Lane center line determination method, device and equipment and readable storage medium

Abstract

The invention discloses a lane center line determination method, device and equipment and a readable storage medium, which are applied to the field of intelligent traffic, and comprise the following steps: determining a first transverse coordinate point corresponding to a first longitudinal sampling point coordinate based on a target model, determining a transverse coordinate point weight corresponding to the first transverse coordinate point, and determining a first longitudinal sampling point coordinate; determining a coordinate point corresponding to the forward lane center line based on the transverse coordinate point weight of the target model and the first transverse coordinate point; performing filtering processing on a second transverse coordinate point determined based on the fourth track model and the second longitudinal sampling point coordinate to obtain a filtered second transverse coordinate point, and determining a coordinate point corresponding to the backward lane center line based on the second longitudinal sampling point coordinate and the filtered second transverse coordinate point, and determining a target lane center line based on the coordinate point corresponding to the forward lane center line and the coordinate point corresponding to the backward lane center line. The lane center line is calculated by fusing various models, so that the lane center line can be accurately determined without completely depending on a sensor.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and in particular to a lane centerline determination method, device, equipment and readable storage medium. Background Art

[0002] Currently, the calculation of lane centerline mostly relies on the road topology given by high-precision maps or the road edge given by visual perception to infer the road centerline. This method is highly dependent on sensors and has high requirements on the geometric characteristics of the road. On sensor-restricted or unstructured roads or tunnels, the sensor's scope is limited, which has a great impact on the assisted driving function and a poor intelligent driving experience. In addition, when the sensor cannot provide accurate information, the road centerline cannot be accurately determined.

[0003] It can be seen that how to accurately determine the lane centerline under sensor limitations is a technical problem that technicians in this field need to solve urgently. Summary of the invention

[0004] In view of this, an object of the present invention is to provide a lane centerline determination method, device, equipment and readable storage medium, which solve the technical problem of low lane centerline determination accuracy in the prior art.

[0005] In order to solve the above technical problems, the present invention provides a method for determining a lane centerline, comprising:

[0006] Determine to select at least one of the first trajectory model, the second trajectory model and the third trajectory model as the target model; wherein the first trajectory model is a first trajectory model determined based on the curvature of the current trajectory of the vehicle; the second trajectory model is a second trajectory model obtained by polynomial fitting based on the coordinates corresponding to the historical trajectory points of the target vehicle; the third trajectory model is a lane centerline model determined based on a visual sensor;

[0007] Determine a first transverse coordinate point corresponding to the coordinate of the first longitudinal sampling point based on the target model, determine a transverse coordinate point weight corresponding to the first transverse coordinate point, and determine a coordinate point corresponding to the center line of the forward lane based on the transverse coordinate point weight of the target model and the first transverse coordinate point;

[0008] Performing filtering processing on the second transverse coordinate point determined based on the fourth trajectory model and the coordinates of the second longitudinal sampling point to obtain a filtered second transverse coordinate point, and determining a coordinate point corresponding to the center line of the rear lane based on the coordinates of the second longitudinal sampling point and the filtered second transverse coordinate point; wherein the fourth trajectory model is a trajectory model obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the vehicle;

[0009] The target lane centerline is determined based on the coordinate point corresponding to the forward lane centerline and the coordinate point corresponding to the backward lane centerline.

[0010] Optionally, before determining to select at least one of the first trajectory model, the second trajectory model and the third trajectory model as the target model, the method further includes:

[0011] Constructing a quadratic function model based on the curvature of the vehicle trajectory to obtain the first dynamic trajectory model;

[0012] Performing filtering processing on the horizontal coordinate points of the historical trajectory points of each cycle of the target vehicle to obtain the filtered horizontal coordinate points corresponding to the target vehicle, and obtaining the historical trajectory points of the target vehicle corresponding to the target vehicle based on the filtered horizontal coordinate points;

[0013] Performing polynomial curve fitting based on the historical trajectory points of the target vehicle to obtain the second trajectory model;

[0014] The third trajectory model is determined based on the rate of change of the lane centerline curvature, the lane centerline curvature, the lane centerline heading angle, and the lane centerline offset distance determined by the visual sensor.

[0015] Optionally, before filtering the horizontal coordinate points of the historical trajectory points of each cycle of the target vehicle to obtain the filtered horizontal coordinate points corresponding to the target vehicle, and obtaining the target vehicle historical trajectory points corresponding to the target vehicle based on the filtered horizontal coordinate points, the method further includes:

[0016] Selecting a preset number of vehicles as the target vehicles based on a target vehicle selection rule; wherein the target vehicle selection rule is a rule determined based on the coordinate values ​​of the vehicles and the time that the vehicles are within the field of view of the self-vehicle;

[0017] The target vehicle selection rule includes:

[0018] The longitudinal coordinate value of the vehicle is within the set longitudinal coordinate value range;

[0019] The difference between the lateral coordinate value of the vehicle and the lateral coordinate value of the ego vehicle is within the set difference threshold;

[0020] The time that the vehicle is in the field of view of the vehicle is greater than the set time threshold;

[0021] The standard deviation of the lateral coordinate values ​​of the vehicle is less than the set standard deviation threshold.

[0022] Optionally, the determining to select at least one of the first trajectory model, the second trajectory model, and the third trajectory model as the target model includes:

[0023] determining to use the first trajectory model, the second trajectory model and the third trajectory model as the target models;

[0024] Accordingly, determining a first transverse coordinate point corresponding to the coordinate of the first longitudinal sampling point based on the target model, and determining a transverse coordinate point weight corresponding to the first transverse coordinate point, and determining a coordinate point corresponding to the center line of the forward lane based on the transverse coordinate point weight of the target model and the first transverse coordinate point, including:

[0025] Determine, based on the first trajectory model, the second trajectory model, and the third trajectory model, a first lateral coordinate point of the vehicle, a first lateral coordinate point of the target vehicle, and a first lateral coordinate point of the sensor corresponding to the coordinates of the first longitudinal sampling point, respectively;

[0026] Based on the first transverse coordinate point of the vehicle, the first transverse coordinate point of the target vehicle and the first transverse coordinate point of the sensor, an average value of the first transverse coordinate point corresponding to the coordinate of each first longitudinal sampling point is determined;

[0027] Based on the first lateral coordinate point of the vehicle, the first lateral coordinate point of the target vehicle, the first lateral coordinate point of the sensor and the average value of the first lateral coordinate point, determining the variance of the lateral coordinate points under three models corresponding to the coordinates of each first longitudinal sampling point;

[0028] Determine a weighted average value of the first transverse coordinate points corresponding to each first longitudinal sampling point based on the variance of the transverse coordinate points;

[0029] Determine the horizontal coordinate point weight corresponding to each first horizontal coordinate point based on the weighted average value;

[0030] Based on the weight of each of the transverse coordinate points, a target first transverse coordinate point corresponding to each first longitudinal sampling point is determined, and based on the first longitudinal sampling point and the target first transverse coordinate point, a coordinate point corresponding to the center line of the forward lane is obtained.

[0031] Optionally, before determining the first horizontal coordinate point corresponding to the first vertical sampling point coordinate based on the target model and determining the horizontal coordinate point weight corresponding to the first horizontal coordinate point, the method further includes:

[0032] Determine if the vision sensor can determine the end point of the lane centerline;

[0033] When the end point of the lane centerline can be determined, setting a maximum point of the coordinates of the first longitudinal sampling point based on the end point determined by the visual sensor;

[0034] When the end point of the lane centerline cannot be determined, the end point of the lane centerline is calculated by linear interpolation of the vehicle speed, and the maximum point of the coordinates of the first longitudinal sampling point is set based on the end point determined by linear interpolation of the vehicle speed.

[0035] Optionally, the process of determining the coordinates of the second longitudinal sampling point includes:

[0036] Determine the minimum point of the center line of the rear lane;

[0037] The coordinates of the second longitudinal sampling point are determined based on the minimum point.

