Lane center reference line generation method and device, electronic equipment and storage medium

Abstract

The invention relates to a lane center reference line generation method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a lane line historical queue, a lane width historical queue and a lane center line historical queue based on preprocessed lane parameter information; a left lane weight and a right lane weight are obtained according to the lane line historical queue, the lane width historical queue and the lane center line historical queue, so that the stability of the lane lines can be judged according to the left lane weight and the right lane weight, and then a target sampling scheme is determined according to the stable lane lines. The method comprises the steps of determining a target sampling scheme of a lane line, determining a lane center line according to the target sampling scheme, obtaining an observation result on the basis of the lane center line, updating the observation result by using Kalman filtering, and correcting a target coefficient of a multi-time polynomial according to the updating result to obtain a final center reference line. The final center reference line can be more accurate, and different types of road conditions can be stably processed.

Description

Technical Field

[0001] The present application relates to the field of automatic driving assistance technology, and in particular to a lane center reference line generation method, device, electronic device and storage medium. Background Art

[0002] The lane center reference line indicates the center line of a lane or road. In the field of autonomous driving, autonomous driving planning and control generally use the center reference line as the basis, so how to generate an accurate center reference line is crucial for autonomous driving. Currently, the generation of the center reference line is generally based on the lane line detected by vision, but the reliability of the lane line is not considered, resulting in a certain error in the generated center reference line. Summary of the invention

[0003] In order to solve the above technical problems or at least partially solve the above technical problems, the present application provides a lane center reference line generation method, device, electronic device and storage medium.

[0004] In a first aspect, the present application provides a method for generating a lane center reference line, the method comprising:

[0005] According to the lane parameter information collected by the vehicle's forward vision, an initial center reference line for the vehicle's travel is obtained; wherein the initial center reference line is expressed based on a multi-order polynomial and a target coefficient;

[0006] Preprocessing the lane parameter information to obtain a lane line history queue, a lane width history queue, and a lane center line history queue;

[0007] Calculating a lane line weight based on the lane line history queue, the lane width history queue and the lane center line history queue; wherein the lane line weight includes a left lane weight and a right lane weight;

[0008] Determining a target sampling scheme from a single-line sampling scheme and a double-line sampling scheme according to the left lane weight and the right lane weight;

[0009] Determine a lane centerline according to the target sampling scheme, and obtain an observation result based on the lane centerline; wherein the observation result includes a plurality of centerline observation values;

[0010] The observation results are updated using Kalman filtering, and the target coefficients of the multi-order polynomial are corrected according to the updated results to obtain the final center reference line.

[0011] Optionally, calculating a lane line weight based on the lane line history queue, the lane width history queue and the lane center line history queue includes:

[0012] Determining a lane influence factor according to the lane line history queue, the lane width history queue and the lane center line history queue; wherein the lane influence factor includes at least two of a lane confidence influence factor, a lane parallelism influence factor, a lane intersection influence factor, a lane length influence factor, a lane curvature influence factor, a lane lateral deviation influence factor and a lane overwidth influence factor;

[0013] The lane line weight is determined according to the product of a plurality of lane influence factors.

[0014] Optionally, the lane line history queue includes a plurality of queue elements represented by lane line information structures, wherein the lane line information structure includes at least a lane line starting point, a lane line end point, a lane line confidence, a lateral deviation from a vehicle coordinate, an angle with a vehicle heading, and a lane line curvature;

[0015] Determining a lane influencing factor according to the lane line history queue, the lane width history queue, and the lane center line history queue includes:

[0016] If the lane line confidence is greater than a preset credibility value, determining the lane confidence influencing factor according to the lane line confidence;

[0017] Calculating the parallelism between the lane line and the historical center line in the lane center line history queue, and determining the lane parallelism influencing factor according to the parallelism;

[0018] Determine the lane intersection influence factor according to the difference between the lateral deviation of the left lane line from the vehicle coordinates and the lateral deviation of the right lane line from the vehicle coordinates;

[0019] Determine the length of the left lane line according to the lane line starting point of the left lane and the lane line end point of the left lane, determine the length of the right lane line according to the lane line starting point of the right lane and the lane line end point of the right lane, and determine the lane length influencing factor according to the left lane line length and the right lane line length;

[0020] Determining the lane curvature influencing factor according to the left lane curvature and the right lane curvature;

[0021] Determine the lane lateral deviation influence factor according to the lateral deviation of the vehicle coordinates from the left lane line and the lateral deviation of the vehicle coordinates from the right lane line;

[0022] The lane overwidth influence factor is determined according to the difference between the lane line and the historical center line.

[0023] Optionally, if the lane line confidence is greater than a preset credibility value, before determining the lane confidence influencing factor according to the lane line confidence, the method further includes:

[0024] The parameters in the lane line information structure are filtered according to a preset filtering rule, and the lane influence factors corresponding to the filtered parameters are set to 0.

[0025] Optionally, the observation results are updated by using Kalman filtering, and the target coefficients of the multi-order polynomial are corrected according to the updated results to obtain the final center reference line, including:

[0026] Determine an observation matrix and an observation noise matrix according to the observation results;

[0027] Calculate the distance change and angle change of the vehicle corresponding to two adjacent observation values;

[0028] Determine the state quantity of the reference line according to the observation matrix, the observation noise matrix, the distance change value and the angle change value;

[0029] Predicting a covariance matrix and a state vector of the state quantity;

[0030] Based on Kalman filtering, the covariance matrix and the state vector are updated to determine the target coefficients of the multi-order polynomial;

[0031] The final center reference line is determined according to the target coefficient and the multi-order polynomial.

