Method for distinguishing front vehicle lane based on radar echo energy target in curve driving

Abstract

The invention discloses a method for distinguishing a front vehicle lane by a target based on radar echo energy in curve driving, which comprises a learning stage and a detection stage, and is characterized in that the learning stage comprises the following steps: step 1, collecting driving sample data in a curve driving process, and constructing a first learning data set; 2, synchronously recording a driving video in the process of collecting the driving sample data; 3, classifying the driving sample data according to the relative lane position relation by using the driving video; step 4, classifying each second learning data set according to intervals divided by the relative speed; 5, performing discrete degree analysis on the millimeter wave radar echo energy value of each third learning data set, and constructing a fourth learning data set; the detection stage comprises the steps of comparing the current discrete degree with the fourth learning data set, and determining the relative lane position relation between the vehicle and the front vehicle. According to the invention, the false detection rate of the front collision early warning system in the curve driving scene can be reduced.

Description

Technical Field

[0001] The invention belongs to the field of automobile active safety, and in particular relates to a method for distinguishing a lane of a preceding vehicle based on a target of radar echo energy during curved road driving. Background Art

[0002] The forward collision warning system based on millimeter-wave radar relies on millimeter-wave radar to detect the relative distance and speed between the vehicle and the vehicle in front, and implements forward collision warning according to the corresponding warning strategy, or actively intervenes to brake and decelerate when necessary. However, this technical idea has a high misjudgment rate in curved driving scenarios. The reason is that when driving on a curve, the vehicle in front of the vehicle in the adjacent lane is often directly in front of the millimeter-wave radar of the vehicle. When the relative distance and speed of the two meet the warning conditions, the warning system or active braking will be triggered. However, in fact, since the two are in different lanes, there is no risk of collision at this time, which is a misjudgment. Therefore, it is necessary to perform lane-level judgment on the relative position relationship between the vehicle in front and the vehicle to overcome the problem of false detection in curved driving scenarios.

[0003] In view of this, the inventor provides a method for distinguishing the lane of a preceding vehicle based on the target of radar echo energy when driving on a curve, so as to solve the above technical problems. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art and propose a method for distinguishing the lane of the preceding vehicle based on the target of radar echo energy during curved driving, so as to solve the problem of false detection of the forward collision warning system based on millimeter-wave radar during curved driving.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] The present invention provides a method for distinguishing a lane of a preceding vehicle based on a target of radar echo energy during curved driving, comprising a learning phase and a detection phase, wherein the learning phase comprises the following steps:

[0007] Step 1: Collect driving sample data during the curve driving process, including the relative speed, relative angle and millimeter wave radar echo energy value between the vehicle and the vehicle in front, to construct a first learning data set;

[0008] Step 2: While collecting driving sample data, the driving video is recorded synchronously, and the lane-level relative position relationship between the preceding vehicle and the vehicle itself is distinguished;

[0009] Step 3: using the driving video, classify the driving sample data of step 1 according to the lane-level relative position relationship to obtain a plurality of second learning data sets;

[0010] Step 4: Classify each of the intervals divided by the relative speed of the second learning data set to obtain a plurality of third learning data sets;

[0011] Step 5: performing a discrete degree analysis on each millimeter wave radar echo energy value of the third learning data set, and applying the discrete degree analysis result to construct a fourth learning data set;

[0012] The detection stage includes: obtaining the discrete degree of the current driving millimeter wave radar echo energy value, comparing the current discrete degree with the fourth learning data set, and determining the lane-level relative position relationship between the vehicle and the vehicle in front;

[0013] Among them, in step 1, the relative speed of the collected driving sample data is that the speed of the vehicle itself is greater than the speed of the front vehicle.

[0014] Furthermore, in step one, the front vehicle includes the front vehicle traveling in the current lane of the vehicle, the left lane of the vehicle, and the right lane of the vehicle.