[0038] Optionally, the process of determining the fourth trajectory model includes:

[0039] Get a set number of historical trajectory points of the vehicle;

[0040] Selecting a target vehicle historical trajectory point from the set number of vehicle historical trajectory points based on the vehicle historical trajectory point selection condition;

[0041] Performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the target vehicle to obtain the fourth trajectory model;

[0042] The conditions for selecting the historical trajectory points of the vehicle include:

[0043] The point whose horizontal coordinate of the trajectory point is greater than the set horizontal coordinate threshold;

[0044] The vertical coordinate at the current moment is greater than the vertical coordinate at the previous moment;

[0045] The vertical coordinate at the current moment is less than zero;

[0046] The longitudinal coordinate at the current moment is less than the minimum value of the longitudinal coordinate of the historical trajectory point of the vehicle;

[0047] The vertical coordinate at the current moment is not equal to the vertical coordinate at the previous moment;

[0048] The difference between the absolute values ​​of the horizontal coordinates at the previous moment and the current moment is greater than the set absolute value difference threshold.

[0049] Optionally, determining the target lane centerline based on the coordinate point corresponding to the centerline of the forward lane and the coordinate point corresponding to the centerline of the backward lane includes:

[0050] Determining a front lane centerline model based on the coordinate points corresponding to the front lane centerline;

[0051] Determine a rear lane centerline model based on the coordinate points corresponding to the rear lane centerline;

[0052] The target lane centerline is determined based on the forward lane centerline model and the backward lane centerline model.

[0053] The present application also provides a lane centerline determination device, comprising:

[0054] A target model determination module is used to determine and select at least one of the first trajectory model, the second trajectory model and the third trajectory model as the target model; wherein the first trajectory model is a first trajectory model determined based on the curvature of the current trajectory of the vehicle; the second trajectory model is a second trajectory model obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the target vehicle; and the third trajectory model is a lane centerline model determined based on a visual sensor;

[0055] a first information determination module, configured to determine a first transverse coordinate point corresponding to the coordinates of the first longitudinal sampling point based on a target model, determine a transverse coordinate point weight corresponding to the first transverse coordinate point, and determine a coordinate point corresponding to a center line of a forward lane based on the transverse coordinate point weight of the target model and the first transverse coordinate point;

[0056] a second information determination module, configured to filter the second lateral coordinate point determined based on the fourth trajectory model and the second longitudinal sampling point coordinates to obtain the filtered second lateral coordinate point, and determine the coordinate point corresponding to the center line of the rear lane based on the second longitudinal sampling point coordinates and the filtered second lateral coordinate point; wherein the fourth trajectory model is a trajectory model obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the vehicle;

[0057] The third information determination module is used to determine the target lane centerline based on the coordinate point corresponding to the forward lane centerline and the coordinate point corresponding to the backward lane centerline.

[0058] The present application also provides a lane centerline determination device, comprising:

[0059] Memory for storing computer programs;

[0060] A processor is used to execute the computer program to implement the steps of the lane centerline determination method as described above.

[0061] The present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the lane centerline determination method as described above are implemented.

[0062] An embodiment of the present invention further provides a computer program product, including a computer program / instruction, characterized in that when the computer program / instruction is executed by a processor, the steps of the lane centerline determination method as described above are implemented.

[0063] It can be seen that the present invention selects at least one of the first trajectory model, the second trajectory model and the third trajectory model as the target model by determining; wherein the first trajectory model is a first trajectory model determined based on the curvature of the current trajectory of the vehicle; the second trajectory model is a second trajectory model obtained by polynomial fitting based on the coordinates corresponding to the historical trajectory points of the target vehicle; the third trajectory model is a lane centerline model determined based on the visual sensor; the first lateral coordinate point corresponding to the first longitudinal sampling point coordinate is determined based on the target model, and the lateral coordinate point weight corresponding to the first lateral coordinate point is determined, and the coordinate point corresponding to the center line of the forward lane is determined based on the lateral coordinate point weight of the target model and the first lateral coordinate point; the second lateral coordinate point determined based on the fourth trajectory model and the second longitudinal sampling point coordinate is filtered to obtain the filtered second lateral coordinate point, and the coordinate point corresponding to the center line of the rear lane is determined based on the second longitudinal sampling point coordinate and the filtered second lateral coordinate point; wherein the fourth trajectory model is a trajectory model obtained by polynomial fitting based on the coordinates corresponding to the historical trajectory points of the vehicle; the center line of the target lane is determined based on the coordinate point corresponding to the center line of the forward lane and the coordinate point corresponding to the center line of the rear lane.

[0064] The beneficial effects of the present invention are as follows: compared with the current method of completely relying on sensors to generate lane centerlines, the present application uses multiple models when determining lane centerlines, among which the first trajectory model, the second trajectory model, and the fourth trajectory model can all be determined without using sensors, so that it is not completely dependent on visual sensors and high-precision maps. In addition, the present application uses the target trajectory model and the fourth trajectory model to determine the centerline of the forward lane and the centerline of the backward lane to obtain the target lane centerline, so that when the sensor cannot provide accurate information, it does not completely rely on the information provided by the visual sensor and the high-precision map to determine the lane centerline, so that the lane centerline can be accurately determined even when the sensor is limited.

[0065] In addition, the present invention also provides a lane centerline determination device, equipment and readable storage medium, which also have the above-mentioned beneficial effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0067] Figure 1 A flow chart of a method for determining a lane centerline provided by an embodiment of the present invention;

[0068] Figure 2 A schematic diagram of the closest point and the farthest point output by a visual sensor provided by an embodiment of the present invention;

[0069] Figure 3 A schematic diagram of an Ackerman steering model provided by an embodiment of the present invention;

[0070] Figure 4 An example flow chart of a method for determining a lane centerline provided by an embodiment of the present invention;

[0071] Figure 5 A trajectory schematic diagram provided for an embodiment of the present invention;

[0072] Figure 6 A schematic diagram of the structure of a lane centerline determination device provided by an embodiment of the present invention;

[0073] Figure 7 A schematic diagram of the structure of a lane centerline determination device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0074] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0075] Please refer to Figure 1 , Figure 1 A flow chart of a method for determining a lane centerline provided by an embodiment of the present invention. The method may include:

[0076] S101, determine to select at least one of a first trajectory model, a second trajectory model and a third trajectory model as a target model; wherein the first trajectory model is a first trajectory model determined based on the curvature of the current trajectory of the vehicle; the second trajectory model is a second trajectory model obtained by performing polynomial fitting based on coordinates corresponding to historical trajectory points of the target vehicle; and the third trajectory model is a lane centerline model determined based on a visual sensor.

[0077] The executor of this embodiment is an electronic device. The electronic device may be a tablet; or the electronic device may also be a computer, etc. The first trajectory model in this embodiment is a model determined based on the curvature of the vehicle's trajectory, so the first trajectory model in this embodiment is the vehicle's trajectory model. The curvature of the vehicle's trajectory in this embodiment refers to the degree to which the curve deviates from a straight line, which mathematically indicates the value of the degree of curvature of the curve at a certain point. The larger the curvature, the greater the degree of curvature of the curve at that point. In this embodiment, a quadratic function model can be constructed based on the curvature of the vehicle's trajectory to obtain the first trajectory model. For example, the first trajectory model can be y=kappa*x 2 Wherein, kappa represents the curvature of the vehicle trajectory, x represents the coordinate value in the x direction, y represents the coordinate value in the y direction, x represents the longitudinal coordinate, and y represents the lateral coordinate. The second trajectory model in this embodiment is a polynomial equation, and the second trajectory model in this embodiment is a historical trajectory model of the target vehicle. The third trajectory model in this embodiment is a model obtained by determining the coefficients of the polynomial equation based on a visual sensor, so the third trajectory model in this embodiment is a visual sensor model.

[0078] It should be further explained that, in order to improve the accuracy of determining the target model, before the above determination of selecting at least one of the first trajectory model, the second trajectory model and the third trajectory model as the target model, the following may also be included:

[0079] Step 1: Construct a quadratic function model based on the curvature of the vehicle trajectory to obtain a dynamic first trajectory model.

