] Based on the above embodiments, the embodiments of this application adopt... Figure 4 The flowchart illustrates in detail how the matching process is performed; please refer to [link / reference]. Figure 4 , Figure 4 yes Figure 1 The illustrated flowchart shows an exemplary embodiment of step S130 in the vehicle positioning method. Specifically, step S130, which matches the current local road information with the local road information within the vehicle's positioning area, includes the following steps:

[0078] S410: Obtain the vehicle's location coordinates.

[0079] Location coordinates refer to the vehicle's position on a navigation map. For example, vehicle location coordinates can be obtained through a satellite positioning system, which can be a GNSS system.

[0080] The vehicle positioning device obtains the vehicle's location on the navigation map.

[0081] S420: Determine the positioning area based on a preset range centered on the vehicle's positioning coordinates.

[0082] The positioning area represents the region where the vehicle's actual location may be. For example, the vehicle's positioning area can be determined based on the operating status of the satellite positioning system. Here, operating status refers to the signal strength of the satellite positioning system. The satellite positioning system can be a GNSS system. For example, a corresponding confidence interval is set according to the operating status of the GNSS system, and the vehicle's positioning area is obtained with the vehicle's positioning coordinates as the center and the corresponding confidence interval as the preset range. For example, when the GNSS signal is weak, the confidence interval will be increased accordingly; when the GNSS signal is strong, the confidence interval will be decreased accordingly.

[0083] The vehicle positioning device sets a corresponding confidence interval based on the working status of the satellite positioning system, and uses the obtained vehicle positioning coordinates as the center to determine the positioning area where the vehicle's actual location may be located, in combination with the corresponding confidence interval.

[0084] S430: Extract local road information in the positioning area. Local road information includes actual road attributes and the center point of the actual local road.

[0085] Actual road attributes refer to the observable road information within the local road information of the location area. For example, actual road attributes can be obtained through a map. Actual road attributes may include roads, road scenes, number of road lanes, traffic signs, traffic lights, ground markings, road slope, road curvature, etc.

[0086] An actual local road refers to a segment of a road within the location area. For example, an actual local road can be any road from the various local road information sets. For example, the location area can be divided into various local road information sets, where the number of each set can be randomly generated. For example, the number could be 2, 3, 4, 5, etc.

[0087] The vehicle positioning device extracts the actual road attributes and the center point of the actual local road from the map based on the vehicle's positioning area.

[0088] S440: Match the current local road center point in the current local road information with the actual local road center point, and match the current road attribute in the current local road information with the actual road attribute.

[0089] Current road attributes refer to road information observable from the vehicle's actual location. For example, current road attributes can be acquired through sensors, such as cameras. Current road attributes can include the road itself, road scene, number of lanes, traffic signs, traffic lights, road markings, road slope, and road curvature. The road scene refers to the specific road shape. Road scenes can include open roads, closed roads, bridges, and tunnels. For example, current road attributes can be represented by the lowercase letter 's'. All observable road attributes are combined to form the current road attribute set. The current road attribute set can be represented by the uppercase letter 'S'.

[0090] In this embodiment, the matching process involves matching the current local road information with the local road information within the positioning area. This includes matching the center point of the current local road information with the center point of the actual local road, and matching the current road attributes in the current local road information with the actual road attributes. For example, the matching calculation method can be an Hidden Markov Model (HMM) algorithm, a particle filter algorithm, or an optimization algorithm.

[0091] For example, a particle filter algorithm is used to match the current local road information with the local road information within the positioning area. Specifically, the current local road information is transformed into a world coordinate system. For example, the current local road information can be transformed into a world coordinate system using the information from each local road.

[0092] For example, the current local road information, when transformed from the vehicle coordinate system to the world coordinate system, satisfies the following equation:

[0093]

[0094] in, This represents the i-th local road center point in the world coordinate system at time k. This represents the j-th local road information within the location area. This represents the i-th current local road center point in the vehicle coordinate system at time k.

[0095] Specifically, after converting the current local road information to the world coordinate system, the current road attributes in the current local road information are matched with the actual road attributes.