[0015] Furthermore, in step 2, the lane-level relative position relationship between the leading vehicle and the vehicle includes the leading vehicle being located in the current lane of the vehicle, the leading vehicle being located in the left lane of the vehicle, and the leading vehicle being located in the right lane of the vehicle.

[0016] Furthermore, in step 3, the process of obtaining the second learning data set is as follows:

[0017] Corresponding to the three types of lane-level relative position relationships, namely, the front vehicle is located in the current lane of the vehicle, the front vehicle is located in the lane on the left of the vehicle, and the front vehicle is located in the lane on the right of the vehicle, the collected driving sample data are divided into three groups to obtain three groups of second learning data sets, which are expressed as Among them, i is the relative position relationship of the three lane levels.

[0018] Furthermore, in step 4, the process of obtaining the third learning data set is as follows:

[0019] First, the relative speed of the vehicle and the preceding vehicle in the driving sample data is divided into intervals, and then the three groups of second learning data sets are classified according to the divided intervals to obtain multiple groups of third learning data sets, which are expressed as Among them, j is the number of divided relative speed intervals x;

[0020] Wherein, the relative velocity is given by v re =v s -v f Indicates that v s is the vehicle speed, v f is the speed of the preceding vehicle, and v s Greater than v f .

[0021] Specifically, since the forward collision warning system only focuses on the vehicle speed v s Greater than the speed of the vehicle ahead v f For such a scenario with collision risk, we only need to consider the case where the relative velocity is positive.

[0022] Furthermore, in step 5, the discrete degree of the millimeter wave radar echo energy value is calculated by the following formula:

[0023]

[0024] Formula (1) describes the radar echo energy collected within the time window t A total of N times, is the radar echo energy variance, that is, the discrete degree of millimeter wave radar echo energy value;

[0025] is the mean echo energy;

[0026] The fourth learning data set is represented as

[0027] Furthermore, the specific process of the detection phase is as follows:

[0028] (1) judging whether the vehicle is currently in a curve based on the gyroscope data built into the forward collision warning system; if the vehicle is currently in a curve, proceeding to step 2; otherwise, not executing step 2;

[0029] (2) Determine the position distribution characteristics of the two vehicles based on the relative angle between the vehicle and the preceding vehicle. If the preceding vehicle and the vehicle meet the preset conditions, proceed to step (3); otherwise, do not execute step (3).

[0030] (3) The discrete degree of the millimeter-wave radar echo energy value of the current vehicle and Compare and determine the lane-level relative position relationship between the vehicle and the vehicle in front. The comparison conditions are as follows:

[0031]

[0032] Furthermore, in step (1), whether the vehicle is currently in a curve driving state is determined based on the gyroscope z-axis angular velocity w z Make a judgment, the judgment conditions are as follows:

[0033]

[0034] in, is the angular velocity threshold, and the counterclockwise direction is the positive direction of the angular velocity.

[0035] Furthermore, in step (2), the preset condition is: when the vehicle turns right, the vehicle in front is located on the left side of the vehicle; or when the vehicle turns left, the vehicle in front is located on the right side of the vehicle.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] The present invention provides a method for distinguishing the lane of a leading vehicle based on a target of radar echo energy during curved driving. Compared with the existing technical solutions, the method takes into account the lane-level relative position relationship between the leading vehicle and the vehicle itself in the perception stage of the forward collision warning system, thereby reducing the false detection rate of the forward collision warning system in the curved driving scene; and in the process of lane-level relative position relationship detection, multiple types of actual driving sample data (relative speed, relative angle, radar echo energy value) are considered to construct a learning data set and the relative position relationship is analyzed based on the discrete degree of radar echo energy, which is conducive to accurately detecting the lane-level relative position relationship between the leading vehicle and the vehicle itself; in addition, during the detection process, the method disclosed by the present invention can be adaptively applied to various types of vehicle models, road curvatures, and the number of road lanes, and has extremely high universality; in actual application, there is no need to upgrade the hardware of the existing forward collision warning system, and lane-level relative position detection can be achieved only by relying on its built-in millimeter-wave radar and gyroscope, so the method disclosed by the present invention does not require the introduction of additional hardware costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings are incorporated in and constitute a part of this specification and, together with the description, serve to explain the principles of the present invention.