[0032] Optionally, determining a target sampling scheme from a single-line sampling scheme and a double-line sampling scheme according to the left lane weight and the right lane weight includes:

[0033] If the left lane weight and the right lane weight are both greater than a first preset threshold, and the deviation between the left lane weight and the right lane weight is less than a second preset threshold, determining that the target sampling scheme is a two-line sampling scheme;

[0034] If the left lane weight and the right lane weight are both greater than the first preset threshold, and the deviation between the left lane weight and the right lane weight is greater than or equal to the second preset threshold, it is determined that the target sampling scheme is a single-line sampling scheme.

[0035] Optionally, the method further comprises:

[0036] If the left lane weight and the right lane weight are both less than or equal to the first preset threshold, obtaining a distance change value and an angle change value of the vehicle movement within a preset time period;

[0037] Determine the state quantity and control matrix;

[0038] Determine a state transfer matrix according to the state quantity and the distance change value;

[0039] Determine a control amount according to the control matrix and the angle change value;

[0040] Determine a prediction equation according to the state quantity, the state transfer matrix, the control matrix and the control quantity;

[0041] Predicting the target coefficient according to the prediction equation;

[0042] The final center reference line is determined according to the target coefficient and the multi-order polynomial.

[0043] In a second aspect, the present application provides a lane center reference line generating device, the device comprising:

[0044] An initial reference line determination module is used to obtain an initial center reference line for the vehicle based on lane parameter information collected by the vehicle's forward vision; wherein the initial center reference line is expressed based on a multi-order polynomial and a target coefficient;

[0045] A preprocessing module, used to preprocess the lane parameter information to obtain a lane line history queue, a lane width history queue and a lane center line history queue;

[0046] A weight calculation module, used to calculate lane line weights based on the lane line history queue, the lane width history queue and the lane center line history queue; wherein the lane line weights include left lane weights and right lane weights;

[0047] A sampling scheme determination module, used to determine a target sampling scheme from a single-line sampling scheme and a double-line sampling scheme according to the left lane weight and the right lane weight;

[0048] An observation module, used to determine the lane centerline according to the target sampling scheme, and obtain an observation result based on the lane centerline; wherein the observation result includes a plurality of centerline observation values;

[0049] The generation module is used to update the observation results using Kalman filtering, and to correct the target coefficients of the multi-order polynomial according to the update results to obtain the final center reference line.

[0050] In a third aspect, the present application provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0051] Memory, used to store computer programs;

[0052] The processor is used to implement the steps of the lane center reference line generation method described in any embodiment of the first aspect when executing the program stored in the memory.

[0053] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the lane center reference line generation method as described in any one of the embodiments of the first aspect.

[0054] Beneficial effects of this application:

[0055] The method provided in the embodiment of the present application obtains an initial center reference line for vehicle travel based on lane parameter information collected by the vehicle's forward vision; wherein the initial center reference line is expressed based on a multi-order polynomial and a target coefficient; the lane parameter information is preprocessed to obtain a lane line history queue, a lane width history queue and a lane center line history queue; the lane line weight is calculated based on the lane line history queue, the lane width history queue and the lane center line history queue; wherein the lane line weight includes a left lane weight and a right lane weight; according to the left lane weight and the right lane weight, a target sampling scheme is determined from a single-line sampling scheme and a double-line sampling scheme; according to the target sampling scheme, the lane center line is determined, and an observation result is obtained based on the lane center line; wherein the observation result includes multiple center line observation values; the observation result is updated using a Kalman filter, and the target coefficient of the multi-order polynomial is corrected according to the updated result to obtain a final center reference line. The method can obtain a lane line history queue, a lane width history queue and a lane center line history queue based on preprocessed lane parameter information, and obtain a left lane weight and a right lane weight according to the lane line history queue, the lane width history queue and the lane center line history queue, so that the stability of the lane line can be judged according to the left lane weight and the right lane weight, and then determine a target sampling scheme according to the stable lane line, and then determine the lane center line according to the target sampling scheme, and obtain observation results based on the lane center line, use Kalman filtering to update the observation results, and correct the target coefficients of the multi-time polynomial according to the updated results to obtain the final center reference line. Due to the target sampling scheme determined based on the stable lane line, the final center reference line can be made more accurate, and different types of road conditions can be stably handled. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0057] In order to more clearly illustrate the embodiments of the present application 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, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0058] Figure 1A system architecture diagram of a lane center reference line generation method provided in one embodiment of the present application;

[0059] Figure 2 A schematic diagram of a process for generating a lane center reference line provided in one embodiment of the present application;

[0060] Figure 3 A logic flow chart of a lane center reference line generation method provided in one embodiment of the present application;

[0061] Figure 4 A schematic diagram of a lane selection process provided for an embodiment of the present application;

[0062] Figure 5 A schematic diagram of the structure of a lane center reference line generating device provided in one embodiment of the present application;

[0063] Figure 6 A schematic diagram of the structure of an electronic device provided in accordance with an embodiment of the present application. DETAILED DESCRIPTION

[0064] The following will describe the implementation methods of the present application with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific implementation methods, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be understood that the preferred embodiments are only for illustrating the present application, not for limiting the scope of protection of the present application.

[0065] The first embodiment of the present application provides a method for generating a lane center reference line. The method can be applied to Figure 1 The system architecture shown in the figure includes at least a data acquisition module 101 and a data processing module 102, and the data acquisition module 101 and the data processing module 102 establish a communication connection. Specifically, the system architecture can be a vehicle, and the type of vehicle is not limited, for example, it can be a fuel vehicle, a pure electric vehicle, a hybrid vehicle or a fuel cell vehicle, etc.