[0080] Step 2: Filter the horizontal coordinate points of the historical trajectory points of the target vehicle in each cycle to obtain the filtered horizontal coordinate points corresponding to the target vehicle, and obtain the historical trajectory points of the target vehicle corresponding to the target vehicle based on the filtered horizontal coordinate points.

[0081] This embodiment can perform a first-order low-pass filtering process based on the horizontal coordinate points of the historical trajectory points of the target vehicle and the horizontal coordinate points of the previous cycle to obtain the filtered horizontal coordinate points. y0 = ay + (1-a) y'; where a is the filter coefficient at the previous moment (for example, it can be 0.9, 0.85, 0.95, etc.), y0 and y' respectively represent the horizontal coordinate points of the current cycle and the horizontal coordinate points of the previous cycle. The historical trajectory points of the target vehicle are determined by recording a certain number of longitudinal coordinate points and their corresponding filtered horizontal coordinate points.

[0082] Step 3: Perform polynomial curve fitting based on the historical trajectory points of the target vehicle to obtain the second trajectory model.

[0083] This embodiment can perform polynomial curve fitting based on the historical trajectory points of the target vehicle to obtain the historical trajectory equation of the target vehicle; y=c3x 3 +c2x2 +c1x 3 +c0.

[0084] Step 4: Determine a third trajectory model based on the rate of change of the lane centerline curvature, the lane centerline curvature, the lane centerline heading angle, and the lane centerline offset distance determined by the visual sensor.

[0085] In this embodiment, the third trajectory model can be determined based on the change rate of the lane centerline curvature, the lane centerline curvature, the lane centerline heading angle and the lane centerline offset distance, which can be y=a3x 3 +a2x 2 +a1x 3 +a0, where a3 represents the rate of change of the lane centerline curvature, a2 represents the lane centerline curvature, a1 represents the lane centerline orientation angle, and a0 represents the lane centerline offset distance.

[0086] This embodiment improves the accuracy of determining various models in the target model by providing a specific method for determining various models in the target model, thereby improving the accuracy of subsequent calculation of the lane centerline.

[0087] It should be further explained that, in order to improve the accuracy of the target vehicle determination, the horizontal coordinate points of the historical trajectory points of each cycle of the target vehicle are filtered to obtain the filtered horizontal coordinate points corresponding to the target vehicle, and before obtaining the target vehicle historical trajectory points corresponding to the target vehicle based on the filtered horizontal coordinate points, it can also include: selecting a preset number of vehicles as target vehicles based on the target vehicle selection rule; wherein the target vehicle selection rule is a rule determined based on the coordinate value of the vehicle and the time when the vehicle exists in the field of vision of the vehicle. The coordinate value of the vehicle in this embodiment may refer to the limiting conditions of the longitudinal coordinate value and the transverse coordinate value of the vehicle. The target vehicle selection rule in this embodiment includes: the longitudinal coordinate value of the vehicle is within the set longitudinal coordinate value range; the difference between the transverse coordinate value of the vehicle and the transverse coordinate value of the vehicle is within the set difference threshold; the time when the vehicle exists in the field of vision of the vehicle is greater than the set time threshold; the standard deviation of the transverse coordinate value of the vehicle is less than the set standard deviation threshold. This embodiment does not limit the specific longitudinal coordinate value range, for example, the longitudinal coordinate value range can be between 4 meters and 120 meters; or the longitudinal coordinate value range can also be between 4 meters and 150 meters. This embodiment does not limit the specific difference threshold. For example, the difference threshold in this embodiment can be 1.5 meters and -1.5 meters; or the difference threshold in this embodiment can also be 1.4 meters to -1.5 meters, that is, the difference is between 1.5 meters and -1.5 meters, or the difference is between 1.4 meters and -1.5 meters. The time threshold in this embodiment can be 0.06 seconds, that is, not less than 0.06 seconds; or the time threshold can also be 0.07 seconds. This embodiment does not limit the specific standard deviation threshold. The standard deviation in this embodiment needs to be less than the standard deviation threshold. For example, the standard deviation threshold in this embodiment can be 0.05; or the standard deviation threshold in this embodiment can be 0.04. This embodiment improves the accuracy of target vehicle determination by setting a specific target vehicle selection method.

[0088] S102, determining a first transverse coordinate point corresponding to the coordinates of the first longitudinal sampling point based on the target model, and determining a transverse coordinate point weight corresponding to the first transverse coordinate point, and determining a coordinate point corresponding to the center line of the forward lane based on the transverse coordinate point weight of the target model and the first transverse coordinate point.

[0089] This embodiment does not limit the specific target model. For example, the target model in this embodiment can be any one of the first trajectory model, the second trajectory model and the third trajectory model; or the target model in this embodiment can be at least two of the first trajectory model, the second trajectory model and the third trajectory model; or the target model in this embodiment can be three models of the first trajectory model, the second trajectory model and the third trajectory model. This embodiment does not limit the specific method for determining the coordinates of the first longitudinal sampling point. The coordinates of the first longitudinal sampling point in this embodiment are the coordinates related to the center line of the lane. For example, this embodiment can first determine the starting point and the end point of the coordinates of the first longitudinal sampling point, and then determine the coordinates of the first longitudinal sampling point. This embodiment does not limit the specific way of the weight of the lateral coordinate point corresponding to the first lateral coordinate. For example, this embodiment can determine the weight of the lateral coordinate point based on demand; or this embodiment can determine the weight of the lateral coordinate point based on the variance. In this embodiment, the process of determining the coordinate point corresponding to the center line of the forward lane based on the lateral coordinate point weight of the target model and the first lateral coordinate point can be to determine the target lateral coordinate point based on the lateral coordinate point weight and the first lateral coordinate point, and determine the center line of the forward lane based on the target lateral coordinate point and the coordinates of the first longitudinal sampling point. The center line of the forward lane in this embodiment refers to a reference line corresponding to the current first trajectory model, that is, the current first trajectory model and the fourth trajectory model are used as a method to distinguish the center of the forward lane from the center line of the rear lane. In this embodiment, the determination of each coordinate point can be a point determined based on the rear axle center of the vehicle.

[0090] It should be further explained that when the first trajectory model, the second trajectory model and the third trajectory model are determined as the target models, the first lateral coordinate point corresponding to the coordinate of the first longitudinal sampling point is determined based on the target model, and the lateral coordinate point weight corresponding to the first lateral coordinate point is determined, and the coordinate point corresponding to the center line of the forward lane is determined based on the lateral coordinate point weight of the target model and the first lateral coordinate point, which may include:

[0091] S1021, determining, based on the first trajectory model, the second trajectory model, and the third trajectory model, a first lateral coordinate point of the vehicle, a first lateral coordinate point of the target vehicle, and a first lateral coordinate point of the sensor corresponding to the coordinates of the first longitudinal sampling point, respectively.

[0092] This embodiment can perform forward longitudinal distance sampling, sampling a longitudinal coordinate at a certain distance (for example, 2 meters) to obtain the first longitudinal sampling point coordinates. For example, the longitudinal sampling point coordinates in this embodiment can be Δx, 2Δx, 3Δx, 4Δx..., nΔx. According to the first longitudinal sampling point coordinates and the three trajectory models (trajectory line 1, trajectory line 2 and trajectory line 3 represent the first trajectory model, the second trajectory model and the third trajectory model respectively), the first transverse coordinate points under the first longitudinal sampling point coordinates are calculated respectively, as shown in Table 1: Table 1 is a schematic table of first transverse coordinate points provided by an embodiment of the present invention.

[0093] Table 1 A schematic table of the first horizontal coordinate point

[0094]

[0095] S1022: Determine an average value of the first transverse coordinate points corresponding to the coordinates of each first longitudinal sampling point based on the first transverse coordinate point of the own vehicle, the first transverse coordinate point of the target vehicle, and the first transverse coordinate point of the sensor.

[0096] The average values ​​of the first horizontal coordinate points determined in this embodiment are shown in Table 2. Table 2 is a schematic table of average values ​​of the first horizontal coordinate points provided in an embodiment of the present invention.

