Front vehicle motion state identification method and device, equipment and medium

Abstract

The embodiment of the invention provides a front vehicle motion state recognition method, device and equipment and a medium, and relates to the technical field of intelligent driving, and the method comprises the steps: setting a reference line of a Frenet coordinate system, and calculating the transverse distance from a target vehicle to the reference line based on the reference line; least square estimation is carried out on the transverse distance, and a transverse speed estimation value and a least square estimation variance are calculated; calibrating the state area based on the reference line and the width of the target vehicle; when the absolute value of the transverse speed estimation value is smaller than the least square estimation variance and the transverse distance is in a riding line area, the target is in a riding line state; when the absolute value of the transverse speed estimation value is larger than or equal to the least square estimation variance and the transverse distance is in the cut-in or cut-out area, the probability value of the target vehicle in the cut-in or cut-out area is calculated through the transverse distance, and the target operation state is recognized. According to the scheme, the recognition accuracy of the vehicle motion state is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent driving, and particularly relates to a method, device, equipment and medium for identifying the motion state of a vehicle ahead. Background Art

[0002] The current target detection and judgment mainly include three solutions. The first is the ACC (Adaptive Cruise Control) that uses only the front millimeter-wave radar to achieve detection and control functions; the second is the ACC that uses only the front-view camera for detection and control functions; the third is the ACC that combines the front millimeter-wave radar and the front-view camera fusion solution to achieve detection and control.

[0003] From the analysis of the hardware configuration, the front millimeter-wave radar can well detect the distance and speed characteristics of the front target, and the front-view camera can better detect the shape of the front target, road curvature characteristics, etc. Obviously, the ACC implemented by the front-view camera and the single radar can more accurately identify the road environment and target characteristics. However, from the cost analysis, it is not difficult to see that the ACC solution equipped with a front-view camera requires a higher price configuration. Therefore, the current mainstream solution for implementing the single ACC function in the market is to use a single radar. However, among the problems reflected in the after-sales market, most of the feedback on the ACC usage problems focuses on the algorithm problems of target judgment. Among them, the more typical ones are the cut-in and cut-out of targets, and the disappearance of targets when entering and exiting curves.

[0004] What needs to be considered in the cut-in and cut-out working conditions is how to quickly and accurately identify or release the target during the cut-in and cut-out process. In addition, the riding-line working condition is also difficult to identify, especially when the lateral speed of the target is inaccurate, it will be more difficult. Therefore, the following factors need to be considered:

[0005] The overlap between the vehicle and the vehicle ahead: The camera and the millimeter-wave radar need to accurately identify the target position, especially in the transverse direction. At present, the general camera and millimeter-wave radar have poor recognition of the lateral speed, and it is easy to misjudge the cut-in and cut-out of the target vehicle by relying solely on the perceived information.

[0006] The lateral acceleration of the vehicle ahead relative to the vehicle: The lateral acceleration is used to predict when the target vehicle can be fully regarded as an identification target relative to the vehicle. However, at present, most cameras and millimeter-wave radars cannot provide lateral acceleration information. The ACC solution equipped with a front-view camera for measuring lateral acceleration requires a higher price configuration.

[0007] The lane line recognition information of the camera: For the curve scene, especially the measurement of medium and long-distance curves will be inaccurate, resulting in an easy misjudgment of the cut-out and cut-in states based only on the geometric relationship between the target vehicle and the lane line. Summary of the Invention

[0008] In view of this, an embodiment of the present invention provides a method for identifying the motion state of a preceding vehicle to solve the technical problem in the prior art that the inaccurate measurement of the relative geometric position relationship between the target vehicle and the lane line by the in-vehicle camera and the inaccurate lateral speed measurement error result in inaccurate identification of the cut-in, cut-out, and straddling states. The method includes:

[0009] Set a reference line in the Frenet coordinate system according to the presence of the lane line;

[0010] Calculate the lateral distance from the target vehicle to the reference line based on the reference line;

[0011] Perform least squares estimation based on the lateral distance, and calculate the estimated value of the lateral speed of the target vehicle relative to the reference line and the least squares estimation variance;

[0012] Calibrate the straddling area, cut-in area, and cut-out area based on the reference line and the width of the target vehicle;

[0013] When the absolute value of the estimated value of the lateral speed is less than the least squares estimation variance, and the lateral distance is within the calibrated straddling area, then identify the target vehicle as being in the straddling state;

[0014] When the absolute value of the estimated value of the lateral speed is greater than or equal to the least squares estimation variance, and the lateral distance is within the calibrated cut-in area or cut-out area, calculate the probability value of the target vehicle being in the cut-in area or cut-out area through the lateral distance, and identify the running state of the target vehicle according to the magnitude of the probability value.

[0015] An embodiment of the present invention also provides a device for identifying the motion state of a preceding vehicle to solve the technical problem in the prior art that the inaccurate measurement of the relative geometric position relationship between the target vehicle and the lane line by the in-vehicle camera and the inaccurate lateral speed measurement error result in inaccurate identification of the cut-in, cut-out, and straddling states. The device includes:

[0016] A reference line setting module for setting a reference line in the Frenet coordinate system according to the presence of the lane line;

[0017] A first calculation module for calculating the lateral distance from the target vehicle to the reference line based on the reference line;

[0018] A second calculation module for performing least squares estimation based on the lateral distance, and calculating the estimated value of the lateral speed of the target vehicle relative to the reference line and the least squares estimation variance;

[0019] An area calibration module for calibrating the straddling area, cut-in area, and cut-out area based on the reference line and the width of the target vehicle;

[0020] A first recognition module, configured to recognize that the target vehicle is in a line-riding state when the absolute value of the estimated lateral speed is less than the least square estimation variance and the lateral distance is within the calibrated line-riding area.

[0021] A second recognition module, configured to calculate a probability value that the target vehicle is in the cut-in area or the cut-out area based on the lateral distance when the absolute value of the estimated lateral speed is greater than or equal to the least square estimation variance and the lateral distance is within the calibrated cut-in area or cut-out area, and recognize the running state of the target vehicle according to the magnitude of the probability value.

[0022] An embodiment of the present invention further provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above-mentioned arbitrary method for recognizing the motion state of the vehicle ahead is implemented, so as to solve the technical problem in the prior art that the recognition of the cut-in, cut-out, and line-riding states is inaccurate due to inaccurate measurement of the relative geometric position relationship between the target vehicle and the lane line by the in-vehicle camera and inaccurate lateral speed measurement error.