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0040] Figure 1 It is a flow chart of the learning phase of the method for distinguishing the lane of the preceding vehicle based on the target of radar echo energy of the present invention;

[0041] Figure 2 It is a flow chart of the detection phase of the method for distinguishing the lane of the preceding vehicle based on the target of radar echo energy of the present invention. DETAILED DESCRIPTION

[0042] Here, exemplary embodiments will be described in detail, and the embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are only examples consistent with some aspects of the present invention described in detail in the appended claims.

[0043] See also Figure 1 , Figure 2 The embodiment of the present invention provides a method for distinguishing a lane of a preceding vehicle based on a target of radar echo energy when driving on a curve, comprising a learning phase and a detection phase, wherein the learning phase comprises the following steps:

[0044] Step 1: Based on the millimeter-wave radar built into the forward collision warning system, driving sample data during the curve driving process is collected, including the relative speed, relative angle and millimeter-wave radar echo energy value between the vehicle and the vehicle in front, so as to construct a first learning data set;

[0045] Step 2: During the process of collecting driving sample data, the vehicle-mounted driving recorder synchronously records the driving video, and the lane-level relative position relationship between the preceding vehicle and the vehicle itself is manually distinguished based on the video;

[0046] Step 3: using the driving video, classify the driving sample data of step 1 according to the lane-level relative position relationship to obtain a plurality of second learning data sets;

[0047] Step 4: Classify each of the intervals divided by the relative speed of the second learning data set to obtain a plurality of third learning data sets;

[0048] Step 5: performing a discrete degree analysis on each millimeter wave radar echo energy value of the third learning data set, and applying the discrete degree analysis result to construct a fourth learning data set;

[0049] The detection stage includes: obtaining the discrete degree of the current driving millimeter wave radar echo energy value, comparing the current discrete degree with the fourth learning data set, and determining the lane-level relative position relationship between the vehicle and the vehicle in front;

[0050] Among them, in step 1, the relative speed of the collected driving sample data is that the speed of the vehicle itself is greater than the speed of the front vehicle.

[0051] It should be noted that this application only considers the situation where the speed of the vehicle is higher than that of the front vehicle, in which case there is a risk of rear-end collision. Taking into account that when driving on a curve, as the vehicle approaches the front vehicle, the different lane-level relative position relationships between the front vehicle and the vehicle will cause the body area (reflection area) of the front vehicle exposed to the millimeter-wave radar of the vehicle to produce different change characteristics. Conventionally, when driving on a left turn, as the vehicle approaches the front vehicle, if the front vehicle is in the current lane of the vehicle, the reflection area of ​​the front vehicle will hardly change. If the vehicle is in the right lane of the vehicle, the reflection area of ​​the front vehicle will show nonlinear change characteristics. According to the basic principle of millimeter-wave radar, the size of the reflection area of ​​the front vehicle will directly affect the size of the radar echo energy value, so the discrete degree of the echo energy value can be used to characterize the lane-level relative position relationship between the front vehicle and the vehicle.

[0052] In order to make the above characterization relationship more accurate, considering that the relative speed between the two vehicles will directly affect the change rate of the reflection area of ​​the front vehicle, and further affect the change rate of the radar echo energy value, it is necessary to learn the discrete degree of radar echo energy in different relative speed intervals.

[0053] Furthermore, in step one, the front vehicle includes the front vehicle traveling in the current lane of the vehicle, the lane on the left of the vehicle, and the lane on the right of the vehicle.