[0097] Table 2 Schematic table of the average value of the first horizontal coordinate point

[0098]

[0099] S1023, based on the first lateral coordinate point of the ego vehicle, the first lateral coordinate point of the target vehicle, the first lateral coordinate point of the sensor, and the average value of the first lateral coordinate point, determine the variance of the lateral coordinate points under the three models corresponding to the coordinates of each first longitudinal sampling point.

[0100] The variances of the lateral coordinate points corresponding to the three trajectory models in this embodiment are shown in Table 3. Table 3 is a schematic table of the variances of the lateral coordinate points provided in an embodiment of the present invention.

[0101] Table 3 Schematic diagram of the variance of a horizontal coordinate point

[0102] Δx 2Δx 3Δx 4Δx 5Δx 6Δx 7Δx …… nΔx Track line 1 <![CDATA[σ 11 ]]> <![CDATA[σ 12 ]]> <![CDATA[σ 13 ]]> <![CDATA[σ 14 ]]> <![CDATA[σ 15 ]]> <![CDATA[σ 16 ]]> <![CDATA[σ 17 ]]> …… <![CDATA[σ 1n ]]> Track line 2 <![CDATA[σ 21 ]]> <![CDATA[σ 22 ]]> <![CDATA[σ 23 ]]> <![CDATA[σ 24 ]]> <![CDATA[σ 25 ]]> <![CDATA[σ2x]]> <![CDATA[σ 27 ]]> …… <![CDATA[σ 2n ]]> Track line 3 <![CDATA[σ 31 ]]> <![CDATA[σ 32 ]]> <![CDATA[σ 33 ]]> <![CDATA[σ 34 ]]> <![CDATA[σ 35 ]]> <![CDATA[σ 36 ]]> <![CDATA[σ 37 ]]> …… <![CDATA[σ 3n ]]>

[0103] S1024: Determine a weighted average of first transverse coordinate points corresponding to each first longitudinal sampling point based on the variance of the transverse coordinate points.

[0104] It is understandable that in statistics, each data point has a corresponding weight value, which determines the importance of the data point in calculating the average value. By multiplying the value of each data point by the weight value, summing the results and dividing them by the sum of all weight values, the weighted average value can be obtained. Here, the inverse variance weighting method is used, that is, the weight of the data is inversely proportional to the variance, and the weighted average value of each first horizontal coordinate point is determined.

[0105] S1025: Determine a horizontal coordinate point weight corresponding to each first horizontal coordinate point based on the weighted average value.

[0106] This embodiment uses the weighted average value as the weight of the horizontal coordinate point. This embodiment provides a specific method for determining the weight, thereby improving the accuracy of weight determination.

[0107] S1026: Determine a target first transverse coordinate point corresponding to each first longitudinal sampling point based on the weight of each transverse coordinate point, and obtain a coordinate point corresponding to the center line of the forward lane based on the coordinates of the first longitudinal sampling point and the target first transverse coordinate point.

[0108] In this embodiment, the target first transverse coordinate point corresponding to the first longitudinal sampling point coordinate can be obtained, and the two are in a one-to-one correspondence. Through the above processing, this embodiment can improve the accuracy of determining the coordinate point corresponding to the center line of the forward lane.

[0109] It should be further explained that, in order to improve the accuracy of determining the coordinates of the first longitudinal sampling point, before determining the first lateral coordinate point corresponding to the coordinates of the first longitudinal sampling point based on the target model, it may also include: determining whether the visual sensor can determine the end point of the lane centerline; when it is determined that the end point of the lane centerline can be determined, the maximum point of the coordinates of the first longitudinal sampling point is set based on the end point determined by the visual sensor; when it is determined that the end point of the lane centerline cannot be determined, the end point of the lane centerline is calculated by linear interpolation of the vehicle speed, and the maximum point of the coordinates of the first longitudinal sampling point is set based on the end point determined by linear interpolation of the vehicle speed. The linear difference of the speed in this embodiment refers to determining the distance corresponding to the current speed based on the given speed and the corresponding distance, and taking the distance as the end point. For ease of understanding, please refer to Table 3, which is a linear interpolation table of speed provided in an embodiment of the present invention. For example, the position corresponding to the speed of 75 is determined as the end point, and the distance value corresponding to the end point is determined based on the speed of 60, the speed of 80, the distance d2 and the distance d3.

[0110] Table 4 A linear interpolation table of vehicle speed

[0111] speed distance 20 d0 40 d1 60 d2 80 d3 100 d4 120 d5

[0112] For easier understanding, please refer to Figure 2 , Figure 2A schematic diagram of the nearest point and the farthest point (end point) output by a visual sensor provided in an embodiment of the present invention; wherein the nearest point can be used as the starting point of the first longitudinal sampling point coordinates, or the farthest point can be used as the end point of the first longitudinal sampling point coordinates, and d-low in the figure represents the nearest point, and d-up represents the farthest point.

[0113] S103, filtering the second lateral coordinate point determined based on the fourth trajectory model and the coordinates of the second longitudinal sampling point to obtain the filtered second lateral coordinate point, and determining the coordinate point corresponding to the center line of the rear lane based on the coordinates of the second longitudinal sampling point and the filtered second lateral coordinate point; wherein the fourth trajectory model is a trajectory model obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the vehicle.

[0114] The fourth trajectory model in this embodiment can be a model obtained by polynomial fitting based on the coordinates corresponding to the historical trajectory points of the vehicle. This embodiment does not limit the specific process of obtaining the fourth trajectory model. For example, this embodiment can select historical trajectory points to obtain target historical trajectory points, and obtain the fourth trajectory model based on the target historical trajectory points; or this embodiment can also directly determine the fourth trajectory model based on the historical trajectory points of the vehicle. This embodiment does not limit the specific determination method of the coordinates of the second longitudinal sampling point. For example, the coordinates of the second longitudinal sampling point in this embodiment can be determined based on the end point related to the determined rear lane centerline; or this embodiment can be based on the second longitudinal sampling point coordinates. The coordinates can be determined based on the starting point and end point related to the rear lane centerline. In order to suppress noise, this embodiment does not limit the specific method of filtering the second lateral coordinate point. For example, this embodiment can be based on the current second lateral coordinate point (which can be understood as the currently used historical second lateral coordinate point) and the second lateral coordinate point of the previous cycle. , To obtain the filtered second lateral coordinate point.

[0115] It should be further explained that the process of determining the coordinates of the second longitudinal sampling point may include: determining the minimum point of the center line of the rear lane; determining the coordinates of the second longitudinal sampling point based on the minimum point. This embodiment does not limit the specific minimum value. For example, the minimum value may be -20; or the minimum value may also be -50, where the negative sign is because it is the center line of the rear lane. This embodiment may use the minimum value as The end point of the second longitudinal sampling point is determined based on the end point of the second longitudinal sampling point and the set interval distance. As the set sampling distance (the sampling distance corresponding to the second longitudinal sampling point), the coordinates of the second longitudinal sampling point can be expressed as -Δx, -2Δx, -3Δ, -4Δx..., -nΔx. This embodiment provides a specific method for determining the left coordinate of the second longitudinal sampling point, thereby improving the accuracy of determining the coordinates of the second longitudinal sampling point.

[0116] It should be further explained that, in order to improve the accuracy of determining the fourth trajectory model, the above-mentioned polynomial fitting based on the coordinates corresponding to the historical trajectory points of the vehicle to obtain the fourth trajectory model may include:

[0117] S1, obtain a set number of self-vehicle historical trajectory points; select from the set number of self-vehicle historical trajectory points based on the self-vehicle historical trajectory point selection conditions to obtain the target self-vehicle historical trajectory point. The self-vehicle historical trajectory point selection conditions include: the horizontal coordinate of the trajectory point is greater than the set horizontal coordinate threshold; the longitudinal coordinate at the current moment is greater than the longitudinal coordinate at the previous moment; the longitudinal coordinate at the current moment is less than zero; the longitudinal coordinate at the current moment is less than the minimum value of the longitudinal coordinate of the self-vehicle historical trajectory point; the longitudinal coordinate at the current moment is not equal to the value of the longitudinal coordinate at the previous moment; the difference between the absolute values ​​of the horizontal coordinates at the previous moment and the current moment is greater than the set absolute value difference threshold.