[0096] For example, the calculation for matching the current road attributes and actual road attributes in the current local road information satisfies the following formula:

[0097]

[0098] in, This represents the probability value of matching the current road attribute s of the current local road information at time k with the j-th local road information within the positioning area. Let a represent the i-th current local road center point in the vehicle coordinate system at time k. s b represents the state transition probability model coefficient corresponding to the current road attribute s. s This represents the observation probability model coefficients corresponding to the current road attribute s. This represents the standard deviation of the dynamic noise model corresponding to the current road attribute s. This represents the mean of the dynamic noise model corresponding to the current road attribute s. This represents the minimum distance obtained by matching the i-th current local road center point in the world coordinate system at time k with the actual local road center point.

[0099] After calculating the matching probability value of each current road attribute, the matching probability values ​​of each current road attribute are weighted to obtain the target matching probability value between the current local road information and each local road information.

[0100] For example, the target matching probability value is calculated according to the following formula:

[0101]

[0102] Among them, score j w represents the target matching probability value between the j-th local road information and the current local road information within the positioning area. s This represents the preset weights of the current road attributes. The probability value of matching the current road attribute s of the current local road information at time k with the j-th local road information within the positioning area is represented.

[0103] After the vehicle positioning device converts the current local road information to the world coordinate system, it matches the current local road center point with the actual local road center point, as well as the current road attributes in the current local road information with the actual road attributes, to obtain the target matching probability value between the current local road information and the local road information within the positioning area.

[0104] As can be seen, the vehicle positioning method in this embodiment obtains the vehicle's positioning coordinates; determines the positioning area within a preset range centered on the vehicle's positioning coordinates; extracts local road information within the positioning area, including actual road attributes and the center point of the actual local road; matches the current local road center point in the current local road information with the center point of the actual local road, and matches the current road attributes in the current local road information with the actual road attributes. By performing coarse positioning of the vehicle followed by map matching, the accuracy and efficiency of vehicle positioning can be improved; matching the road center point and road attributes increases the comprehensiveness of the matching, thereby improving the accuracy of vehicle positioning.

[0105] Based on the above embodiments, the embodiments of this application adopt... Figure 5 The flowchart details how to obtain the vehicle's trajectory; please refer to [link / reference]. Figure 5 , Figure 5 yes Figure 1 The illustrated flowchart shows an exemplary embodiment of the vehicle positioning method prior to step S120. Specifically, before step S120, which projects the road center point based on the acquired vehicle trajectory to obtain the vehicle's current local road information, the vehicle positioning method of this embodiment further includes the following steps:

[0106] S510: Based on the target vehicle position in the world coordinate system at the previous moment, the target vehicle speed at the previous moment, the vehicle's attitude information in the world coordinate system at the previous moment, the acceleration error at the previous moment, the angular velocity error at the previous moment, the acceleration at the previous moment, and the angular velocity at the previous moment, determine the initial vehicle position in the world coordinate system at the current moment, the initial vehicle speed at the current moment, the vehicle's attitude information in the world coordinate system at the current moment, the acceleration error at the current moment, the angular velocity error at the current moment, the acceleration at the current moment, and the angular velocity at the current moment.

[0107] The target vehicle position can be a recursive position of the vehicle. For example, the starting position of the vehicle can be determined as the target vehicle position, or the position of the vehicle at the previous moment can be determined as the target vehicle position. For example, the target vehicle position can be represented as p0.

[0108] The initial position can be the vehicle's position at the current moment. For example, the initial position can be obtained from the target vehicle's position. For example, the initial position can be represented as p.

[0109] The target vehicle speed can be a recursive vehicle speed. For example, the target vehicle speed can be the vehicle speed derived from the speed of the vehicle at the previous moment. For example, the target vehicle speed can be represented as v0. In another embodiment, the target vehicle speed can be the vehicle's initial speed. For example, the vehicle's initial speed can be obtained through a sensor. The sensor can be a vehicle wheel speed sensor.

[0110] The initial vehicle speed can be the vehicle's speed at the current moment. For example, the initial vehicle speed can be obtained recursively from the target vehicle speed. For example, the initial vehicle speed can be represented as v.