[0023] An embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program for executing the above-mentioned arbitrary method for recognizing the motion state of the vehicle ahead, so as to solve the technical problem in the prior art that the recognition of the cut-in, cut-out, and line-riding states is inaccurate due to inaccurate measurement of the relative geometric position relationship between the target vehicle and the lane line by the in-vehicle camera and inaccurate lateral speed measurement error.

[0024] Compared with the prior art, the at least one technical solution adopted in the embodiments of the present specification can achieve at least the following beneficial effects: The method of the present application introduces Frenet coordinates and re-estimates the lateral speed based on the least square estimation, rather than simply relying on perception, effectively avoiding misjudgment of cut-in and cut-out recognition; By probabilistically processing cut-in, cut-out, and line-riding to obtain the probability magnitudes of cut-in and cut-out to evaluate the cut-in, cut-out, and line-riding states of the target vehicle, it can effectively avoid misjudgment caused by measurement errors due to over-reliance on position measurement, and can quickly and accurately recognize or release the target during the cut-in and cut-out processes, thereby improving the recognition accuracy of the motion state of the target vehicle. Description of the Drawings

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0026] Figure 1 It is a flowchart of the method for identifying the motion state of the vehicle ahead provided by the embodiment of the present invention;

[0027] Figure 2 It is a schematic diagram of the lateral distance from the target vehicle to the reference line provided by the embodiment of the present invention;

[0028] Figure 3 It is a schematic diagram of the lateral distance array at consecutive time points provided by the embodiment of the present invention;

[0029] Figure 4 It is another flowchart of the method for identifying the motion state of the vehicle ahead provided by the embodiment of the present invention;

[0030] Figure 5 It is a structural block diagram of a computer device provided by the embodiment of the present invention;

[0031] Figure 6 It is a structural block diagram of a device for identifying the motion state of the vehicle ahead provided by the embodiment of the present invention. Detailed implementation manners

[0032] The embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0033] The following uses specific specific examples to illustrate the implementation manners of the present application. Those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The present application can also be implemented or applied through other different specific implementation manners. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without making creative efforts belong to the scope of protection of the present application.

[0034] In the embodiment of the present invention, a method for identifying the motion state of the vehicle ahead is provided. As Figure 1 shown, the method includes:

[0035] Step S101, set a reference line in the Frenet coordinate system according to the presence or absence of lane lines;

[0036] Step S102, calculate the lateral distance from the target vehicle to the reference line based on the reference line;

[0037] Step S103, perform least squares estimation based on the lateral distance, and calculate the estimated value of the lateral speed of the target vehicle relative to the reference line and the least squares estimation variance;

[0038] Step S104: Calibrate the line-riding area, cut-in area, and cut-out area based on the reference line and the width of the target vehicle.

[0039] Step S105: When the absolute value of the estimated lateral speed is less than the least squares estimation variance and the lateral distance is within the calibrated line-riding area, identify the target vehicle as being in the line-riding state.

[0040] Step S106: When the absolute value of the estimated lateral speed is greater than or equal to the least squares estimation variance and the lateral distance is within the calibrated cut-in area or cut-out area, calculate the probability value of the target vehicle being in the cut-in area or cut-out area based on the lateral distance, and identify the running state of the target vehicle according to the magnitude of the probability value.

[0041] In this embodiment, the Frenet coordinate is introduced and the lateral speed is re-estimated based on the least squares estimation, rather than simply relying on perception, effectively avoiding misjudgments in the identification of cut-in and cut-out; by probabilistically processing cut-in, cut-out, and line-riding to obtain the probability magnitudes of cut-in and cut-out to evaluate the cut-in, cut-out, and line-riding states of the target vehicle, it can effectively avoid misjudgments caused by measurement errors due to over-reliance on position measurement, and can quickly and accurately identify or release the target during the cut-in and cut-out processes, thereby improving the recognition accuracy of the running state of the target vehicle.

[0042] In one embodiment, setting the reference line in the Frenet coordinate system according to the presence or absence of lane lines includes:

[0043] When the confidence levels of the lane lines on both the left and right sides are greater than the preset threshold, it is determined that the lane lines exist; otherwise, it is determined that the lane lines do not exist.

[0044] When the lane lines exist, set the reference line in the Frenet coordinate system as the center line of the lane lines on both the left and right sides, and the reference line equation is:

[0045] y = C 0 +C 1 x + C 2 x 2 +C 3 x 3 (1)

[0046] y is the longitudinal coordinate, x is the lateral coordinate, and C 0 、C 1 、C 2 and C 3 are all coefficients. When the lane lines exist, C 0Select half of the sum of the constant term coefficients of the left lane line equation and the right lane line equation, C 1 , C 2 and C 3 are respectively selected as the first-order term coefficient, the second-order term coefficient, and the third-order term coefficient of the lane line equation with a higher confidence level for the left and right lane lines;

[0047] When the lane line does not exist, set the reference line in the Frenet coordinate system as the center line of the vehicle's trajectory, and the reference line equation is:

[0048] y = C 0 + C 1 x + C 2 x 2 (2)

[0049] When the lane line does not exist, C 0 is selected as the position of the center point of the vehicle's front bumper, C 0 = 0, C 1 is selected as the slope of the vehicle's trajectory center line at the center point of the front bumper, C 2 is selected as the curvature of the vehicle's trajectory center line at the center point of the front bumper.

[0050] In specific implementation, the confidence level of the lane line is between 0 and 1. Set a preset threshold. When the value (confidence level) obtained from the front-end perception for evaluating the credibility of the lane line exceeds this preset threshold, that is, the lane line exists; otherwise, the lane line does not exist. For the lane line equations of the left and right lane lines, both are described by a cubic polynomial. If the lane line exists, that is, the confidence levels of the left and right lane lines are both greater than the preset threshold. At this time, select the center line of the left and right lane lines as the reference line. For the reference line equation y = C 0 + C 1 x + C 2 x 2 + C 3 x 3 , C 0 is selected as half of the sum of the constant term coefficients of the left lane line equation and the right lane line equation. When the confidence level of the left lane line is higher than that of the right lane line, C 1 , C 2 and C 3 are respectively selected as the first-order term coefficient, the second-order term coefficient, and the third-order term coefficient of the left lane line equation. When the confidence level of the right lane line is higher than that of the left lane line, C 1 , C 2 and C 3They are respectively selected as the first-order coefficient, second-order coefficient, and third-order coefficient of the right lane line equation. If the lane line does not exist, that is, the confidence levels of the left and right lane lines are both less than the preset threshold, the ego-vehicle trajectory is selected as the reference line. The center line equation of the ego-vehicle trajectory involved in this embodiment is a parabolic curve. Therefore, the reference line equation is a quadratic polynomial y = C 0 + C 1 x + C 2 x 2 .