[0118] The set lateral coordinate threshold in this embodiment may include a lateral coordinate maximum value and a lateral coordinate minimum value, wherein the minimum value may be -1, -1.5, etc.; the lateral coordinate maximum value in this embodiment may be 1, 1.5, etc. This embodiment does not limit the minimum value of the longitudinal coordinate of the historical trajectory point of the vehicle, for example, the minimum value may be 50, 51, etc. The absolute value difference threshold in this embodiment may be 0.05, 0.06, etc. This embodiment does not limit the specific method for determining the historical trajectory point of the vehicle, for example, this embodiment may be determined based on the Ackerman steering model. For ease of understanding, please refer to Figure 3 , Figure 3 , which is a schematic diagram of an Ackerman steering model provided in an embodiment of the present invention. Figure 3 Where δi and δo represent the turning angles of the inner and outer front wheels, respectively. Lb and Lw represent the length and width of the vehicle, respectively. delt-x = egoCurneRadius*sina; delt-y = egoCurneRadius*cosa. Where a represents the curvature radius of the ego vehicle's historical trajectory, delt-x and delt-y represent the offsets of the x and y positions.

[0119] S2, performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the target vehicle to obtain the fourth trajectory model.

[0120] This embodiment performs polynomial fitting based on the above obtained vehicle historical trajectory points, and the equation of the vehicle historical trajectory can be a polynomial equation including cubic term coefficients, quadratic term coefficients, linear term coefficients and constant terms. This embodiment provides a specific process for determining the fourth trajectory model, which improves the accuracy of determining the fourth trajectory model.

[0121] S104: Determine the center line of the target lane based on the coordinate points corresponding to the center line of the forward lane and the coordinate points corresponding to the center line of the backward lane.

[0122] This embodiment does not limit the specific method for determining the target lane centerline. For example, this embodiment can directly determine the target lane centerline based on the coordinate points corresponding to the rear lane centerline and the coordinate points corresponding to the front lane centerline; or this embodiment can also determine the rear lane centerline model based on the coordinate points corresponding to the rear lane centerline, determine the front lane centerline model based on the coordinate points corresponding to the front lane centerline, and determine the target lane centerline based on the front lane center model and the rear lane centerline model.

[0123] It should be further explained that the above-mentioned determination of the target lane centerline based on the coordinate points corresponding to the centerline of the front lane and the coordinate points corresponding to the centerline of the rear lane may include: determining the front lane centerline model based on the coordinate points corresponding to the centerline of the front lane; determining the rear lane centerline model based on the coordinate points corresponding to the centerline of the rear lane; and determining the target lane centerline based on the front lane centerline model and the rear lane centerline model. This embodiment may determine the target lane centerline based on the front lane centerline model and the rear lane centerline model.

[0124] A method for determining a lane centerline provided by an embodiment of the present invention may include: S101, determining to select at least one of a first trajectory model, a second trajectory model, and a third trajectory model as a target model; wherein the first trajectory model is determined based on the curvature of the vehicle trajectory; and a second trajectory model is obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the target vehicle. S102, determining a first transverse coordinate point corresponding to the coordinates of the first longitudinal sampling point based on the target model, and determining a transverse coordinate point weight corresponding to the first transverse coordinate point, and determining a coordinate point corresponding to the centerline of the forward lane based on the transverse coordinate point weight of the target model and the first transverse coordinate point. S103, determining a second transverse coordinate point corresponding to the coordinates of the second longitudinal sampling point based on the fourth trajectory model, and filtering the corresponding second transverse coordinate point of the previous cycle to obtain a filtered second transverse coordinate point; wherein a fourth trajectory model is obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the vehicle. S104, determining a coordinate point corresponding to the centerline of the backward lane based on the coordinates of the second longitudinal sampling point and the filtered second transverse coordinate point, and determining the centerline of the target lane based on the coordinate point corresponding to the centerline of the forward lane and the coordinate point corresponding to the centerline of the backward lane. Compared with the current method of generating lane center lines completely relying on sensors, the present application uses parameters that can be obtained without relying on sensors when generating lane center lines, so that the lane center lines can be accurately determined based on multiple models without relying on sensors. In addition, specific methods for determining the first trajectory model, the second trajectory model, and the third trajectory model are given, which improves the accuracy of model determination; and, the embodiment determines the target vehicle based on set conditions, so that the lane center line can be determined based on the accurate target vehicle; and, the embodiment gives a specific process for determining the coordinate points corresponding to the center line of the forward lane based on the first trajectory model, the second trajectory model, and the third trajectory model, which improves the accuracy of determining the coordinate points corresponding to the center line of the forward lane; and, a specific method for determining the coordinates of the first longitudinal sampling point and the second longitudinal sampling point is given, which improves the accuracy of sampling; and, a specific method for determining the fourth trajectory model is given, which improves the accuracy of determining the fourth trajectory model.

[0125] In order to make the present invention easier to understand, please refer to Figure 4 , Figure 4 A flowchart of a method for determining a lane centerline provided by an embodiment of the present invention may specifically include:

[0126] S201, selecting from the historical trajectory points of the vehicle to obtain the target historical trajectory point of the vehicle, performing polynomial fitting based on the coordinates corresponding to the target historical trajectory point of the vehicle, and obtaining the historical trajectory model of the vehicle (corresponding to the fourth trajectory model above).

[0127] For easier understanding, please refer to Figure 5 , Figure 5 A trajectory diagram provided for an embodiment of the present invention includes the target vehicle's historical trajectory, the vehicle's trajectory, the vehicle's historical trajectory, and the visual lane centerline. The selection conditions in the embodiment of the present invention include: ① a point whose horizontal coordinate is greater than the horizontal coordinate threshold; ② the vertical coordinate at the current moment is greater than the vertical coordinate at the previous moment; ③ the vertical coordinate at the current moment is less than 0; ④ the vertical coordinate at the current moment is less than the minimum value of the vertical coordinate of the vehicle's historical trajectory point; ⑤ the vertical coordinate at the current moment is not equal to the value of the vertical coordinate at the previous moment; ⑥ the difference in the absolute value of the horizontal coordinates at the previous moment and the current moment is greater than the absolute value difference threshold.

[0128] S202, determining a current trajectory model of the vehicle based on the curvature of the vehicle trajectory.

[0129] The current trajectory model of the vehicle in this embodiment corresponds to the above-mentioned trajectory model of the vehicle (corresponding to the above-mentioned first trajectory model).

[0130] S203, determining a target vehicle from adjacent vehicles based on a selection rule, performing a first-order low-pass filtering process on the lateral coordinate points of the historical trajectory points of the target vehicle in the current cycle and the previous cycle, and obtaining filtered historical trajectory points of the target vehicle.

[0131] In this embodiment, the selection rules include: ① the longitudinal coordinate x of the target vehicle is within the set longitudinal coordinate range; ② the difference between the horizontal coordinate of the target vehicle and the horizontal coordinate of the own vehicle is within the set difference threshold; ③ the period value of the target in the field of view is greater than the set time threshold; ④ the standard deviation of the target horizontal coordinate is less than the set standard deviation threshold.

[0132] S204, performing polynomial curve fitting based on the filtered historical trajectory points of the target vehicle to obtain a historical trajectory model of the target vehicle (corresponding to the second trajectory model described above).

[0133] The historical trajectory model of the target vehicle in this embodiment is:

[0134] y3=c3x3 3 +c2x3 2 +c1x3+c0.

[0135] S205 , determining a visual sensor model (corresponding to the third trajectory model described above) based on the data collected by the visual sensor as parameters.

[0136] The visual sensor model in this embodiment is y4=a3x4 3 +a2x4 2 +a1x4+a0; where a3 represents the rate of change of the lane centerline curvature, a2 represents the lane centerline curvature, a1 represents the lane line orientation angle, and a0 represents the lane centerline offset distance.