[0111] Attitude information can represent the vehicle's attitude in the world coordinate system. For example, the vehicle's attitude information in the world coordinate system at the previous moment can be represented as q0, and the vehicle's attitude information in the world coordinate system at the current moment can be represented as q.

[0112] Acceleration error refers to the non-zero value output by an accelerometer when the sensor is stationary. The accelerometer can be an IMU (Induction Mutual Detection Unit). For example, acceleration error can be expressed as α. b It should be noted that the acceleration error does not change with the movement of the vehicle; therefore, the acceleration error at the previous moment is the same as the acceleration error at the current moment.

[0113] Angular velocity error refers to the non-zero value output by an angular velocity sensor when it is stationary. The angular velocity sensor can be an IMU (Induction Mutual Detection and Reduction) sensor. For example, angular velocity error can be expressed as ω. bIt should be noted that the angular velocity error does not change with the movement of the vehicle; therefore, the angular velocity error at the previous moment is the same as the angular velocity error at the current moment.

[0114] Acceleration represents the rate and direction of change of a vehicle's velocity per unit time. For example, acceleration can be acquired using an accelerometer, which can be an IMU (Induction Unit) sensor. For example, acceleration can be expressed as α. m .

[0115] Angular velocity represents the speed and direction of a vehicle's rotation around a certain axis. For example, acceleration can be obtained using an angular velocity sensor, which can be an IMU sensor. For example, angular velocity can be expressed as ω. m .

[0116] For example, the initial position is calculated based on the target vehicle's position, target speed, the vehicle's attitude information in the world coordinate system at the previous moment, the acceleration and acceleration error at the previous moment, satisfying the following formula:

[0117]

[0118] Where p represents the initial position, p0 represents the target vehicle position, v0 represents the target vehicle speed, Δt represents the time interval between the previous and current moments, q0 represents the vehicle's attitude information in the world coordinate system at the previous moment, and α m α represents the acceleration at the previous moment. b denoted by , where g represents the acceleration error and g represents the gravitational acceleration.

[0119] For example, the initial vehicle speed at the current moment is calculated based on the target vehicle speed, the vehicle's attitude information in the world coordinate system at the previous moment, the acceleration and acceleration error at the previous moment, satisfying the following formula:

[0120] v←v0+(q0⊙(α m -α b )+g)Δt

[0121] Where v represents the initial vehicle speed, v0 represents the target vehicle speed, Δt represents the time interval between the previous moment and the current moment, q0 represents the vehicle's attitude information in the world coordinate system at the previous moment, and α m α represents the acceleration at the previous moment. b denoted by , where g represents the acceleration error and g represents the gravitational acceleration.

[0122] For example, the attitude information of the vehicle in the world coordinate system at the current moment is calculated based on the vehicle's attitude information in the world coordinate system at the previous moment, the angular velocity at the previous moment, and the angular velocity error at the previous moment, satisfying the following formula:

[0123]

[0124] Where q represents the vehicle's attitude information in the world coordinate system at the current moment, q0 represents the vehicle's attitude information in the world coordinate system at the previous moment, and ω m ω represents the angular velocity at the previous moment. b Δt represents the angular velocity error, and Δt represents the time interval between the previous moment and the current moment.

[0125] The vehicle positioning device calculates the target vehicle position, target speed, vehicle attitude information, acceleration error, angular velocity error, acceleration, and angular velocity in the world coordinate system at the current moment based on the target vehicle position, target speed, vehicle attitude information, acceleration error, angular velocity error, acceleration, and angular velocity acquired at the previous moment.

[0126] S520: The Jacobian of the vehicle state deviation is determined based on the vehicle's initial position in the world coordinate system at the current moment, the initial vehicle speed at the current moment, the vehicle's attitude information in the vehicle coordinate system at the current moment, the acceleration error at the current moment, and the angular velocity error at the current moment.

[0127] Vehicle state deviation refers to the difference between the actual state and the desired state of a vehicle. For example, vehicle state deviation can be obtained from calculations under ideal conditions and a noise term. The noise term can be predetermined.