[0051] In one embodiment, calculating the lateral distance from the target vehicle to the reference line based on the reference line includes:

[0052] Project the center point x of the vehicle width of the target vehicle onto the reference line to obtain the projection point r;

[0053] Calculate the distance from the center point x of the vehicle width to the projection point r. This distance is the lateral distance l from the target vehicle to the reference line. The calculation formula for the lateral distance l is:

[0054]

[0055] where the coordinate point (x x , y x ) is the coordinate point of the center point x of the vehicle width of the target vehicle in the rectangular coordinate system, the coordinate point (x r , y r ) is the coordinate point of the projection point r in the rectangular coordinate system, and θ r is the angle between the tangent vector of the projection point r on the reference line and the x-axis.

[0056] Refer to Figure 2 for a detailed description of the calculation of the lateral distance. Figure 2 The symbols in are all expressed in the form of vectors. The center point of the vehicle width of the target vehicle is The projection point of this point on the reference line is The distance between the center point of the vehicle width and the projection point is the lateral distance l(s) in the Frenet coordinate system. The coordinate point of the center point of the vehicle width of the target vehicle x , y x ) in the rectangular coordinate system, and the coordinate point of the projection point in the rectangular coordinate system is (x r , y r ). is the tangent vector of the projection point on the reference line and the angle with the x-axis, is the tangent vector of the center point of the vehicle width on the reference line Angle with the x-axis, is the tangent vector of the normal vector, is the tangent vector of the normal vector.

[0057] In one embodiment, the least squares estimation based on the lateral distance to calculate the lateral speed estimation value of the target vehicle relative to the reference line and the least squares estimation variance includes:

[0058] Sample the lateral distance at preset time intervals within a preset time window to obtain a set of measured values l of the lateral distance at consecutive time points i , where i is the i-th time point;

[0059] Obtain the true value L of the lateral distance corresponding to the set of consecutive time points i ;

[0060] Based on the measured values l of the lateral distance at the set of consecutive time points i , construct a coefficient matrix and an observation matrix for least squares estimation;

[0061] Based on the true value L of the lateral distance i , the coefficient matrix and the observation matrix, perform least squares estimation to obtain the lateral speed estimation value and the least squares estimation variance.

[0062] In specific implementation, in actual perception measurement, the accuracy of the position measurement of a moving target is generally higher than that of the target speed measurement. There are relatively large errors in calculating the lateral relative speed of the target in the Frenet coordinate system based on the target perception speed. Therefore, in this embodiment, based on the calculation of l, a preset time window of 1 s and a preset time interval of 100 ms are set for illustration. Specifically, by sampling the l value of the target every 100 ms within a 1-s time window of l, a more accurate lateral distance change rate, that is, the lateral speed relative to the reference line, is obtained by least squares estimation. For this, the coefficient matrix H and the observation matrix Y for calculating the least squares estimation are constructed as shown in the following formula:

[0063]

[0064] where l 1 ~l 10 are the lateral distances li of the target vehicle relative to the reference line in the Frenet coordinate system sampled every 100 ms within a 1-s sliding time window, and the subscript i represents the i-th sampling time point within the time window. l 1 ~l 10 are the measured values l of the lateral speed at a set of consecutive time points i. Specifically, l is the lateral distance measurement value in the Cartesian coordinate system converted to the lateral distance measurement value in the Frenet coordinate system, and L 1 ~L 10 is the true value of the lateral distance in the corresponding Frenet coordinate system, y is the observation matrix of the true value of the lateral distance, and L i+1 -L i is the difference between the true values of the lateral distances of the two adjacent sampling points, and l i+1 -l i is the difference between the measurement values of the lateral distances of the two adjacent sampling points. To ensure the timeliness of the data, the time window slides every fixed period of 100 ms to ensure the timely update of the data and avoid calculation delays. The specific sliding process is as Figure 3 shown.

[0065] Specifically, when implemented, the calculation processes for the lateral velocity estimation value and the least squares estimation variance are as follows:

[0066] When performing the least squares estimation, since the estimation value is easily affected by the lateral ranging error, the lateral velocity estimation value can be regarded as a random variable, and it is assumed that the lateral velocity estimation value satisfies a normal distribution. The true value l i is near the corresponding measurement value L i and satisfies a Gaussian distribution. Then, the mean and variance of the lateral velocity estimation value are calculated as shown in the following formula:

[0067]

[0068] where E is the mean of the corresponding matrix, D is the variance of the corresponding matrix, the error between the lateral distance measurement value l i of each sampling point in the Frenet coordinate system and the corresponding true value L i is ω i , and this error ω i is assumed to be a set of uncorrelated white noises with zero mean, so γ is a white noise random vector. Therefore, under the premise of knowing the vector Y each time, the vector y can also be regarded as a random variable. When Hspeed is based on the true value of the lateral distance, the random vector y is substituted into the least squares estimation to obtain the lateral velocity estimation quantity. The lateral velocity estimation quantity Hspeed can also be regarded as a random variable, and the mean E(Hspeed) and variance D(Hspeed) of the lateral velocity estimation quantity Hspeed are obtained according to formula (6). Hspeed is the lateral velocity estimation quantity calculated based on the measurement value of the lateral distance; σ 2 is the variance of each component of the γ vector, and it is assumed that the components are independent of each other; due to the special structure of the matrix Y, J 9 is not an identity matrix and is a normal matrix or a real symmetric matrix.

[0069] Further, through the above formula, we obtain where μ s and are the mean value E(Hspeed) and variance D(Hspeed) respectively, which is the least squares estimation variance. Then is the conditional probability distribution of the true lateral speed under the premise of obtaining the lateral speed estimation value hspeed. This is a posterior probability, and it is assumed that this probability distribution conforms to the normal distribution. hspeed is as shown in formula (6). hspeed is the estimation value calculated by substituting the measured value of the lateral distance into the least squares formula. Formula (6) shows that E(Hspeed) is equal to hspeed.