[0137] S206, determining an end point of the center line of the forward lane based on a visual sensor or a vehicle speed linear interpolation table, and determining a first longitudinal sampling point coordinate based on the end point of the center line of the forward lane.

[0138] In this embodiment, ① the visual sensor is used to determine whether the lane line quality meets the requirements. If so, the starting point and end point of the lane center line in the visual sensor are used as the starting point and end point of the front lane center line; ② if not, the starting point and end point of the front lane center line are calculated by linear interpolation of the vehicle speed.

[0139] S207, based on the coordinates of the first longitudinal sampling point, determine the first lateral coordinate points corresponding to the three trajectory models, namely, the vehicle current trajectory model, the target historical trajectory model, and the visual sensor model, and calculate the average value of the first lateral coordinate points.

[0140] S208, determining the variance of the first horizontal coordinate point according to the first horizontal coordinate point of each trajectory line model and the average value of the first horizontal coordinate point, determining the weight of the weighted average value based on the variance of the first horizontal coordinate point, and determining the target first horizontal coordinate corresponding to each first longitudinal sampling point coordinate based on the weight.

[0141] S209: Determine the coordinates of the center line of the forward lane based on the first lateral coordinate of the target and the first longitudinal sampling point coordinates.

[0142] S210, determining the coordinates of the second longitudinal sampling point based on the minimum value of the center line of the rear lane, and performing a first-order low-pass filtering process based on the coordinates of the second longitudinal sampling point and the vehicle historical trajectory points of the current historical period and the corresponding vehicle historical trajectory points of the previous period corresponding to the vehicle historical trajectory model, to obtain a filtered second lateral coordinate point.

[0143] The vehicle history trajectory model in this embodiment corresponds to the fourth trajectory model mentioned above.

[0144] S211, determining the coordinates of the center line of the rear lane based on the filtered second lateral coordinate point and the second longitudinal sampling point coordinates.

[0145] S212: Determine the center line of the target lane based on the coordinates of the center line of the front lane and the coordinates of the center line of the rear lane.

[0146] The embodiment of the present invention utilizes multiple trajectory line models to perform forward and backward trajectory sampling, and integrates multiple models to calculate the lane center line, thereby improving the accuracy of determining the target lane center line in situations such as intersections with low lane line quality or no lane lines, unstructured roads, and curves.

[0147] The following is an introduction to a lane centerline determination device provided in an embodiment of the present invention. The lane centerline determination device described below and the lane centerline determination method described above can be referenced to each other.

[0148] Please refer to Figure 6 , Figure 6 A schematic diagram of a structure of a lane centerline determination device provided by an embodiment of the present invention may include:

[0149] The target model determination module 100 is used to determine and select at least one of the first trajectory model, the second trajectory model and the third trajectory model as the target model; wherein the first trajectory model is determined based on the curvature of the vehicle trajectory; and the second trajectory model is obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the target vehicle;

[0150] A first information determination module 200 is used to determine a first transverse coordinate point corresponding to the coordinate of the first longitudinal sampling point based on a target model, determine a transverse coordinate point weight corresponding to the first transverse coordinate point, and determine a coordinate point corresponding to a center line of a forward lane based on the transverse coordinate point weight of the target model and the first transverse coordinate point;

[0151] The second information determination module 300 is used to determine the second transverse coordinate point corresponding to the coordinate of the second longitudinal sampling point based on the fourth trajectory model, and the corresponding second transverse coordinate point of the previous cycle for filtering to obtain the filtered second transverse coordinate point; wherein the fourth trajectory model is obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the vehicle;

[0152] The third information determination module 400 is used to determine the coordinate point corresponding to the center line of the rear lane based on the coordinates of the second longitudinal sampling point and the filtered second lateral coordinate point, and to determine the center line of the target lane based on the coordinate point corresponding to the center line of the front lane and the coordinate point corresponding to the center line of the rear lane.

[0153] Further, based on the above embodiment, the lane centerline determination device may further include:

[0154] A first trajectory model determination module, configured to construct a quadratic function model based on the curvature of the vehicle trajectory to obtain the first dynamic trajectory model;

[0155] The target vehicle historical trajectory point determination module is used to filter the horizontal coordinate points of the historical trajectory points of each cycle of the target vehicle to obtain the filtered horizontal coordinate points corresponding to the target vehicle, and obtain the target vehicle historical trajectory points corresponding to the target vehicle based on the filtered horizontal coordinate points;

[0156] A second trajectory model determination module, configured to perform polynomial curve fitting based on the historical trajectory points of the target vehicle to obtain the second trajectory model;

[0157] The third trajectory model determination module is used to determine the third trajectory model based on the rate of change of the lane centerline curvature, the lane centerline curvature, the lane centerline heading angle and the lane centerline offset distance determined by the visual sensor.

[0158] Further, based on the above embodiment, the lane centerline determination device may further include:

[0159] A target vehicle determination module is used to select a preset number of vehicles as the target vehicles based on a target vehicle selection rule; wherein the target vehicle selection rule is a rule determined based on the coordinate value of the vehicle and the time the vehicle exists in the field of view of the own vehicle; wherein the target vehicle selection rule may include: the longitudinal coordinate value of the vehicle is within a set longitudinal coordinate value range; the difference between the lateral coordinate value of the vehicle and the lateral coordinate value of the own vehicle is within a set difference threshold; the time the vehicle exists in the field of view of the own vehicle is greater than a set time threshold; the standard deviation of the lateral coordinate value of the vehicle is less than a set standard deviation threshold.

[0160] Further, based on any of the above embodiments, the target model determination module 100 includes:

[0161] a target model determining unit, configured to determine the first trajectory model, the second trajectory model and the third trajectory model as the target model;

[0162] Accordingly, the first information determination module 200 includes:

[0163] a transverse coordinate point determination unit, configured to determine, based on the first trajectory model, the second trajectory model, and the third trajectory model, a first transverse coordinate point of the vehicle, a first transverse coordinate point of the target vehicle, and a first transverse coordinate point of the sensor corresponding to the coordinates of the first longitudinal sampling point, respectively;

[0164] an average value determining unit, for determining an average value of the first transverse coordinate point corresponding to the coordinate of each first longitudinal sampling point based on the first transverse coordinate point of the self-vehicle, the first transverse coordinate point of the target vehicle and the first transverse coordinate point of the sensor;

[0165] A variance determination unit, configured to determine the variance of the lateral coordinate points under three models corresponding to the coordinates of each first longitudinal sampling point based on the first lateral coordinate point of the ego vehicle, the first lateral coordinate point of the target vehicle, the first lateral coordinate point of the sensor, and an average value of the first lateral coordinate point;

[0166] a weighted average value determining unit, configured to determine a weighted average value of first transverse coordinate points corresponding to each first longitudinal sampling point based on the variance of the transverse coordinate points;

[0167] A weight determination unit, configured to determine a horizontal coordinate point weight corresponding to each first horizontal coordinate point based on the weighted average value;

[0168] The coordinate point determination unit corresponding to the center line of the forward lane is used to determine the target first transverse coordinate point corresponding to each first longitudinal sampling point based on the weight of each transverse coordinate point, and obtain the coordinate point corresponding to the center line of the forward lane based on the first longitudinal sampling point and the target first transverse coordinate point.

[0169] Further, based on any of the above embodiments, the lane centerline determination device may further include:

[0170] A judgment module, used to determine whether the visual sensor can determine the end point of the lane centerline;

[0171] A first first longitudinal sampling point coordinate setting module, configured to set a maximum value point of the first longitudinal sampling point coordinate based on the end point determined by the visual sensor when an end point of the lane centerline is determined;

[0172] The second first longitudinal sampling point coordinate setting module is used to calculate the end point of the lane centerline by linear interpolation of the vehicle speed when it is determined that the end point of the lane centerline cannot be determined, and set the maximum point of the first longitudinal sampling point coordinate based on the end point determined by linear interpolation of the vehicle speed.