[0066] Next, based on the system architecture, the lane center reference line generation method is described in detail. Figure 2 , the lane center reference line generation method includes:

[0067] Step 201, obtaining an initial center reference line for vehicle travel based on lane parameter information collected by the vehicle's forward vision; wherein the initial center reference line is expressed based on a multi-order polynomial and a target coefficient.

[0068] The vehicle front view refers to the camera and radar configured in front of the vehicle to collect lane parameter information. The center position of the front end of the vehicle is established as the origin, the vehicle's forward direction is the x-axis, and the direction perpendicular to the vehicle's forward direction to the left is the y-axis. The initial center reference line can be expressed by the following cubic polynomial:

[0069]

[0070] Among them, when the multinomial is a cubic polynomial, the target coefficients include c0, c1, c2, and c3. c0 represents the lateral deviation value between the vehicle and the center line of the road, c1 represents the angle between the vehicle and the center line of the road, c2 represents the lane curvature, and c3 represents the curvature change rate. The more accurate the target coefficient is, the more accurate the calculated center reference line is.

[0071] Step 202 , pre-processing the lane parameter information to obtain a lane line history queue, a lane width history queue, and a lane center line history queue.

[0072] The sampling period can be set, for example, the sampling period is set to 50ms, that is, the system samples every 50ms to obtain the lane parameter information under the current time frame, pre-processes the lane parameter information under the current frame, and stores it in the parameter history queue according to parameter classification, so as to obtain the lane line history queue, lane width history queue width_queue and lane center line history queue ref_line_queue, etc. Specifically, the lane line history queue may include the left lane line history queue reLine_queue and the right lane line history queue riLine_queue. The lane parameter information collected in each time frame can be stored in the queue as a queue element. The lane line history queue may include all the queue elements collected by the lane line information structure. The lane line information structure may include the lane line starting point start_x, the lane line end_x, the lane line confidence line_conf, the lateral deviation from the vehicle coordinate line_c0, the angle with the vehicle heading line_c1 and the lane line curvature line_c2, etc. Of course, it may also include the lane line curvature change rate line_c3, etc. The lane line confidence line_conf is a constant between 0 and 1. The larger the value, the higher the confidence.

[0073] Step 203, calculating lane line weights based on the lane line history queue, the lane width history queue and the lane center line history queue; wherein the lane line weights include left lane weights and right lane weights.

[0074] In one embodiment, the lane line weight is calculated based on the lane line history queue, the lane width history queue and the lane center line history queue, including: determining the lane influence factor according to the lane line history queue, the lane width history queue and the lane center line history queue; wherein the lane influence factor includes at least two of the lane confidence influence factor, the lane parallelism influence factor, the lane intersection influence factor, the lane length influence factor, the lane curvature influence factor, the lane lateral deviation influence factor and the lane overwidth influence factor; determining the lane line weight according to the product of multiple lane influence factors.

[0075] In this embodiment, lane influence factors can be determined according to the lane line history queue, lane width history queue and lane centerline history queue, such as lane confidence influence factor kconf, lane parallelism influence factor kp, lane intersection influence factor kx, lane length influence factor krange, lane curvature influence factor kc, lane lateral deviation influence factor kd and lane excess width influence factor kw, etc. The lane line weight k can be determined according to the product of the obtained multiple lane influence factors, such as respectively obtaining the left lane line weight and the right lane line weight, so that the stability of the left lane line and the right lane line can be judged according to the obtained left lane weight and right lane weight.

[0076] In one embodiment, determining a lane influencing factor based on a lane line history queue, a lane width history queue, and a lane center line history queue includes:

[0077] If the lane line confidence is greater than the preset credibility value, the lane confidence influence factor is determined according to the lane line confidence. If the lane line confidence coefficient is greater than the preset credibility value (ie, the lowest credible confidence value), the lane confidence influence factor kconf can be set to be equal to the lane line confidence line_conf.

[0078] Calculate the parallelism between the lane line and the historical center line in the lane center line history queue, and determine the lane parallelism influencing factor based on the parallelism. Calculate the parallelism between the lane line and the historical center line, fit a historical reference line equation by historical information in the historical queue, and calculate the longitudinal movement distance △x between two adjacent moments in the historical queue in sequence. i , and according to the currently acquired lane line parameters, take the start_x point as the starting point, and the interval △x i Select points on the two lines. The number of points can be set to a maximum of 30, and the minimum must be greater than 2. Press the vertical △x i Distance calculation: the lateral difference between the lane line and the historical center trajectory △y i , and divide it into segments with slopes △y i / △x iStore the deviation array dev_list, calculate the standard deviation δ of dev_list, and if δ is smaller, the parallelism is better. Set the parallelism influence factor kp=1 for the side with better parallelism in the left and right lane lines, and kp=0.8 for the side with worse parallelism.

[0079] The lane intersection influence factor is determined based on the difference between the lateral deviation of the left lane line and the lateral deviation of the right lane line. To determine the intersection and merging trend, according to the lane line parameters, perform a preview every 5m and calculate the line_c0 difference of the lane lines on both sides. If the difference is less than or equal to 0, it means that there is a convergence point in the lane lines. At this time, the lane line on the side with poor parallelism is set to have its lane intersection influence factor kx=0, and the side with good parallelism is set to kx=1.

[0080] The length of the left lane is determined according to the starting point of the left lane and the end point of the left lane, the length of the right lane is determined according to the starting point of the right lane and the end point of the right lane, and the lane length influence factor is determined according to the length of the left lane and the length of the right lane. The lane length influence factor krange is calculated, where the lane length can be expressed as (end_x-start_x), and the length of the longer lane of the left and right lanes is max_length, and finally krange = (end_x-start_x) / max_length.