[0128] The Jacobian refers to the first-order partial derivative matrix of the vehicle state deviation. For example, the Jacobian of the vehicle state deviation can be expressed as F... X Among them, F X Satisfy the following formula:

[0129]

[0130] Among them, F X Let I represent the vehicle state deviation, r represent the rotation matrix corresponding to the vehicle's attitude information in the world coordinate system at the previous moment, and α represent the vehicle state deviation. m α represents the acceleration at the previous moment. b Δt represents the acceleration error, and Δt represents the time interval between the previous moment and the current moment.

[0131] The vehicle positioning device obtains the Jacobian of the vehicle state deviation based on the vehicle's initial position in the world coordinate system at the current moment, the initial vehicle speed at the current moment, the vehicle's attitude information in the world coordinate system at the current moment, the acceleration error at the current moment, the angular velocity error at the current moment, and a preset noise term.

[0132] S530: Determine the target vehicle position at the current moment based on the Jacobian of the vehicle state deviation, the preset Jacobian, and the preset noise parameters.

[0133] The preset Jacobian refers to the partial derivative matrix of the perturbation vector. The perturbation vector can be a preset error value. For example, the preset Jacobian can be expressed as F... i Among them, F i Satisfy the following formula:

[0134]

[0135] Among them, F i I represents the predefined Jacobian, and I represents the identity matrix.

[0136] The preset noise parameter refers to the error in the sensor's output data. For example, the sensor may include an IMU sensor. For example, the error may originate from temperature variations, scaling factor errors, etc.

[0137] The vehicle state covariance is obtained by combining the Jacobian of the vehicle state deviation, the preset Jacobian, and the preset noise parameters. It should be noted that the vehicle state covariance changes with the vehicle's movement during the recursive process of the vehicle trajectory. For example, the vehicle state covariance at the current moment, calculated from the vehicle state covariance at the previous moment, satisfies the following formula:

[0138]

[0139] Where P represents the vehicle state covariance at the current moment, and F... X P0 represents the vehicle state deviation, and P0 represents the vehicle state covariance at the previous moment. F represents the transpose of the vehicle state deviation. i Indicates the presupposition of Jacobi, Q i Indicates the preset noise parameters. This indicates the transpose of the preset Jacobian.

[0140] The vehicle positioning device determines the vehicle state covariance based on the Jacobian of the vehicle state deviation, the preset Jacobian, and the preset noise parameters, and determines the target vehicle position in the actual state based on the vehicle state covariance.

[0141] As can be seen, the vehicle positioning method in this embodiment determines the initial vehicle position, initial vehicle speed, attitude information, acceleration error, angular velocity error, acceleration, and angular velocity in the world coordinate system at the current moment based on the acquired target vehicle position, target vehicle speed, attitude information, acceleration error, angular velocity error, acceleration, and angular velocity at the previous moment. It then determines the Jacobian of the vehicle state deviation based on the initial vehicle position, initial vehicle speed, attitude information, acceleration error, and angular velocity at the current moment. Finally, it determines the target vehicle position at the current moment based on the Jacobian of the vehicle state deviation, a preset Jacobian, and preset noise parameters. Combining real-time error status during the vehicle trajectory recursion process improves the accuracy of the vehicle trajectory.

[0142] Based on the above embodiments, the embodiments of this application adopt... Figure 6 The flowchart details how to adjust the vehicle's initial speed; please refer to [link / reference]. Figure 6 , Figure 6 yes Figure 5 The illustrated flowchart shows an exemplary embodiment following step S530 in the vehicle positioning method. Specifically, step S530, after determining the target vehicle position at the current moment based on the Jacobian of the vehicle state deviation, a preset Jacobian, and preset noise parameters, includes the following steps:

[0143] S610: Get the vehicle's latest speed.

[0144] The latest vehicle speed refers to the vehicle's speed at the current moment. For example, the latest vehicle speed can be obtained through sensors, such as wheel speed odometers. For example, the latest vehicle speed can be represented as v. c .

[0145] The vehicle positioning device obtains the vehicle's latest speed based on the vehicle's sensors.