[0070] Let d(Hspeed) be the root mean square of D(Hspeed). When E(Hspeed) takes values of 2d(Hspeed), d(Hspeed), 0, -d(Hspeed), and -2d(Hspeed), the probability values of Hspeed taking values in the interval [-d(Hspeed), d(Hspeed)] are: 0.1573, 0.4772, 0.6827, 0.4772, 0.1573 respectively. Since the perceived lateral distance measurement is relatively accurate, the calibrated values of the root mean square σ of each component of the γ vector in practice are generally small. Therefore, when the least squares estimation value E(Hspeed) is within the range of [-d(Hspeed), d(Hspeed)], the probability that Hspeed may take values in [-d(hspeed), d(hspeed)] is greater than or equal to 0.4772, that is, it is considered that the magnitude of the lateral speed of the target in the Frenet coordinate system is zero at this time. Thus, through the above steps, the lateral speed estimation value in the Frenet coordinate system has been probabilized, and the probability distribution satisfies the normal distribution.

[0071] In one embodiment, the method further includes:

[0072] Probabilize the lateral speed estimation value, and the probability distribution of the lateral speed estimation value satisfies the normal distribution.

[0073] In this embodiment, the probability theory method is used to probabilize the lateral speed of the target.

[0074] In one embodiment, the calibration of the straddling area, the cutting-in area, and the cutting-out area based on the reference line and the width of the target vehicle includes:

[0075] Set area segmentation points, and the area segmentation points are C 0 -(5 / 8)wd, C 0-(3 / 8)wd, C 0 , C 0 +(3 / 8)wd and C 0 +(5 / 8)wd;

[0076] Based on the region segmentation points, calibrate the region where |l - C 0 | ≤ (3 / 8)wd as the straddle region, and calibrate the region where l - C 0 | ≤ (5 / 8)wd as the cut-in region or the cut-out region, where wd is the width of the target vehicle.

[0077] In specific implementation, the judgment of the cut-in, cut-out, and straddle of the target vehicle is based on the judgment of the geometric relationship of the target relative to the reference line. For example, when the reference line is the center line of the left and right lane lines, for the recognition of the cut-in, cut-out, and straddle states of the left or right target, in this embodiment, the region is divided with the left or right lane line as the center, and the positions of the left or right lane lines are determined by the constant term coefficients of the left lane line equation or the right lane line equation respectively. The region is divided into four parts C 0 -(5 / 8)wd ~ C 0 -(3 / 8)wd, C 0 -(3 / 8)wd ~ C 0 , C 0 ~ C 0 +(3 / 8)wd, C 0 +(3 / 8)wd ~ C 0 +(5 / 8)wd, define the region where |l - C 0 | ≤ (3 / 8)wd as the straddle area, and |l - C 0 | ≤ (5 / 8)wd as the cut-in or cut-out region, where wd is the width of the target vehicle. For judging cut-in and cut-out, not only does the target need to be currently in the cut-in or cut-out region, but the target also needs to leave the cut-in or cut-out region within the calibrated time. Table 1 takes the right lane line as an example, and according to the cut-in or cut-out region l - C 0 | ≤ (5 / 8)wd, five segmentation points are divided, and different cut-in or cut-out leaving time constraints are respectively corresponding when the target lateral distance is near the segmentation points. The left lane line is the same, just in the opposite direction.

[0078] Table 1 Cut-in or cut-out leaving time constraints

[0079]

[0080] It should be noted that the time constraints in Table 1 are only examples, and in specific implementation, the set time can be adjusted according to the situation. When the target lateral distance is within the intervals divided by the segmentation points in Table 1, the corresponding cut-in and cut-out times are obtained by linear interpolation of the cut-in and cut-out times at the segmentation points.

[0081] During specific implementation, when the absolute value of the estimated lateral speed is less than the least squares estimation variance and the lateral distance is within the calibrated lane straddling area, the target vehicle is identified as being in the lane straddling state, which specifically includes the following steps:

[0082] First, perform least squares estimation on the constructed matrix H and Y to obtain an estimated lateral speed hspeed. If the absolute value of hspeed is less than the least squares estimation variance, according to the conclusion of the above steps, at this time, the lateral speed of the target vehicle in the Frenet coordinate system is 0 m / s, that is, the target vehicle is laterally stationary relative to the reference line. If the target lateral distance is within the range of |l - C 0 | ≤ (3 / 8)wd, it is in the lane straddling area state. If the target lateral distance is not within the range of |l - C 0 | ≤ (3 / 8)wd, then the target vehicle is neither in the lane straddling nor in the cutting-in and cutting-out states. Conversely, if the absolute value of the estimated lateral speed hspeed is greater than the least squares estimation variance and the target lateral distance is within the range of |l - C 0 | ≤ (5 / 8)wd, it jumps to the judgment of cutting-in and cutting-out. If the lateral speed is not 0 m / s and the target lateral distance is also not within the range of |l - C 0 | ≤ (5 / 8)wd, then the motion state of the target vehicle is neither lane straddling nor cutting-in and cutting-out.

[0083] In one embodiment, referring to Figure 4 , calculating the probability value of the target vehicle being in the cutting-in area or the cutting-out area based on the lateral distance, and identifying the running state of the target vehicle according to the magnitude of the probability value, includes:

[0084] Let the lateral distance moved by the target vehicle within the cutting-in time Cut lnT be Cut lnD, and the lateral distance moved by the target vehicle within the cutting-out time CutoutT be CutoutD. Both Cut lnD and CutoutD follow a normal distribution, and the cumulative distribution function of Cut lnD and the cumulative distribution function of CutoutD are respectively obtained;

[0085] Based on the cumulative distribution function of Cut lnD and the cumulative distribution function of CutoutD, calculate the probability value of the target vehicle being in the cutting-in area or the cutting-out area;

[0086] When |l - C Lf | < |l - C Ri |, the target vehicle is close to the left lane line. When |l - C Lf | < |l - C Ri |, the target vehicle is close to the right lane line, CLf is the constant term coefficient of the left lane line equation, C Ri is the constant term coefficient of the right lane line equation;

[0087] When the target vehicle approaches the left lane line and the estimated lateral speed is greater than positive one times the least squares estimation variance, if the probability value P(l + CutInD > C Lf +(5 / 8)wd) > 0.6, it is recognized that the target vehicle is in the cut-in state. If P(l + CutInD > C Lf +(5 / 8)wd) ≤ 0.6 and |l - C Lf | ≤ (3 / 8)wd, it is recognized that the target vehicle is in the on-line state. Otherwise, the target vehicle is not in the cut-in state, cut-out state, or on-line state;