[0081] The lane curvature influence factor is determined based on the curvature of the left lane and the curvature of the right lane. The curvature influence factor kc is calculated. If the curvature of the lane line on one side is greater than 1.2 times the curvature of the lane line on the other side, the lane curvature influence factor kc on that side is 1 / 1.2, and kc on the other side is 1.

[0082] The lane lateral deviation influence factor is determined according to the lateral deviation of the vehicle coordinates from the left lane line and the lateral deviation of the vehicle coordinates from the right lane line. The lateral deviation is denoted as dev, and the side with a larger lateral deviation is denoted as max_dev. Then, the lateral deviation influence factor kd=1 is calculated for the side with a smaller lateral deviation, and kd=dev / max_dev for the side with a larger lateral deviation.

[0083] Determine the lane over-width influence factor based on the difference between the lane line and the historical center line. Calculate the over-width lane and widened lane influence factor kw. According to the historical center trajectory, preview three points at 5m intervals forward, and calculate the deviation between the left and right lane lines and the center line in turn. If the deviation value is greater than 4 meters, the lane line of the vehicle is judged to be widened or over-width, and the lane over-width influence factor kw of the lane line on this side is set to 0, and the lane over-width influence factor kw of the lane line on the non-widened side is set to 1.

[0084] In this embodiment, the lane confidence factor kconf, lane parallelism factor kp, lane intersection factor kx, lane length factor krange, lane curvature factor kc, lane lateral deviation factor kd, and lane overwidth factor kw of the left and right lanes are calculated respectively, so that the stability of the left and right lanes can be determined by combining multi-dimensional parameters. For example, the final lane weight is obtained by multiplying each influencing factor, that is, k = kconf*kp*kx*krange*kc*kd*kw.

[0085] In one embodiment, if the lane line confidence is greater than a preset credibility value, before determining the lane confidence influence factor according to the lane line confidence, the method also includes: filtering the parameters in the lane line information structure according to preset filtering rules, and setting the lane influence factor corresponding to the filtered parameters to 0.

[0086] In this embodiment, the obviously unreasonable lane line values ​​can be filtered out first, and their weights can be reset to 0 to ensure that the line will not be selected. For example, if a lane line on a certain side does not exist, the weight of the lane line on that side is directly reset to 0. In addition, if the line_c0 jump exceeds 1.2 times the amplitude, the lane line start_x is greater than 20 meters, and the lane line confidence line_conf is lower than 0.4, it can also be regarded as an unavailable state, and its weight is also set to 0 to ensure that a more stable and reliable lane line on one side is selected. If the lane lines on both sides are unavailable, their weights are set to 0. If the influence factor of a certain line is set to 0, the weight of the lane line on that side must also be greater than 0, so that there is no need to calculate other influence factors, reducing the amount of calculation and improving the response speed.

[0087] Step 204 , determining a target sampling scheme from a single-line sampling scheme and a double-line sampling scheme according to the left lane weight and the right lane weight.

[0088] In one embodiment, a target sampling scheme is determined from a single-line sampling scheme and a double-line sampling scheme based on the left lane weight and the right lane weight, including: if the left lane weight and the right lane weight are both greater than a first preset threshold, and the deviation between the left lane weight and the right lane weight is less than a second preset threshold, determining that the target sampling scheme is a double-line sampling scheme; if the left lane weight and the right lane weight are both greater than the first preset threshold, and the deviation between the left lane weight and the right lane weight is greater than or equal to the second preset threshold, determining that the target sampling scheme is a single-line sampling scheme.

[0089] In this embodiment, the first preset threshold value may be, for example, 0, and the second preset threshold value may be, for example, 0.1, that is, if the left lane weight and the right lane weight are both greater than 0, and the weight difference between the left and right lane lines is within 0.1, a two-line sampling scheme is adopted. If the left lane weight and the right lane weight are both greater than 0, and the weight difference between the left and right lane lines is greater than 0.1, a single-line sampling is performed based on the lane line on the side with a higher weight.

[0090] Furthermore, if the left lane weight and the right lane weight are both less than or equal to the first preset threshold value, obtain the distance change value and the angle change value of the vehicle movement within the preset time length; determine the state quantity and the control matrix; determine the state transfer matrix according to the state quantity and the distance change value; determine the control quantity according to the control matrix and the angle change value; determine the prediction equation according to the state quantity, the state transfer matrix, the control matrix and the control quantity; predict the target coefficient according to the prediction equation; determine the final center reference line according to the target coefficient and the multi-order polynomial.

[0091] In this embodiment, if the lane lines on both sides are unavailable or cannot be accurately identified for some reason, the driving center line will be predicted based on the historical driving trajectory. For example, the lane center line trajectory parameters within 5 seconds can be set when the lane line disappears or is unavailable. The distance change value dis and angle change value anq caused by the vehicle movement within 5 seconds are obtained. Combined with the prediction equation, the prediction of various parameters of the lane center line is realized.

[0092] Specifically, the initial center reference line equation is as follows:

[0093]

[0094] The state quantity is: X = [c0 c12 × c26 × c3] T

[0095] State transition matrix:

[0096] Control Matrix:

[0097] Control volume:

[0098] Then, the prediction equation is: X p =F×X+B×U

[0099] Where F is the state transfer matrix, B is the control matrix, and U is the control amount. Through the prediction process, the c0 prediction value of the initial center reference line equation can be obtained:

[0100]

[0101] The c1 prediction value of the lane line equation:

[0102] c1=c1+2dis*c2+3dis 2 *c3+dis 3 -anq

[0103] The c2 prediction value of the lane line equation:

[0104] c2=c2+3dis*c3

[0105] Then, the obtained target coefficient value is substituted into the initial center reference line equation to obtain the final center reference line.