[0146] S620: Calculate the observed Jacobian of vehicle speed based on the vehicle speed vector of the latest vehicle speed.

[0147] The vehicle speed vector refers to the latest vehicle speed in the vehicle coordinate system. For example, the vehicle speed vector of the latest vehicle speed in the vehicle coordinate system can be represented as V. m =[v c 0 0] T .

[0148] The vehicle speed observation Jacobian refers to the matrix used to update and correct the initial vehicle speed. For example, the vehicle speed observation Jacobian can be obtained from the vehicle speed vector. For example, the vehicle speed observation Jacobian can be represented as H, where H satisfies the following equation:

[0149] H = [0 3×3 R T [R T V m ] × J r 0 3×3 0 3×3 ]

[0150] Where H represents the vehicle speed observation Jacobian, 0 3×3 Let V represent a 3×3 zero matrix, R represent the rotation vector at the current time step, and V represent the zero matrix. m The velocity vector represents the latest vehicle speed, and the right Jacobian represents the rotation vector at the current moment.

[0151] The vehicle positioning device calculates the velocity vector of the vehicle's latest speed and obtains the vehicle speed observation Jacobian based on the velocity vector of the latest speed.

[0152] In another embodiment, the vehicle trajectory can be optimized by estimating the vehicle's trajectory using sensors. For example, the sensors may be visual odometry (VIO), laser odometry (LIO), or laser visual odometry (LVIO).

[0153] For example, estimating vehicle trajectory using Visual Odometry (VIO) specifically includes: extracting features from the acquired target image to obtain image feature points and feature point descriptors. For example, ORB and SUPERPOINT can be used for image feature point extraction. Feature points from different frames of the target image are then matched based on the feature point descriptors. A sliding window of preset length is set, and the vehicle trajectory is calculated within the preset length window. Specifically, the vehicle trajectory is optimized by minimizing the error in the recursive process of vehicle trajectory estimation and minimizing the reprojection error. Here, the reprojection error refers to the distance between the two-dimensional point obtained by projecting a three-dimensional point onto the image plane and the actually observed two-dimensional point.

[0154] For example, the minimized computation satisfies the following equation:

[0155]

[0156] Where T represents the position of the vehicle inside the sliding window. This represents the error in recursively estimating the vehicle's trajectory based on the IMU sensor and wheel speed odometer between two adjacent target images, where Log: This represents the mapping from a Lie group to a vector space. This represents the reprojection error of the k-th feature point in the i-th frame of the target image.

[0157] The vehicle positioning device adjusts the initial vehicle speed at the current moment based on the Jacobian observed by the vehicle speed. In addition, the vehicle positioning device can also estimate the vehicle trajectory based on sensors and corresponding algorithms, thereby obtaining an optimized vehicle trajectory.

[0158] As can be seen, the vehicle positioning method in this application obtains the latest vehicle speed; calculates the vehicle speed observation Jacobian based on the speed vector of the latest vehicle speed; and adjusts the initial vehicle speed at the current moment based on the vehicle speed observation Jacobian to obtain the target vehicle speed at the current moment. Therefore, by optimizing the derived initial speed using the vehicle's body speed obtained by sensors, a more accurate vehicle speed can be ensured. Furthermore, the vehicle trajectory can also be optimized using sensors and corresponding algorithms to obtain a more accurate vehicle trajectory.

[0159] To elaborate on the application of this application to the vehicle positioning method, Figure 7 The flowchart shown below provides further explanation, as detailed below:

[0160] The vehicle positioning device obtains the vehicle's previous state through sensors and recursively derives the current state from that state, thus obtaining the recursive trajectory, i.e., the vehicle's travel trajectory. It acquires target images through sensors such as cameras and performs target detection to obtain the pixel coordinates of the road center point. The pixel coordinates of the road center point are then calculated to obtain the road center point in the vehicle coordinate system. Road point attributes, i.e., the current local road attributes, are obtained through target image detection. A real-time local road map, i.e., the current local road information, is constructed by combining the vehicle's travel trajectory and the road center point. The vehicle's positioning area is obtained based on satellite navigation, and this area is divided into various local road information sections. The road center points and road attributes within the positioning area are obtained from the map, i.e., the map road skeleton. The current local road information is matched with the local road information within the positioning area to obtain the vehicle's target location.