[0173] Further, based on any of the above embodiments, the lane centerline determination device may further include:

[0174] A minimum value point determination module, used to determine the minimum value point of the center line of the rear lane;

[0175] The second longitudinal sampling point coordinate determining module is used to determine the second longitudinal sampling point coordinate based on the minimum value point.

[0176] Further, based on any of the above embodiments, the second information determination module 300 may include:

[0177] A set number of ego vehicle historical trajectory point determination unit, used to obtain a set number of ego vehicle historical trajectory points;

[0178] A target self-vehicle historical trajectory point determination unit, used for selecting a target self-vehicle historical trajectory point from the set number of self-vehicle historical trajectory points based on the self-vehicle historical trajectory point selection condition to obtain a target self-vehicle historical trajectory point;

[0179] a fourth trajectory model determination unit, configured to perform polynomial fitting based on coordinates corresponding to historical trajectory points of the target vehicle to obtain the fourth trajectory model;

[0180] The conditions for selecting the historical trajectory points of the vehicle include:

[0181] The point whose horizontal coordinate of the trajectory point is greater than the set horizontal coordinate threshold;

[0182] The vertical coordinate at the current moment is greater than the vertical coordinate at the previous moment;

[0183] The vertical coordinate at the current moment is less than zero;

[0184] The longitudinal coordinate at the current moment is less than the minimum value of the longitudinal coordinate of the historical trajectory point of the vehicle;

[0185] The vertical coordinate at the current moment is not equal to the vertical coordinate at the previous moment;

[0186] The difference between the absolute values ​​of the horizontal coordinates at the previous moment and the current moment is greater than the set absolute value difference threshold.

[0187] Further, based on any of the above embodiments, the third information determination module 400 may include:

[0188] A forward lane centerline model determination unit, configured to determine a forward lane centerline model based on a coordinate point corresponding to the forward lane centerline;

[0189] A rear lane centerline model determination unit, configured to determine a rear lane centerline model based on coordinate points corresponding to the rear lane centerline;

[0190] A target lane centerline determination unit is used to determine the target lane centerline based on the forward lane centerline model and the backward lane centerline model.

[0191] It should be noted that the order of the modules and units in the above lane centerline determination device can be changed without affecting the logic.

[0192] The lane centerline determination device provided by the embodiment of the present invention may include: a target model determination module 100, used to determine and select at least one of a first trajectory model, a second trajectory model and a third trajectory model as a target model; wherein the first trajectory model is determined based on the curvature of the vehicle trajectory; a polynomial fitting is performed based on the coordinates corresponding to the historical trajectory points of the target vehicle to obtain the second trajectory model; a first information determination module 200, used to determine a first lateral coordinate point corresponding to the coordinates of the first longitudinal sampling point based on the target model, and determine a lateral coordinate point weight corresponding to the first lateral coordinate point, and determine the forward vehicle based on the lateral coordinate point weight of the target model and the first lateral coordinate point. The coordinate point corresponding to the center line of the lane; the second information determination module 300 is used to determine the second lateral coordinate point corresponding to the coordinate of the second longitudinal sampling point based on the fourth trajectory model, and the corresponding second lateral coordinate point of the previous cycle is filtered to obtain the filtered second lateral coordinate point; wherein, the fourth trajectory model is obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the vehicle; the third information determination module 400 is used to determine the coordinate point corresponding to the center line of the rear lane based on the coordinates of the second longitudinal sampling point and the filtered second lateral coordinate point, and determine the target lane center line based on the coordinate point corresponding to the center line of the front lane and the coordinate point corresponding to the center line of the rear lane. Compared with the current method of generating the lane center line completely relying on sensors, the present application uses parameters that can be obtained without relying on sensors when generating the lane center line, so that the lane center line can be accurately determined based on multiple models without relying on sensors.

[0193] A lane centerline determination device provided in an embodiment of the present invention is introduced below. The lane centerline determination device described below and the lane centerline determination method described above can be referenced to each other.

[0194] Please refer to Figure 7 , Figure 7 A schematic diagram of a lane centerline determination device provided in an embodiment of the present invention may include:

[0195] A memory 10, used for storing computer programs;

[0196] The processor 20 is used to execute a computer program to implement the above-mentioned method for determining the lane centerline.

[0197] The memory 10 , the processor 20 , and the communication interface 30 all communicate with each other via a communication bus 40 .

[0198] In the embodiment of the present invention, the memory 10 is used to store one or more programs, and the program may include program code, and the program code includes computer operation instructions. In the embodiment of the present invention, the memory 10 may store programs for implementing the following functions:

[0199] Determine and select at least one of the first trajectory model, the second trajectory model and the third trajectory model as the target model; wherein the first trajectory model is determined based on the curvature of the vehicle trajectory; and the second trajectory model is obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the target vehicle.

[0200] A first lateral coordinate point corresponding to the coordinates of the first longitudinal sampling point is determined based on the target model, and a lateral coordinate point weight corresponding to the first lateral coordinate point is determined. A coordinate point corresponding to the center line of the forward lane is determined based on the lateral coordinate point weight of the target model and the first lateral coordinate point.

[0201] Based on the fourth trajectory model, the second lateral coordinate point corresponding to the coordinate of the second longitudinal sampling point and the corresponding second lateral coordinate point of the previous cycle are determined for filtering to obtain the filtered second lateral coordinate point; wherein, polynomial fitting is performed based on the coordinates corresponding to the historical trajectory points of the vehicle to obtain the fourth trajectory model.

[0202] The coordinate point corresponding to the center line of the rear lane is determined based on the coordinates of the second longitudinal sampling point and the filtered second lateral coordinate point, and the center line of the target lane is determined based on the coordinate point corresponding to the center line of the front lane and the coordinate point corresponding to the center line of the rear lane.

[0203] In a possible implementation, the memory 10 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function, etc.; the data storage area may store data created during use.

[0204] In addition, the memory 10 may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include an NVRAM. The memory stores an operating system and operating instructions, executable modules or data structures, or a subset thereof, or an extended set thereof, wherein the operating instructions may include various operating instructions for implementing various operations. The operating system may include various system programs for implementing various basic tasks and processing hardware-based tasks.

[0205] The processor 20 may be a central processing unit (CPU), an application specific integrated circuit, a digital signal processor, a field programmable gate array or other programmable logic device, a microprocessor or any conventional processor, etc. The processor 20 may call a program stored in the memory 10 .

[0206] The communication interface 30 may be an interface of a communication module, and is used to connect to other devices or systems.

[0207] Of course, it should be noted that Figure 7 The structure shown does not constitute a limitation on the lane centerline determination device in the embodiment of the present invention. In actual applications, the lane centerline determination device may include: Figure 7 More or fewer components than shown, or combinations of certain components.

[0208] The following is an introduction to a readable storage medium provided in an embodiment of the present invention. The readable storage medium described below and the method for determining the lane center line described above can be referenced to each other.

[0209] The present invention also provides a readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above-mentioned method for determining the lane center line are implemented.

[0210] The readable storage medium may include: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0211] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0212] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0213] Finally, it should be noted that, in this article, relationships such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variations are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0214] The above is a detailed introduction to a lane centerline determination method, device, equipment and readable storage medium provided by the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for determining a lane centerline, characterized in that: include: Determine to select at least one of the first trajectory model, the second trajectory model and the third trajectory model as the target model; wherein the first trajectory model is a first trajectory model determined based on the curvature of the current trajectory of the vehicle; the second trajectory model is a second trajectory model obtained by polynomial fitting based on the coordinates corresponding to the historical trajectory points of the target vehicle; the third trajectory model is a lane centerline model determined based on a visual sensor; Determine a first transverse coordinate point corresponding to the coordinate of the first longitudinal sampling point based on the target model, determine a transverse coordinate point weight corresponding to the first transverse coordinate point, and determine a coordinate point corresponding to the center line of the forward lane based on the transverse coordinate point weight of the target model and the first transverse coordinate point; Performing filtering processing on the second transverse coordinate point determined based on the fourth trajectory model and the coordinates of the second longitudinal sampling point to obtain a filtered second transverse coordinate point, and determining a coordinate point corresponding to the center line of the rear lane based on the coordinates of the second longitudinal sampling point and the filtered second transverse coordinate point; wherein the fourth trajectory model is a trajectory model obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the vehicle; The target lane centerline is determined based on the coordinate point corresponding to the forward lane centerline and the coordinate point corresponding to the backward lane centerline.