[0106] Step 205, determine the lane centerline according to the target sampling scheme, and obtain an observation result based on the lane centerline; wherein the observation result includes multiple centerline observation values.

[0107] When a single-line sampling scheme is selected as the line selection result, the system will first offset the selected lane line by half the lane width to determine the position of the lane centerline, and then observe based on the lane centerline to obtain the observation result.

[0108] When the line selection result is a two-line sampling scheme, the overlapping part between the two lane lines is calculated according to the start_x and end_x of the lane line, and the lane center line is determined, and then observation is performed based on the lane center line to obtain the observation result.

[0109] Step 206, using Kalman filtering to update the observation results, and correcting the target coefficients of the polynomial according to the updated results to obtain the final center reference line.

[0110] The method can obtain a lane line history queue, a lane width history queue and a lane center line history queue based on preprocessed lane parameter information, and obtain a left lane weight and a right lane weight according to the lane line history queue, the lane width history queue and the lane center line history queue, so that the stability of the lane line can be judged according to the left lane weight and the right lane weight, and then determine a target sampling scheme according to the stable lane line, and then determine the lane center line according to the target sampling scheme, and obtain observation results based on the lane center line, use Kalman filtering to update the observation results, and correct the target coefficients of the multi-time polynomial according to the updated results to obtain the final center reference line. Due to the target sampling scheme determined based on the stable lane line, the final center reference line can be made more accurate, and different types of road conditions can be stably handled.

[0111] In one embodiment, the observation results are updated using Kalman filtering, and the target coefficients of the polynomial are corrected based on the updated results to obtain the final center reference line, including: determining the observation matrix and the observation noise matrix based on the observation results; calculating the distance change value and the angle change value of the vehicle corresponding to two adjacent observation values; determining the state quantity of the reference line based on the observation matrix, the observation noise matrix, the distance change value and the angle change value; predicting the covariance matrix and the state vector of the state quantity; updating the covariance matrix and the state vector based on Kalman filtering to determine the target coefficients of the polynomial; and determining the final center reference line based on the target coefficients and the multi-order polynomial.

[0112] In this embodiment, the observation quantity y is an n-order diagonal matrix composed of the lateral positions of all sampling points. The observation matrix is:

[0113]

[0114] The observation noise matrix is:

[0115]

[0116] According to the vehicle history information queue, the distance dis moved and the changed heading angle anq within the two sampling cycles are calculated.

[0117] The state quantity of the output reference line is X lane =[c0, c1, c2, c3], the state transfer equation between time t and time t+1 can be obtained as:

[0118]

[0119] c 2，t+1 =c 2，t +Δdis×c 3，t

[0120] Further use the Kalman prediction formula to predict the state quantity X lane And the covariance matrix p is predicted as follows:

[0121]

[0122] where x k represents the system state vector, x k-1 is the state vector of the system at the previous moment, A is the state transfer matrix that converts the state at time t-1 to the state at time t, and B is the control input matrix that converts the motion measurement value u k-1 The effect of is mapped to the state vector, p represents the covariance matrix, which represents the uncertainty of the system, and Q is the process noise covariance matrix.

[0123] After the above prediction process is completed, the Kalman filter update formula is used to perform Kalman update, and finally the entire Kalman filter process is completed. Substitute the calculated target coefficient into the initial center reference line equation to obtain the final output center reference line as follows:

[0124]

[0125] In this embodiment, the selection weight of unreasonable lane lines and those with poor stability is reduced, thereby ensuring the stability of the center line. In addition, the calculation and estimation of the width of the current lane and the lane ahead can ensure stable driving in the lane ahead. The lane line weight is calculated by combining a variety of factors that affect the stability of the center reference line, thereby improving the stability of the center reference line.

[0126] In a specific embodiment, the logic flow chart of the lane center reference line generation method is as follows: Figure 3 ,include:

[0127] Step S1, obtain the forward-looking lane line and initialize the center reference line.

[0128] Get the forward perception lane line, establish the center position of the front end of the vehicle as the origin, the vehicle's forward direction as the x-axis, and the direction perpendicular to the vehicle's forward direction to the left as the y-axis. The initial center reference line can be expressed by the following cubic polynomial:

[0129]

[0130] When the multinomial is a cubic polynomial, the target coefficients include c0, c1, c2, and c3. c0 represents the lateral deviation between the vehicle and the center line of the road, c1 represents the angle between the vehicle and the center line of the road, c2 represents the lane curvature, and c3 represents the curvature change rate.

[0131] Step S2: lane line preprocessing, setting lane line and center reference line history queue and lane width queue, and setting sampling period.

[0132] First, initialize the historical queue, set the centerline historical queue ref_line_queue, set the left and right lane line historical queues as reLine_queue and riLine_queue respectively, and the lane width historical queue width_queue. Each element in the queue is a structure containing all the information of that element. Taking the lane line as an example, the lane line information is output in the form of lane line parameters in the forward view. The element information contained in the lane line information structure mainly includes parameters such as the start point start_x of the lane line, the end point end_x of the lane line, the lane line confidence line_conf, the lateral deviation line_c0 from the vehicle coordinate, the angle line_c1 with the vehicle heading, the lane line curvature line_c2, and the lane line curvature change rate line_c3. The sampling period can be set. For example, set the sampling period to 50 ms, that is, the system samples once every 50 ms to obtain the lane parameter information in the current time frame.