[0088] When the target vehicle approaches the left lane line and the estimated lateral speed is less than or equal to negative one times the least squares estimation variance, if the probability value P(l + CutoutD < C Lf -(5 / 8)wd) > 0.6, it is recognized that the target vehicle is in the cut-out state. If P(l + CutoutD < C Lf -(5 / 8)wd) ≤ 0.6 and |l - C Lf | ≤ (3 / 8)wd, it is recognized that the target vehicle is in the on-line state. Otherwise, the target vehicle is not in the cut-in state, cut-out state, or on-line state;

[0089] When the target vehicle approaches the right lane line and the estimated lateral speed is greater than or equal to positive one times the least squares estimation variance, if the probability value P(l + CutoutD > C Ri +(5 / 8)wd) > 0.6, it is recognized that the target vehicle is in the cut-out state. If P(l + CutoutD > C Ri +(5 / 8)wd) ≤ 0.6 and |l - C Ri | ≤ (3 / 8)wd, it is recognized that the target vehicle is in the on-line state. Otherwise, the target vehicle is not in the cut-in state, cut-out state, or on-line state;

[0090] When the target vehicle approaches the right lane line and the estimated lateral speed is less than negative one times the least squares estimation variance, if the probability value P(l + CutInD < C Ri -(5 / 8)wd) > 0.6, it is recognized that the target vehicle is in the cut-in state. If P(l + CutInD < C Ri -(5 / 8)wd) ≤ 0.6 and |l - CRi If |≤(3 / 8)wd, it is recognized that the target vehicle is in the straddling state; otherwise, the target vehicle is not in the cut-in state, cut-out state, or straddling state.

[0091] During specific implementation, referring to Figure 4 , set C Lf as the constant term coefficient of the left lane line equation, and C Ri as the constant term coefficient of the right lane line equation. When |l - C Lf | < |l - C Ri |, it indicates that the target vehicle is close to the left lane line. Then, the judgment of the target vehicle's cut-in, cut-out, and straddling is bound to the left lane line, and the following judgment continues:

[0092] 1) If the estimated lateral speed hspeed obtained by performing least squares estimation calculation based on the constructed matrix H and Y is positive (right is positive and left is negative, that is, the estimated lateral speed is greater than one positive multiple of the least squares estimation variance), then the target vehicle may be in the cut-in state. Since where μ s is hspeed, is D(Hspeed), then multiplying Hspeed by the corresponding cut-in time in Table 1 is the lateral distance moved by the target vehicle within the cut-in time. Since Hspeed is a random variable, the lateral distance moved by the target vehicle within the cut-in time is also a random variable. Let the lateral distance moved by the target vehicle within the cut-in time be Cut l nD, and the cut-in time be set as Cut l nT. Cut l nD follows a normal distribution, that is where μ d is the product of hspeed and Cut l T, is the product of D(Hspeed) and the square of Cut l nT;

[0093] 2) Cut l nD follows a normal distribution, that is Then, let the cumulative distribution function of Cut l nD be Use the cumulative distribution function to calculate the cut-in probability. Let hd i st be the current lateral distance of the target vehicle, that is, calculate P(hd i st + μ d > C Lf +(5 / 8)wd). If P(hd i st + μ d > C Lf +(5 / 8)wd) > 0.6, it is recognized as cut-in. Specifically, hd i st is the current lateral distance, μ d is the cut-in time Cut l nT multiplied by hspeed (estimated lateral speed), and hd i st + μ dis the lateral distance of the target vehicle after cutting in, then P(hd ist+μ d >C Lf +(5 / 8)wd) is the probability that the target crosses this split point after cutting in. If the probability that the target crosses this split point after cutting in is greater than the set value 0.6, it is considered that the target has a cutting action. d >C Lf +(5 / 8)wd)≤0.6 and |hd i st-C Lf |≤(3 / 8)wd, it is identified as riding the line, otherwise it is neither cutting in nor riding the line, nor cutting out;

[0094] 3) If the lateral speed estimate hspeed calculated by the least squares estimation based on the constructed matrices H and Y is negative (positive on the right and negative on the left, that is, the lateral speed estimate is less than or equal to negative one times the least squares estimation variance), the target may be cut out, because where μ s is hspeed, D(Hspeed), then Hspeed multiplied by the corresponding cut-in time in Table 1 is the lateral distance moved by the target vehicle during the cut-out time. Since Hspeed is a random variable, the lateral distance moved by the target vehicle during the cut-out time is also a random variable. Let the lateral distance moved by the target vehicle during the cut-out time be CutoutD, and the cut-out time be CutoutT. CutoutD obeys the normal distribution, that is, where μ d is the product of hspeed and CutoutT, It is the product of D(Hspeed) and CutoutT square;

[0095] 4) CutoutD follows a normal distribution, that is Then let the cumulative distribution function of CutoutD be The cut-out probability is calculated using the cumulative distribution function, that is, P(hd ist+μ d <C Lf -(5 / 8)wd), if P(hdist+μ d <C Lf -(5 / 8)wd)>0.6, it is identified as cut out. If P(hdist+μ d <C Lf -(5 / 8)wd)≤0.6 and |hdist-C Lf |≤(3 / 8)wd, it is identified as riding the line, otherwise it is neither cutting out nor riding the line, nor cutting in.

[0096] In specific implementation, when |lC Lf |>|lCRi When it indicates that the target vehicle is close to the right lane line, the judgment of the target vehicle's cut-in, cut-out, and lane-riding is bound to the right lane line. Then, a similar judgment process as above is continued, and no repeated elaboration will be made here.

[0097] In the embodiments of the present application, a relatively accurate lateral speed is obtained through the least square estimation of the lateral distance in the Frenet coordinate system, which solves the problem that the camera generally cannot accurately measure the target lateral speed in the curved road scenario. The probability theory method is used to probabilize the target lateral speed, and the probability values of the preceding vehicle's cut-in, cut-out, and lane-riding are calculated in real time through the target's current lateral distance. Whether the front vehicle target is in the state of cut-in, cut-out, or lane-riding is arbitrated according to the probability size. This method solves the problem of inaccurate identification of cut-in, cut-out, and lane-riding caused by inaccurate measurement of the relative geometric position relationship between the target vehicle and the lane line by the in-vehicle camera and inaccurate lateral speed measurement error.

[0098] In this embodiment, a computer device is provided, as Figure 5 shown, including a memory 501, a processor 502, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above-mentioned arbitrary method for identifying the motion state of the preceding vehicle is implemented.