[0161] Please see Figure 8 , Figure 8This is a schematic diagram of the structure of an embodiment of the vehicle positioning device of this application. The vehicle positioning device 80 includes a sensing module 810, a projection module 820, and a matching module 830; the sensing module 810 is used to perform road perception processing on the acquired target image to obtain the road center point in the vehicle coordinate system; the projection module 820 is used to project the road center point according to the acquired vehicle trajectory to obtain the current local road information of the vehicle; the matching module 830 is used to match the current local road information with the local road information in the vehicle's positioning area.

[0162] In the above-described scheme, the vehicle positioning device of this application performs road perception processing on the acquired target image to obtain the road center point in the vehicle coordinate system; it then projects the road center point based on the acquired vehicle trajectory to obtain the vehicle's current local road information; it matches the current local road information with the local road information within the vehicle's positioning area to obtain target road information that matches the current local road information, and determines the road center point in the target road information as the vehicle's target location. Thus, by combining the vehicle trajectory and the road center point to generate the current local road information, and matching the current local road information with the local road information within the vehicle's positioning area, a more accurate vehicle location information is obtained, improving the accuracy of vehicle positioning.

[0163] The functions of each module can be found in the vehicle positioning method implementation examples, and will not be repeated here.

[0164] To implement the vehicle positioning method of the above embodiments, this application proposes another electronic device, please refer to [link / reference needed]. Figure 9 , Figure 9 This is a schematic diagram of an embodiment of the vehicle positioning method electronic device provided in this application.

[0165] Electronic device 90 includes memory 910 and processor 920, wherein memory 910 and processor 920 are coupled together.

[0166] The memory 910 is used to store program data, and the processor 920 is used to execute the program data to implement the vehicle positioning method of the above embodiment.

[0167] In this embodiment, processor 920 can also be referred to as CPU (Central Processing Unit). Processor 920 may be an integrated circuit chip with signal processing capabilities. Processor 920 may also be a general-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component. The general-purpose processor may be a microprocessor, or processor 920 may be any conventional processor.

[0168] This application also provides a computer-readable storage medium 1000, such as... Figure 10 As shown, the computer-readable storage medium 1000 is used to store program data 1100, which, when executed by a processor, is used to implement the vehicle positioning method as described in the method embodiments of this application.

[0169] The methods involved in the vehicle positioning method embodiments of this application, when implemented as software functional units and sold or used as independent products, can be stored in a device, such as a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0170] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A vehicle positioning method, characterized in that, The method includes: The acquired target image is processed by road perception to obtain the road center point in the vehicle coordinate system; The road center point is projected based on the acquired vehicle trajectory to obtain the vehicle's current local road information. Specifically, the road center point in the vehicle coordinate system at each time moment is projected onto the corresponding vehicle trajectory to obtain the projected vehicle trajectory. The vehicle coordinate system at each time moment is determined based on the vehicle's actual position. The road center points at other times in the projected vehicle trajectory are then projected onto the vehicle coordinate system of the target time's road center point to obtain the road centerline. The current local road information is constructed based on the road centerline and the lane lines corresponding to the road center points at each time moment. The current local road information is matched with the local road information within the vehicle's positioning area. The center point of the road in the target local road information that matches the current local road information within the positioning area is determined as the target location of the vehicle; The step of matching the current local road information with the local road information within the vehicle's positioning area includes: Obtain the location coordinates of the vehicle; The positioning area is determined by a preset range centered on the vehicle's positioning coordinates; Extract local road information from the positioning area, where the local road information includes actual road attributes and the center point of the actual local road; The system matches the center point of the current local road in the current local road information with the center point of the actual local road, and matches the current road attributes in the current local road information with the actual road attributes. The current road attributes include one or more of the following: road, road scene, number of road lanes, traffic signs, traffic lights, ground markings, road slope, and road curvature.