[0133] Step S3, update the historical sampling points and calculate the lane width.

[0134] Obtain the new sampling points, calculate and update the lane width. Set the current lane width curr_width as the lane width formed by the selected left and right lines at the current position, and set the forward lane width fro_width as the lane width formed by the selected left and right lines at 2*speed ahead, where speed is the current vehicle speed. Denote the historical lane width of the previous sampling period as his_width.

[0135] If there are lane lines on both sides, and fro_width>curr_width and fro_width>4.0 are satisfied, it is considered that the forward lane is widened to an extra-wide lane. At this time, set the lane width curr_width as the historical lane width.

[0136] If there are lane lines on both sides, and 2.8<fro_width<1.2*curr_width and fro_width<4.0 are satisfied, it is considered that the vehicle is currently driving in a normal lane. At this time, set the lane width curr_width as the absolute value of the difference between the left and right lane lines line_c0.

[0137] If there is only a single-side lane line or the lane line disappears, then set the lane width curr_width as the historical lane width.

[0138] To ensure the smoothness of the lane width change, perform first-order low-pass filtering on the lane width curr_width to obtain the lane width width and store it in the historical lane width queue width_queue.

[0139] Step S4: Set the lane line stability weight. Ensure that the initial weights are the same. For example, set the lane line initial weight k, where the left lane line initial weight k l =1, initial weight k of the right lane line r =1.

[0140] Step S5: Lane line selection: A reasonable lane line sampling scheme is selected to make the center line calculation result more stable.

[0141] The lane selection process is shown in the figure below: Figure 4 , the multi-factor weight accumulation method is used to calculate the lane line selectable weight, and further obtain the selectable lane lines, including:

[0142] Filter out obviously unreasonable lane lines and reset their weights to 0. First, filter out obviously unreasonable lane line values ​​and reset their weights to 0 to ensure that the line will not be selected. For example, if a lane line on a certain side does not exist, the weight of the lane line on that side is directly reset to 0. In addition, if the line_c0 jump exceeds 1.2 times the amplitude, the lane line start_x is greater than 20 meters, and the lane line confidence line_conf is less than 0.4, it can also be regarded as an unavailable state. Similarly, its weight is set to 0 to ensure that a more stable and reliable lane line on one side is selected. If both sides of the lane line are unavailable, their weights are set to 0. If the influence factor of a certain line is set to 0, the weight of the lane line on that side must also be greater than 0, so there is no need to calculate other influence factors.

[0143] The confidence influence factor is calculated based on the original confidence of the lane line. If the lane line confidence coefficient is greater than the preset confidence value (ie, the lowest credible confidence value), the lane confidence influence factor kconf can be set to be equal to the lane line confidence line_conf.

[0144] Calculate the parallelism between the lane line and the historical center line. Calculate the parallelism between the lane line and the historical center line, fit a historical reference line equation to the historical information in the historical queue, and calculate the longitudinal movement distance △x between two adjacent moments in the historical queue in sequence. i , and according to the currently acquired lane line parameters, take the start_x point as the starting point, and the interval △x i Select points on the two lines. The number of points can be set to a maximum of 30, and the minimum must be greater than 2. Press the vertical △x i Distance calculation: the lateral difference between the lane line and the historical center trajectory △y i , and divide it into segments with slopes △y i / △x i Store the deviation array dev_list, calculate the standard deviation δ of dev_list, and if δ is smaller, the parallelism is better. Set the parallelism influence factor kp=1 for the side with better parallelism in the left and right lane lines, and kp=0.8 for the side with worse parallelism.

[0145] Determine the intersection and merging trend. According to the lane line parameters, perform a preview every 5m and calculate the line_c0 difference of the lane lines on both sides. If the difference is less than or equal to 0, it means that there is a convergence point of the lane lines. At this time, the lane line on the side with poor parallelism is set to have its lane intersection influence factor kx=0, and the side with good parallelism is set to kx=1.

[0146] Calculate the lane line length influence factor. Calculate the lane line length influence factor krange, where the lane line length can be expressed as (end_x-start_x), and the length of the longer lane line on the left and right lane lines is max_length, and finally krange = (end_x-start_x) / max_length.

[0147] Calculate the curvature influence factor. Calculate the curvature influence factor kc. If the curvature of the lane line on one side is greater than 1.2 times the curvature of the lane line on the other side, the lane curvature influence factor kc on that side is 1 / 1.2, and kc on the other side is 1.

[0148] Calculate the lateral deviation influence factor. Let the lateral deviation be dev, and the side with larger lateral deviation be max_dev. Then calculate the lateral deviation influence factor kd=1 on the side with smaller lateral deviation, and kd=dev / max_dev on the side with larger lateral deviation.

[0149] Calculate the extra-wide influence factor. Calculate the extra-wide lane and widened lane influence factor kw. According to the historical center trajectory, preview three points at 5m intervals forward, and calculate the deviation between the left and right lane lines and the center line in turn. If the deviation value is greater than 4 meters, the lane line of the vehicle is judged to be widened or extra-wide, and the lane extra-wide influence factor kw of the lane line on this side is set to 0, and the lane extra-wide influence factor kw of the lane line on the non-widened side is set to 1.

[0150] The final lane line weight is obtained by accumulating various influencing factors. The final lane line weight is obtained by accumulating various influencing factors, that is, k = kconf*kp*kx*krange*kc*kd*kw.

[0151] The lane sampling scheme is selected based on the calculated lane weights. If the left lane weight and the right lane weight are both greater than 0, and the difference between the left and right lane weights is within 0.1, a two-line sampling scheme is adopted. If the left lane weight and the right lane weight are both greater than 0, and the difference between the left and right lane weights is greater than 0.1, a single-line sampling is performed based on the lane with the higher weight.