[0099] Specifically, the computer device can be a computer terminal, a server, or a similar computing device.

[0100] In this embodiment, a computer-readable storage medium is provided, and the computer-readable storage medium stores a computer program for executing the above-mentioned arbitrary method for identifying the motion state of the preceding vehicle.

[0101] Specifically, the computer-readable storage medium includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer-readable storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory, or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD), or other optical storage, magnetic cassette tapes, magnetic disk storage, or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible by a computing device. As defined herein, computer-readable storage media do not include transitory computer-readable media, such as modulated data signals and carrier waves.

[0102] Based on the same inventive concept, an apparatus for identifying the motion state of a preceding vehicle is also provided in an embodiment of the present invention, as described in the following embodiments. Since the principle of the apparatus for identifying the motion state of a preceding vehicle to solve the problem is similar to that of the method for identifying the motion state of a preceding vehicle, the implementation of the apparatus for identifying the motion state of a preceding vehicle can refer to the implementation of the method for identifying the motion state of a preceding vehicle, and the repeated parts will not be described again. As used hereinafter, the term "unit" or "module" may be a combination of software and / or hardware that can achieve a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0103] Figure 6 is a structural block diagram of the apparatus for identifying the motion state of a preceding vehicle according to an embodiment of the present invention, as Figure 6 shown, including: a reference line setting module 601, a first calculation module 602, a second calculation module 603, a region calibration module 604, a first identification module 605, and a second identification module 606. The following describes this structure.

[0104] The reference line setting module 601 is configured to set a reference line in the Frenet coordinate system according to the presence of lane lines;

[0105] The first calculation module 602 is configured to calculate the lateral distance from the target vehicle to the reference line based on the reference line;

[0106] The second calculation module 603 is configured to perform a least squares estimation based on the lateral distance, and calculate the estimated value of the lateral speed of the target vehicle relative to the reference line and the least squares estimation variance;

[0107] The region calibration module 604 is configured to calibrate a straddling region, a cutting-in region, and a cutting-out region based on the reference line and the width of the target vehicle;

[0108] The first identification module 605 is configured to identify the target vehicle as being in a straddling state when the estimated value of the lateral speed is less than the least squares estimation variance and the lateral distance is within the calibrated straddling region;

[0109] The second identification module 606 is configured to calculate the probability value of the target vehicle being in the cutting-in region or the cutting-out region based on the lateral distance when the estimated value of the lateral speed is greater than or equal to the least squares estimation variance and the lateral distance is within the calibrated cutting-in region or cutting-out region, and identify the running state of the target vehicle according to the magnitude of the probability value.

[0110] In one embodiment, the reference line setting module 601 is further configured to:

[0111] When the confidence levels of the left and right lane lines are both greater than a preset threshold, it is determined that the lane line exists; otherwise, it is determined that the lane line does not exist.

[0112] When the lane line exists, the reference line in the Frenet coordinate system is set as the center line of the left and right lane lines, and the reference line equation is y = C 0 +C 1 x+C 2 x 2 +C 3 x 3 , where y is the longitudinal coordinate, x is the lateral coordinate, and C 0 , C 1 , C 2 and C 3 are all coefficients. When the lane line exists, C 0 is selected as half of the sum of the constant term coefficients of the left lane line equation and the right lane line equation, and C 1 , C 2 and C 3 are respectively selected as the first-order term coefficient, the second-order term coefficient, and the third-order term coefficient of the lane line equation with a higher confidence level among the left and right lane lines.

[0113] When the lane line does not exist, the reference line in the Frenet coordinate system is set as the center line of the vehicle's own trajectory, and the reference line equation is y = C 0 +C 1 x+C 2 x 2 . When the lane line does not exist, C 0 is selected as the position of the center point of the front bumper of the vehicle itself, C 0 = 0, and C 1 is selected as the slope of the center line of the vehicle's own trajectory at the center point of the front bumper, and C 2 is selected as the curvature of the center line of the vehicle's own trajectory at the center point of the front bumper.

[0114] In one embodiment, the first calculation module 602 is further configured to:

[0115] Project the vehicle width center point x of the target vehicle onto the reference line to obtain the projection point r;

[0116] Calculate the distance from the vehicle width center point x to the projection point r, and this distance is the lateral distance l of the target vehicle from the reference line. The calculation formula for the lateral distance l is:

[0117]

[0118] where the coordinate point (x x ,y x) is the coordinate point of the center point x of the vehicle width of the target vehicle in the rectangular coordinate system, and the coordinate point (x r , y r ) is the coordinate point of the projection point r in the rectangular coordinate system, and θ r is the included angle between the tangent vector of the projection point r on the reference line and the x-axis.

[0119] In one embodiment, the second calculation module 603 is further configured to:

[0120] Sample the lateral distance at preset time intervals within a preset time window to obtain a set of measured values l of the lateral distance at consecutive time points i , where i is the i-th time point;

[0121] Obtain the true value L of the lateral distance corresponding to the set of consecutive time points i ;

[0122] Based on the measured values l of the lateral distance at the set of consecutive time points i , construct a coefficient matrix and an observation matrix for least squares estimation;

[0123] Based on the true value L of the lateral distance i , the coefficient matrix and the observation matrix, perform least squares estimation to obtain the estimated value of the lateral speed and the least squares estimation variance.

[0124] In one embodiment, the device further includes:

[0125] A probability module for probabilizing the estimated value of the lateral speed, and the probability distribution of the estimated value of the lateral speed satisfies a normal distribution.

[0126] In one embodiment, the area calibration module 604 is further configured to:

[0127] Set area segmentation points, and the area segmentation points are respectively C 0 -(5 / 8)wd, C 0 -(3 / 8)wd, C 0 , C 0 +(3 / 8)wd and C 0 +(5 / 8)wd;

[0128] Based on the area segmentation points, calibrate the area where |l - C 0 | ≤ (3 / 8)wd as the straddling line area, and calibrate the area where l - C 0 | ≤ (5 / 8)wd as the cutting-in area or the cutting-out area, where wd is the width of the target vehicle.