[0054] Furthermore, in step 2, there are three types of lane-level relative position relationships, including the front vehicle being located in the current lane of the vehicle, the front vehicle being located in the left lane of the vehicle, and the front vehicle being located in the right lane of the vehicle.

[0055] Furthermore, in step 3, the process of obtaining the second learning data set is as follows: first, using the driving video synchronously collected in step 2, the actual driving sample data collected in step 1 is divided into 3 groups: The first classification is completed, and three sets of second learning data sets are obtained, corresponding to the three relative position relationships of the front vehicle in the current lane of the vehicle, the front vehicle in the left lane of the vehicle, and the front vehicle in the right lane of the vehicle.

[0056] Furthermore, in step 4, the process of obtaining the third learning data set is as follows: considering the influence of the relative speed of the two vehicles (the vehicle itself and the front vehicle) on the rate of change of the radar echo energy, the relative speed of the vehicle itself and the front vehicle of the driving sample data is first divided into intervals, and then the three groups of the second learning data sets are classified according to the divided intervals to obtain multiple groups of third learning data sets, which are expressed as Among them, j is the number of divided relative speed intervals x;

[0057] Wherein, the relative velocity is given by v re =v s -v f Indicates that v s is the vehicle speed, v f is the speed of the preceding vehicle, and v s Greater than v f .

[0058] Optionally, the relative speed interval division result between the vehicle and the preceding vehicle is:

[0059]

[0060] Among them, let v re (km / h) is the relative speed between the preceding vehicle and the vehicle itself. There are 4 groups of different relative speed ranges.

[0061] Since the forward collision warning system only focuses on the vehicle speed vs Greater than the speed of the vehicle ahead v f For such a scenario with collision risk, we only need to consider the case where the relative velocity is positive as shown in equation (1).

[0062] After completing the above two classifications, the four sets of third learning data sets are represented as follows:

[0063]

[0064] Since the discrete degree of radar echo energy can be used to characterize the lane-level relative position relationship between the preceding vehicle and the vehicle, in step 5, the discrete degree of the millimeter-wave radar echo energy value is calculated by the following formula:

[0065]

[0066] Formula (3) describes the radar echo energy collected within the time window t A total of N times, is the echo energy variance, that is, the discrete degree of millimeter wave radar echo energy value;

[0067] is the mean echo energy;

[0068] The fourth learning data set is represented as Among them, i is the relative position relationship of the three lane levels, and j is the number of divided relative speed intervals x.

[0069] Specifically, the fourth learning data set in this embodiment Corresponding to formula (2), use To characterize the relative position relationship of three lane levels.

[0070] Furthermore, the specific process of the detection phase is as follows:

[0071] (1) judging whether the vehicle is currently in a curve based on the gyroscope data built into the forward collision warning system; if the vehicle is currently in a curve, proceeding to step 2; otherwise, not executing step 2;

[0072] Optionally, use a gyroscope sensor to collect vehicle angular velocity data. Usually, angular velocity data will have three components (x, y, z), but for vehicle turning, the angular velocity of the z-axis (vertical axis) is the main focus; then, the collected angular velocity data is filtered to remove noise. Optionally, a low-pass filter is used. Finally, according to the vehicle characteristics and experimental data, the angular velocity threshold is set The counterclockwise direction is defined as the positive direction of angular velocity. According to the angular velocity w of the gyroscope z axis z Determine the driving state of the vehicle, and the judgment conditions are as follows:

[0073]

[0074] (2) Determine the position distribution characteristics of the two vehicles based on the relative angle between the vehicle and the preceding vehicle. If the preceding vehicle and the vehicle meet the preset conditions, proceed to step (3); otherwise, do not execute step (3).