[0152] Step S6: Lane center reference line calculation: The Kalman filter algorithm is used to update the center line in the single-line sampling, double-line sampling, and lane line disappearance scenarios.

[0153] When a single-line sampling scheme is selected as the line selection result, the system will first offset the selected lane line by half the lane width to determine the position of the lane centerline, and then observe based on the lane centerline to obtain the observation result.

[0154] When the line selection result is a two-line sampling scheme, the overlapping part between the two lane lines is calculated according to the start_x and end_x of the lane line, and the lane center line is determined, and then observation is performed based on the lane center line to obtain the observation result.

[0155] If the lane lines on both sides are unavailable or cannot be accurately identified for some reason, the driving center line will be predicted based on the historical driving trajectory. For example, the lane center line trajectory parameters within 5 seconds can be set when the lane line disappears or is unavailable. The distance change value dis and angle change value anq caused by the vehicle movement within 5 seconds are obtained. Combined with the prediction equation, the parameters of the lane center line are predicted.

[0156] In this embodiment, the selection weight of unreasonable lane lines and those with poor stability is reduced, thereby ensuring the stability of the center line. In addition, the calculation and estimation of the width of the current lane and the lane ahead can ensure stable driving in the lane ahead. The lane line weight is calculated by combining a variety of factors that affect the stability of the center reference line, thereby improving the stability of the center reference line.

[0157] Based on the same technical concept, the second embodiment of the present application provides a lane center reference line generating device, such as Figure 5 , the device comprises:

[0158] The initial reference line determination module 501 is used to obtain the initial center reference line of the vehicle according to the lane parameter information collected by the vehicle's forward vision; wherein the initial center reference line is expressed based on a multi-order polynomial and a target coefficient;

[0159] A preprocessing module 502 is used to preprocess the lane parameter information to obtain a lane line history queue, a lane width history queue and a lane center line history queue;

[0160] A weight calculation module 503, configured to calculate a lane line weight based on the lane line history queue, the lane width history queue and the lane center line history queue; wherein the lane line weight includes a left lane weight and a right lane weight;

[0161] A sampling scheme determination module 504 is used to determine a target sampling scheme from a single-line sampling scheme and a double-line sampling scheme according to the left lane weight and the right lane weight;

[0162] An observation module 505 is used to determine the lane centerline according to the target sampling scheme, and obtain an observation result based on the lane centerline; wherein the observation result includes multiple centerline observation values;

[0163] The generation module 506 is used to update the observation results using Kalman filtering, and to correct the target coefficients of the multi-order polynomial according to the update results to obtain the final center reference line.

[0164] The device can obtain a lane line history queue, a lane width history queue and a lane center line history queue based on preprocessed lane parameter information, and obtain a left lane weight and a right lane weight according to the lane line history queue, the lane width history queue and the lane center line history queue, so that the stability of the lane line can be judged according to the left lane weight and the right lane weight, and then a target sampling scheme is determined according to the stable lane line, and then the lane center line is determined according to the target sampling scheme, and an observation result is obtained based on the lane center line, and the observation result is updated by using Kalman filtering, and the target coefficient of the multi-time polynomial is corrected according to the updated result to obtain a final center reference line. Due to the target sampling scheme determined based on the stable lane line, the final center reference line can be made more accurate, and different types of road conditions can be stably processed.

[0165] like Figure 6 As shown, the third embodiment of the present application provides an electronic device, including a processor 111, a communication interface 112, a memory 113 and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114.

[0166] Memory 113, used for storing computer programs;

[0167] In one embodiment, the processor 111 is used to implement the lane center reference line generation method provided by any one of the aforementioned method embodiments when executing the program stored in the memory 113 .

[0168] The memory and processor in the above electronic device communicate through the communication bus and the communication interface. The communication bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The communication bus can be divided into an address bus, a data bus, a control bus, etc.

[0169] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0170] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0171] A fourth embodiment of the present application provides a computer-readable medium having a non-volatile program code executable by a processor.

[0172] Optionally, in an embodiment of the present application, a computer-readable medium is configured to store program code for a processor to execute the above method.

[0173] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, and this embodiment will not be described in detail here.

[0174] When the embodiments of the present application are specifically implemented, reference may be made to the above-mentioned embodiments, which have corresponding technical effects.

[0175] It is understood that the embodiments described herein may be implemented in hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit may be implemented in one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), general purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions of the present application, or a combination thereof.

[0176] For software implementation, the technology of this article can be implemented by a unit that performs the functions of this article. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or outside the processor.

[0177] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can 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 this application.

[0178] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0179] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0180] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0181] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0182] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks, or optical disks.

[0183] It should be noted that, in this article, relational terms such as "first" and "second" are only used 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", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including 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. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0184] The above embodiments are only preferred embodiments for fully illustrating the present application, and the protection scope of the present application is not limited thereto. Any equivalent substitution or change made by a person skilled in the art based on the present application is within the protection scope of the present application.

Claims

1. A method for generating a lane center reference line, characterized in that: The method comprises: According to the lane parameter information collected by the vehicle's forward vision, an initial center reference line for the vehicle's travel is obtained; wherein the initial center reference line is expressed based on a multi-order polynomial and a target coefficient; Preprocessing the lane parameter information to obtain a lane line history queue, a lane width history queue, and a lane center line history queue; Calculating a lane line weight based on the lane line history queue, the lane width history queue and the lane center line history queue; wherein the lane line weight includes a left lane weight and a right lane weight; Determining a target sampling scheme from a single-line sampling scheme and a double-line sampling scheme according to the left lane weight and the right lane weight; Determine a lane centerline according to the target sampling scheme, and obtain an observation result based on the lane centerline; wherein the observation result includes a plurality of centerline observation values; The observation results are updated using Kalman filtering, and the target coefficients of the multi-order polynomial are corrected according to the updated results to obtain the final center reference line.