[0129] In one embodiment, the second recognition module 606 is further configured to:

[0130] Let the lateral distance traveled by the target vehicle within the cut-in time Cut lnT be Cut lnD, and the lateral distance traveled by the target vehicle within the cut-out time CutoutT be CutoutD. Both Cut lnD and CutoutD follow a normal distribution. The cumulative distribution function of Cut lnD and the cumulative distribution function of CutoutD are obtained respectively;

[0131] Based on the cumulative distribution function of Cut lnD and the cumulative distribution function of CutoutD, calculate the probability value that the target vehicle is in the cut-in area or the cut-out area;

[0132] When |l - C Lf | < |l - C Ri |, the target vehicle is close to the left lane line. When |l - C Lf | < |l - C Ri |, the target vehicle is close to the right lane line. C Lf is the constant term coefficient of the left lane line equation, and C Ri is the constant term coefficient of the right lane line equation;

[0133] When the target vehicle is close to the left lane line and the estimated lateral speed is greater than positive one times the least squares estimation variance, if the probability value P(l + CutInD > C Lf +(5 / 8)wd) > 0.6 that the target vehicle is in the cut-in area or the cut-out area, it is identified that the target vehicle is in the cut-in state. If P(l + CutInD > C Lf +(5 / 8)wd) ≤ 0.6 and |l - C Lf | ≤ (3 / 8)wd, it is identified that the target vehicle is in the straddling state. Otherwise, the target vehicle is not in the cut-in state, cut-out state, or straddling state;

[0134] When the target vehicle is close to the left lane line and the estimated lateral speed is less than or equal to negative one times the least squares estimation variance, if the probability value P(l + CutoutD < C Lf -(5 / 8)wd) > 0.6 that the target vehicle is in the cut-in area or the cut-out area, it is identified that the target vehicle is in the cut-out state. If P(l + CutoutD < C Lf -(5 / 8)wd) ≤ 0.6 and |l - C Lf | ≤ (3 / 8)wd, it is identified that the target vehicle is in the straddling state. Otherwise, the target vehicle is not in the cut-in state, cut-out state, or straddling state;

[0135] When the target vehicle approaches the right lane line and the estimated lateral speed is greater than or equal to positive one times the least squares estimation variance, if the probability value P(l + CutoutD > C Ri +(5 / 8)wd) > 0.6, the target vehicle is identified as being in the cut-out state. If P(l + CutoutD > C Ri +(5 / 8)wd) ≤ 0.6 and |l - C Ri | ≤ (3 / 8)wd, the target vehicle is identified as being in the on-line state. Otherwise, the target vehicle is not in the cut-in state, cut-out state, or on-line state;

[0136] When the target vehicle approaches the right lane line and the estimated lateral speed is less than negative one times the least squares estimation variance, if the probability value P(l + CutInD < C Ri -(5 / 8)wd) > 0.6, the target vehicle is identified as being in the cut-in state. If P(l + CutInD < C Ri -(5 / 8)wd) ≤ 0.6 and |l - C Ri | ≤ (3 / 8)wd, the target vehicle is identified as being in the on-line state. Otherwise, the target vehicle is not in the cut-in state, cut-out state, or on-line state.

[0137] The embodiments of the present invention achieve the following technical effects: The method of the present application introduces Frenet coordinates and re-estimates the lateral speed based on the least squares estimation, rather than simply relying on perception, effectively avoiding misjudgment in the recognition of cut-in and cut-out; By probabilistically processing cut-in, cut-out, and on-line to obtain the probability magnitudes of cut-in and cut-out to evaluate the cut-in, cut-out, and on-line states of the target vehicle, it can effectively avoid misjudgment caused by measurement errors due to over-reliance on position measurement, and can accurately identify or release the target quickly and precisely during the cut-in and cut-out processes, thereby improving the recognition accuracy of the motion state of the target vehicle.

[0138] Obviously, those skilled in the art should understand that the above-mentioned modules or steps of the embodiments of the present invention can be implemented by a general-purpose computing device. They can be concentrated on a single computing device or distributed on a network composed of multiple computing devices. Optionally, they can be implemented by program codes executable by the computing device, so that they can be stored in a storage device and executed by the computing device. And in some cases, the steps shown or described can be executed in a different order than here, or they can be separately made into individual integrated circuit modules, or multiple modules or steps among them can be made into a single integrated circuit module to implement. Thus, the embodiments of the present invention are not limited to any specific combination of hardware and software.

[0139] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the embodiments of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for identifying the motion state of a preceding vehicle, characterized in that: include: According to the existence of lane lines, set the reference line in the Frenet coordinate system; Calculating a lateral distance from the target vehicle to the reference line based on the reference line; Performing least squares estimation based on the lateral distance, calculating an estimated value of the lateral speed of the target vehicle relative to the reference line and a least squares estimation variance; Based on the reference line and the width of the target vehicle, calibrate the line-riding area, the cut-in area and the cut-out area; When the absolute value of the lateral velocity estimation value is less than the least square estimation variance, and the lateral distance is in the calibrated line-riding area, the target vehicle is identified as being in a line-riding state; When the absolute value of the lateral velocity estimate is greater than or equal to the least squares estimation variance, and the lateral distance is in the calibrated cut-in area or cut-out area, the probability value of the target vehicle being in the cut-in area or the cut-out area is calculated by the lateral distance, and the running state of the target vehicle is identified according to the size of the probability value.

2. The method for identifying the motion state of a preceding vehicle according to claim 1, characterized in that: The step of setting a reference line in the Frenet coordinate system according to the presence of the lane line includes: When the confidence of the lane lines on the left and right sides are both greater than the preset threshold, it is determined that the lane line exists, otherwise it is determined that the lane line does not exist; When the lane line exists, the reference line in the Frenet coordinate system is set to the center line of the lane lines on the left and right sides, and the reference line equation is y=C0+C1x+C2x 2 +C3x 3 , y is the longitudinal coordinate, x is the transverse coordinate, C0, C1, C2 and C3 are all coefficients. When the lane line exists, C0 is selected as half of the sum of the constant coefficients of the left lane line equation and the right lane line equation, and C1, C2 and C3 are respectively selected as the linear coefficient, quadratic coefficient and cubic coefficient of the lane line equation with higher confidence on the left and right sides; When the lane line does not exist, the reference line in the Frenet coordinate system is set to the center line of the vehicle trajectory, and the reference line equation is y=C0+C1x+C2x 2 When the lane line does not exist, C0 is selected as the position of the front protection center point of the vehicle, C0=0, C1 is selected as the slope of the center line of the vehicle trajectory at the front protection center point, and C2 is selected as the curvature of the center line of the vehicle trajectory at the front protection center point.