2. The vehicle positioning method according to claim 1, characterized in that, The step of performing road perception processing on the acquired target image to obtain the road center point in the vehicle coordinate system includes: The road center point in the target image is determined based on the feature similarity between the target features of each detected target in the target image and the features of the preset road center point. The coordinate information of the lane lines in the vehicle coordinate system is determined based on the pixel coordinates of the road center point in the target image. The road center point in the vehicle coordinate system is determined based on the coordinate information of the lane line in the vehicle coordinate system.

3. The vehicle positioning method according to claim 2, characterized in that, The coordinate information includes the three-dimensional point coordinates of the lane line in the vehicle coordinate system. The step of determining the coordinate information of the lane line in the vehicle coordinate system based on the pixel coordinates of the road center point in the target image includes: Calculate the product of the pixel coordinate transpose of the road center point in the target image with the preset camera intrinsic parameters, the camera pixel scale, and the first preset camera extrinsic parameters; The sum of the product and the second preset camera extrinsic parameters is used as the three-dimensional point coordinates of the lane line in the vehicle coordinate system.

4. The vehicle positioning method according to claim 1, characterized in that, Before the step of projecting the road center point based on the acquired vehicle trajectory to obtain the vehicle's current local road information, the method further includes: Based on the target vehicle position in the world coordinate system at the previous moment, the target vehicle speed at the previous moment, the vehicle's attitude information in the world coordinate system at the previous moment, the acceleration error at the previous moment, the angular velocity error at the previous moment, the acceleration at the previous moment, and the angular velocity at the previous moment, the initial vehicle position, the initial vehicle speed, the attitude information, the acceleration error, the angular velocity error, the acceleration, and the angular velocity at the current moment are determined in the world coordinate system. The Jacobian of the vehicle state deviation is determined based on the vehicle's initial position in the world coordinate system at the current moment, the initial vehicle speed at the current moment, the vehicle's attitude information in the world coordinate system at the current moment, the acceleration error at the current moment, and the angular velocity error at the current moment. The target vehicle position of the vehicle at the current moment is determined based on the Jacobian of the vehicle state deviation, the preset Jacobian, and the preset noise parameters.

5. The vehicle positioning method according to claim 4, characterized in that, After the step of determining the target vehicle position of the vehicle at the current moment based on the Jacobian of the vehicle state deviation, a preset Jacobian, and preset noise parameters, the method further includes: Obtain the latest speed of the vehicle; The vehicle speed observation Jacobian is calculated based on the vehicle speed vector of the latest vehicle speed; The initial vehicle speed at the current moment is adjusted based on the observed Jacobian speed to obtain the target vehicle speed at the current moment.

6. A vehicle positioning device, characterized in that, include: The perception module is used to perform road perception processing on the acquired target image to obtain the road center point in the vehicle coordinate system. The projection module is used to project the road center point based on the acquired vehicle trajectory to obtain the current local road information of the vehicle. Specifically, the road center point in the vehicle coordinate system at each time moment is projected onto the vehicle trajectory at the corresponding time moment to obtain the projected vehicle trajectory. The vehicle coordinate system at each time moment is determined based on the actual position of the vehicle. The road center points at other times in the projected vehicle trajectory are projected onto the vehicle coordinate system where the road center point at the target time is located to obtain the road centerline. The current local road information is constructed based on the road centerline and the lane lines corresponding to the road center points at each time moment. A matching module is used to match the current local road information with local road information within the vehicle's positioning area; wherein, the positioning coordinates of the vehicle are obtained; a positioning area is determined with the vehicle's positioning coordinates as the center of a preset range; local road information within the positioning area is extracted, the local road information including actual road attributes and the center point of the actual local road; the center point of the current local road in the current local road information is matched with the center point of the actual local road; and the current road attributes in the current local road information are matched with the actual road attributes, the current road attributes including one or more of the following: road, road scene, number of road lanes, traffic signs, traffic lights, ground markings, road slope, and road curvature.

7. An electronic device, characterized in that, include: A memory and a processor, wherein the memory stores program instructions, and the processor retrieves the program instructions from the memory to execute the vehicle positioning method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, include: The system stores program data, which, when executed by a processor, is used to implement the vehicle positioning method as described in any one of claims 1-5.

Images