2. The method according to claim 1, characterized in that Calculating a lane line weight based on the lane line history queue, the lane width history queue, and the lane center line history queue includes: Determine a lane influence factor according to the lane line history queue, the lane width history queue and the lane center line history queue; wherein the lane influence factor includes at least two of a lane confidence influence factor, a lane parallelism influence factor, a lane intersection influence factor, a lane length influence factor, a lane curvature influence factor, a lane lateral deviation influence factor and a lane overwidth influence factor; The lane line weight is determined according to the product of a plurality of lane influence factors.

3. The method according to claim 2, characterized in that The lane line history queue includes a plurality of queue elements represented by lane line information structures, wherein the lane line information structure includes at least a lane line starting point, a lane line end point, a lane line confidence, a lateral deviation from a vehicle coordinate, an angle with a vehicle heading, and a lane line curvature; Determining a lane influencing factor according to the lane line history queue, the lane width history queue, and the lane center line history queue includes: If the lane line confidence is greater than a preset credibility value, determining the lane confidence influencing factor according to the lane line confidence; Calculating the parallelism between the lane line and the historical center line in the lane center line history queue, and determining the lane parallelism influencing factor according to the parallelism; Determine the lane intersection influence factor according to the difference between the lateral deviation of the left lane line from the vehicle coordinates and the lateral deviation of the right lane line from the vehicle coordinates; Determine the length of the left lane line according to the lane line starting point of the left lane and the lane line end point of the left lane, determine the length of the right lane line according to the lane line starting point of the right lane and the lane line end point of the right lane, and determine the lane length influencing factor according to the left lane line length and the right lane line length; Determining the lane curvature influencing factor according to the left lane curvature and the right lane curvature; Determine the lane lateral deviation influence factor according to the lateral deviation of the vehicle coordinates from the left lane line and the lateral deviation of the vehicle coordinates from the right lane line; The lane overwidth influence factor is determined according to the difference between the lane line and the historical center line.

4. The method according to claim 3, characterized in that If the lane line confidence is greater than a preset credibility value, before determining the lane confidence influencing factor according to the lane line confidence, the method further includes: The parameters in the lane line information structure are filtered according to a preset filtering rule, and the lane influence factors corresponding to the filtered parameters are set to 0.

5. The method according to claim 1, characterized in that The observation results are updated by using Kalman filtering, and the target coefficients of the multi-order polynomial are corrected according to the updated results to obtain the final center reference line, including: Determine an observation matrix and an observation noise matrix according to the observation results; Calculate the distance change and angle change of the vehicle corresponding to two adjacent observation values; Determine the state quantity of the reference line according to the observation matrix, the observation noise matrix, the distance change value and the angle change value; Predicting a covariance matrix and a state vector of the state quantity; Based on Kalman filtering, the covariance matrix and the state vector are updated to determine the target coefficients of the multi-order polynomial; The final center reference line is determined according to the target coefficient and the multi-order polynomial.

6. The method according to claim 1, characterized in that According to the left lane weight and the right lane weight, a target sampling scheme is determined from a single-line sampling scheme and a double-line sampling scheme, including: If the left lane weight and the right lane weight are both greater than a first preset threshold, and the deviation between the left lane weight and the right lane weight is less than a second preset threshold, determining that the target sampling scheme is a two-line sampling scheme; If the left lane weight and the right lane weight are both greater than the first preset threshold, and the deviation between the left lane weight and the right lane weight is greater than or equal to the second preset threshold, it is determined that the target sampling scheme is a single-line sampling scheme.

7. The method according to claim 6, characterized in that The method further comprises: If the left lane weight and the right lane weight are both less than or equal to the first preset threshold, obtaining a distance change value and an angle change value of the vehicle movement within a preset time period; Determine the state quantity and control matrix; Determine a state transfer matrix according to the state quantity and the distance change value; Determine a control amount according to the control matrix and the angle change value; Determine a prediction equation according to the state quantity, the state transfer matrix, the control matrix and the control quantity; Predicting the target coefficient according to the prediction equation; The final center reference line is determined according to the target coefficient and the multi-order polynomial.

8. A lane center reference line generating device, characterized in that: The device comprises: An initial reference line determination module is used to obtain an initial center reference line for the vehicle based on lane parameter information collected by the vehicle's forward vision; wherein the initial center reference line is expressed based on a multi-order polynomial and a target coefficient; A preprocessing module, used to preprocess the lane parameter information to obtain a lane line history queue, a lane width history queue and a lane center line history queue; A weight calculation module, used to calculate lane line weights based on the lane line history queue, the lane width history queue and the lane center line history queue; wherein the lane line weights include left lane weights and right lane weights; A sampling scheme determination module, used to determine a target sampling scheme from a single-line sampling scheme and a double-line sampling scheme according to the left lane weight and the right lane weight; An observation module, used to determine the lane centerline according to the target sampling scheme, and obtain an observation result based on the lane centerline; wherein the observation result includes a plurality of centerline observation values; The generation module is used to update the observation results using Kalman filtering, and to correct the target coefficients of the multi-order polynomial according to the update results to obtain the final center reference line.

9. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, for implementing the method according to any one of claims 1 to 7 when executing a program stored in a memory.

10. A computer-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 method according to any one of claims 1 to 7 is implemented.

Images