2. The method for determining the lane centerline according to claim 1, characterized in that: Before determining to select at least one of the first trajectory model, the second trajectory model and the third trajectory model as the target model, the method further includes: Constructing a quadratic function model based on the curvature of the vehicle trajectory to obtain the first dynamic trajectory model; Performing filtering processing on the horizontal coordinate points of the historical trajectory points of each cycle of the target vehicle to obtain the filtered horizontal coordinate points corresponding to the target vehicle, and obtaining the historical trajectory points of the target vehicle corresponding to the target vehicle based on the filtered horizontal coordinate points; Performing polynomial curve fitting based on the historical trajectory points of the target vehicle to obtain the second trajectory model; The third trajectory model is determined based on the rate of change of the lane centerline curvature, the lane centerline curvature, the lane centerline heading angle, and the lane centerline offset distance determined by the visual sensor.

3. The method for determining the lane centerline according to claim 2, characterized in that: Before filtering the horizontal coordinate points of the historical trajectory points of the target vehicle in each cycle to obtain the filtered horizontal coordinate points corresponding to the target vehicle, and obtaining the historical trajectory points of the target vehicle corresponding to the target vehicle based on the filtered horizontal coordinate points, the method further includes: Selecting a preset number of vehicles as the target vehicles based on a target vehicle selection rule; wherein the target vehicle selection rule is a rule determined based on the coordinate values ​​of the vehicles and the time that the vehicles are within the field of view of the self-vehicle; The target vehicle selection rule includes: The longitudinal coordinate value of the vehicle is within the set longitudinal coordinate value range; The difference between the lateral coordinate value of the vehicle and the lateral coordinate value of the ego vehicle is within the set difference threshold; The time that the vehicle is in the field of view of the vehicle is greater than the set time threshold; The standard deviation of the lateral coordinate values ​​of the vehicle is less than the set standard deviation threshold.

4. The method for determining the lane centerline according to claim 1, characterized in that: The determining and selecting at least one of the first trajectory model, the second trajectory model, and the third trajectory model as the target model includes: determining to use the first trajectory model, the second trajectory model and the third trajectory model as the target models; Accordingly, determining a first transverse coordinate point corresponding to the coordinate of the first longitudinal sampling point based on the target model, and determining a transverse coordinate point weight corresponding to the first transverse coordinate point, and determining a coordinate point corresponding to the center line of the forward lane based on the transverse coordinate point weight of the target model and the first transverse coordinate point, including: Determine, based on the first trajectory model, the second trajectory model, and the third trajectory model, a first lateral coordinate point of the vehicle, a first lateral coordinate point of the target vehicle, and a first lateral coordinate point of the sensor corresponding to the coordinates of the first longitudinal sampling point, respectively; Based on the first transverse coordinate point of the vehicle, the first transverse coordinate point of the target vehicle and the first transverse coordinate point of the sensor, an average value of the first transverse coordinate point corresponding to the coordinate of each first longitudinal sampling point is determined; Based on the first lateral coordinate point of the vehicle, the first lateral coordinate point of the target vehicle, the first lateral coordinate point of the sensor and the average value of the first lateral coordinate point, determining the variance of the lateral coordinate points under three models corresponding to the coordinates of each first longitudinal sampling point; Determine a weighted average value of the first transverse coordinate points corresponding to each first longitudinal sampling point based on the variance of the transverse coordinate points; Determine the horizontal coordinate point weight corresponding to each first horizontal coordinate point based on the weighted average value; Based on the weight of each of the transverse coordinate points, a target first transverse coordinate point corresponding to each first longitudinal sampling point is determined, and based on the first longitudinal sampling point and the target first transverse coordinate point, a coordinate point corresponding to the center line of the forward lane is obtained.

5. The method for determining the lane centerline according to claim 1, characterized in that: Before determining the first horizontal coordinate point corresponding to the first longitudinal sampling point coordinate based on the target model and determining the horizontal coordinate point weight corresponding to the first horizontal coordinate point, the method further includes: Determine if the vision sensor can determine the end point of the lane centerline; When the end point of the lane centerline can be determined, setting a maximum point of the coordinates of the first longitudinal sampling point based on the end point determined by the visual sensor; When the end point of the lane centerline cannot be determined, the end point of the lane centerline is calculated by linear interpolation of the vehicle speed, and the maximum point of the coordinates of the first longitudinal sampling point is set based on the end point determined by linear interpolation of the vehicle speed.

6. The method for determining the lane centerline according to claim 1, characterized in that: The process of determining the coordinates of the second longitudinal sampling point includes: Determine the minimum point of the center line of the rear lane; The coordinates of the second longitudinal sampling point are determined based on the minimum point.

7. The method for determining the lane centerline according to claim 1, characterized in that: The process of determining the fourth trajectory model includes: Get a set number of historical trajectory points of the vehicle; Selecting a target vehicle historical trajectory point from the set number of vehicle historical trajectory points based on the vehicle historical trajectory point selection condition; Performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the target vehicle to obtain the fourth trajectory model; The conditions for selecting the historical trajectory points of the vehicle include: The point whose horizontal coordinate of the trajectory point is greater than the set horizontal coordinate threshold; The vertical coordinate at the current moment is greater than the vertical coordinate at the previous moment; The vertical coordinate at the current moment is less than zero; The longitudinal coordinate at the current moment is less than the minimum value of the longitudinal coordinate of the historical trajectory point of the vehicle; The vertical coordinate at the current moment is not equal to the vertical coordinate at the previous moment; The difference between the absolute values ​​of the horizontal coordinates at the previous moment and the current moment is greater than the set absolute value difference threshold.

8. The method for determining the lane centerline according to any one of claims 1 to 7, characterized in that: Determining the target lane centerline based on the coordinate point corresponding to the centerline of the forward lane and the coordinate point corresponding to the centerline of the backward lane includes: Determining a front lane centerline model based on the coordinate points corresponding to the front lane centerline; Determine a rear lane centerline model based on the coordinate points corresponding to the rear lane centerline; The target lane centerline is determined based on the forward lane centerline model and the backward lane centerline model.

9. A device for determining a lane centerline, characterized in that: include: A target model determination module is used to determine and select at least one of the first trajectory model, the second trajectory model and the third trajectory model as the target model; wherein the first trajectory model is a first trajectory model determined based on the curvature of the current trajectory of the vehicle; the second trajectory model is a second trajectory model obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the target vehicle; and the third trajectory model is a lane centerline model determined based on a visual sensor; a first information determination module, configured to determine a first transverse coordinate point corresponding to the coordinates of the first longitudinal sampling point based on a target model, determine a transverse coordinate point weight corresponding to the first transverse coordinate point, and determine a coordinate point corresponding to a center line of a forward lane based on the transverse coordinate point weight of the target model and the first transverse coordinate point; a second information determination module, configured to filter the second lateral coordinate point determined based on the fourth trajectory model and the second longitudinal sampling point coordinates to obtain the filtered second lateral coordinate point, and determine the coordinate point corresponding to the center line of the rear lane based on the second longitudinal sampling point coordinates and the filtered second lateral coordinate point; wherein the fourth trajectory model is a trajectory model obtained by performing polynomial fitting based on the coordinates corresponding to the historical trajectory points of the vehicle; The third information determination module is used to determine the target lane centerline based on the coordinate point corresponding to the forward lane centerline and the coordinate point corresponding to the backward lane centerline.

10. A lane centerline determination device, characterized in that: include: Memory for storing computer programs; A processor, configured to execute the computer program to implement the steps of the lane centerline determination method as described in any one of claims 1 to 8.

11. A readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for determining the lane center line according to any one of claims 1 to 8 are implemented.

Images