3. The method for identifying the motion state of a preceding vehicle according to claim 2, characterized in that: The calculating the lateral distance from the target vehicle to the reference line based on the reference line includes: Projecting the center point x of the vehicle width of the target vehicle onto the reference line to obtain a projection point r; Calculate the distance from the vehicle width center point x to the projection point r, which is the lateral distance l from the target vehicle to the reference line. The calculation formula of the lateral distance l is: Among them, the coordinate point (x x ,y x ) is the coordinate point of the vehicle width center point x of the target vehicle in the rectangular coordinate system, and the coordinate point (x r ,y r ) is the coordinate point of the projection point r in the rectangular coordinate system, θ r is the angle between the tangent vector of the projection point r on the reference line and the x-axis.

4. The method for identifying the motion state of a preceding vehicle according to claim 1, characterized in that: The performing least squares estimation based on the lateral distance to calculate the lateral speed estimation value of the target vehicle relative to the reference line and the least squares estimation variance includes: The lateral distance is sampled at preset time intervals within a preset time window to obtain a set of lateral distance measurement values ​​at consecutive time points. i , i is the i-th time point; Get the true value L of the lateral distance corresponding to the set of continuous time points i ; Based on the measured value l of the lateral distance of the set of consecutive time points i , construct the coefficient matrix and observation matrix for least squares estimation; Based on the true value of the lateral distance L i , the coefficient matrix and the observation matrix perform least squares estimation to obtain the lateral velocity estimate and the least squares estimation variance.

5. The method for identifying the motion state of a preceding vehicle according to claim 1, characterized in that: The method further comprises: The lateral velocity estimation value is probabilized, and the probability distribution of the lateral velocity estimation value satisfies a normal distribution.

6. The method for identifying the motion state of a preceding vehicle according to claim 3, characterized in that: The calibrating of the line-riding area, the cut-in area, and the cut-out area based on the reference line and the width of the target vehicle includes: Setting area division points, the area division points are C0-(5 / 8)wd, C0-(3 / 8)wd, C0, C0+(3 / 8)wd and C0+(5 / 8)wd; Based on the area segmentation point, the area marked with |l-C0|≤(3 / 8)wd is the line riding area, and the area marked with |l-C0|≤(5 / 8)wd is the cut-in area or the cut-out area, wherein wd is the width of the target vehicle.

7. The method for identifying the motion state of a preceding vehicle according to claim 6, characterized in that: The step of calculating the probability value of the target vehicle being in the cut-in area or the cut-out area by using the lateral distance, and identifying the running state of the target vehicle according to the magnitude of the probability value, includes: Assume that the lateral distance moved by the target vehicle within the cut-in time CutlnT is CutlnD, and the lateral distance moved by the target vehicle within the cut-out time CutoutT is CutoutD, and both CutlnD and CutoutD obey normal distribution, and obtain the cumulative distribution function of CutlnD and the cumulative distribution function of CutoutD respectively; Based on the cumulative distribution function of CutInD and the cumulative distribution function of CutoutD, calculating the probability value of the target vehicle being in the cut-in area or the cut-out area; When | lC Lf |<|lC Ri |, the target vehicle approaches the left lane line, when |lC Lf |<|lC Ri |, the target vehicle approaches the right lane line, C Lf is the constant coefficient of the left lane line equation, C Ri is the constant term coefficient of the right lane line equation; When the target vehicle is close to the left lane line and the lateral velocity estimate is greater than positive one times the least squares estimate variance, if the probability value P(l+CutInD>C Lf +(5 / 8)wd)>0.6, the target vehicle is identified as being in the cut-in state. If P(l+CutInD>C Lf +(5 / 8)wd)≤0.6 and |lC Lf |≤(3 / 8)wd, the target vehicle is identified as being in a line-riding state; otherwise, the target vehicle is not in a cut-in state, a cut-out state, or a line-riding state; When the target vehicle is close to the left lane line and the lateral speed estimate is less than or equal to negative one times the least squares estimate variance, if the target vehicle is in the cut-in area or the cut-out area, the probability value P(l+CutoutD <C Lf -(5 / 8)wd)>0.6, then the target vehicle is identified as being in the cut-out state. If P(l+CutoutD <C Lf -(5 / 8)wd)≤0.6 and |lC Lf |≤(3 / 8)wd, the target vehicle is identified as being in a line-riding state; otherwise, the target vehicle is not in a cut-in state, a cut-out state, or a line-riding state; When the target vehicle is close to the right lane line and the lateral velocity estimate is greater than or equal to positive one times the least squares estimate variance, if the probability value P(l+CutoutD>CutoutD) of the target vehicle is in the cut-in area or the cut-out area Ri +(5 / 8)wd)>0.6, the target vehicle is identified as being in the cut-out state. If P(l+CutoutD>C Ri +(5 / 8)wd)≤0.6 and |lC Ri |≤(3 / 8)wd, the target vehicle is identified as being in a line-riding state; otherwise, the target vehicle is not in a cut-in state, a cut-out state, or a line-riding state; When the target vehicle is close to the right lane line and the lateral speed estimate is less than negative one times the least squares estimate variance, the probability value P(l+CutInD <C Ri -(5 / 8)wd)>0.6, then the target vehicle is identified as being in the cut-in state. If P(l+CutInD <C Ri -(5 / 8)wd)≤0.6 and |lC Ri |≤(3 / 8)wd, the target vehicle is identified as being in a line-riding state; otherwise, the target vehicle is not in a cut-in state, a cut-out state, or a line-riding state.

8. A device for identifying the motion state of a preceding vehicle, characterized in that: include: The reference line setting module is used to set the reference line in the Frenet coordinate system according to the existence of the lane line; A first calculation module, configured to calculate a lateral distance from a target vehicle to the reference line based on the reference line; A second calculation module, configured to perform a least squares estimation based on the lateral distance, and calculate an estimated value of a lateral speed of the target vehicle relative to the reference line and a least squares estimation variance; An area calibration module, for calibrating a line-riding area, a cut-in area, and a cut-out area based on the reference line and the width of the target vehicle; A first identification module, configured to identify that the target vehicle is in a line-riding state when the lateral speed estimation value is less than the least squares estimation variance and the lateral distance is in the calibrated line-riding area; The second identification module is used to calculate the probability value of the target vehicle being in the cut-in area or the cut-out area through the lateral distance when the lateral speed estimation value is greater than or equal to the least squares estimation variance and the lateral distance is in the calibrated cut-in area or the cut-out area, and identify the running state of the target vehicle according to the size of the probability value.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for identifying the motion state of a leading vehicle according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program for executing the preceding vehicle motion state recognition method according to any one of claims 1 to 7.

Images