[0075] Specifically, the relative angle α between the two vehicles detected by the millimeter wave radar can be used to reflect the position distribution characteristics of the two vehicles. Optionally, the forward direction of the vehicle is set as the angle starting axis, and the counterclockwise angle is a positive value. Therefore, the position distribution characteristics of the front vehicle and the vehicle are determined by formula (5):

[0076]

[0077] It should be noted that the position distribution characteristics of the leading vehicle and the ego vehicle described in formula (5) are not the relative lane positions of the two vehicles. For example, when α>0, the leading vehicle is located on the left side of the ego vehicle. At this time, the leading vehicle may be in the current lane of the ego vehicle or in the lane on the left side of the ego vehicle.

[0078] The preset condition is: when the vehicle turns right, the vehicle in front is on the left side of the vehicle; or when the vehicle turns left, the vehicle in front is on the right side of the vehicle. Specifically, it can be expressed by the following formula:

[0079]

[0080] (3) The discrete degree of the millimeter-wave radar echo energy value of the current vehicle and Compare and determine the lane-level relative position relationship between the vehicle and the vehicle in front. The comparison conditions are as follows:

[0081]

[0082] in, is the current radar echo energy variance, and the corresponding relative speed is

[0083] when When , the specific lane-level relative position relationship between the preceding vehicle and the vehicle itself is determined by the following formula:

[0084]

[0085] In summary, the conditions under which the front vehicle may cause the forward collision warning system to falsely alarm are as follows:

[0086]

[0087] In the embodiment of the present invention, when the front vehicle is located in the current lane of the vehicle, compared with other lane-level relative positions (the front vehicle is located in the left lane of the vehicle, the front vehicle is located in the right lane of the vehicle), the radar echo energy presents different numerical characteristics, which can be characterized by the energy variance, as shown in formula (3) of this embodiment. Therefore, in order to further determine the relative lane position between the front vehicle and the vehicle, it is necessary to first use the relative speed v of the two vehicles fed back by the millimeter-wave radar re And determine the relative speed interval according to formula (1); at the same time, calculate the current radar echo energy variance in the time window t according to formula (3) and compare it with the fourth learning data set of radar echo energy statistics in the corresponding relative speed interval in the test phase For comparison, the result of the judgment of formula (6) in the embodiment is matched with the lane-level position relationship in the test phase. For example, when the vehicle turns right and the front vehicle is located on the left side of the vehicle, the lane-level position relationship corresponding to the test phase is that the front vehicle is located on the left lane of the vehicle. The relative lane positions of the front vehicle and the vehicle are determined by comparison to avoid causing false alarms to the system.

[0088] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention.

[0089] It should be understood that the present invention is not limited to what has been described above and that various modifications and changes may be made without departing from its scope. The scope of the present invention is limited only by the appended claims.

Claims

1. A method for distinguishing the lane of a vehicle ahead based on radar echo energy during driving on a curve, characterized in that: It includes a learning phase and a detection phase, wherein the learning phase includes the following steps: Step 1: Collect driving sample data during the curve driving process, including the relative speed, relative angle and millimeter wave radar echo energy value between the vehicle and the vehicle in front, to construct a first learning data set; Step 2: While collecting driving sample data, the driving video is recorded synchronously, and the lane-level relative position relationship between the preceding vehicle and the vehicle itself is distinguished; Step 3: using the driving video, classify the driving sample data of step 1 according to the lane-level relative position relationship to obtain a plurality of second learning data sets; Step 4: Classify each of the intervals divided by the relative speed of the second learning data set to obtain a plurality of third learning data sets; Step 5: performing a discrete degree analysis on each millimeter wave radar echo energy value of the third learning data set, and applying the discrete degree analysis result to construct a fourth learning data set; The detection stage includes: obtaining the discrete degree of the current driving millimeter wave radar echo energy value, comparing the current discrete degree with the fourth learning data set, and determining the lane-level relative position relationship between the vehicle and the vehicle in front; Among them, in step 1, the relative speed of the collected driving sample data is that the speed of the vehicle itself is greater than the speed of the front vehicle.

2. The method for distinguishing the lane of the vehicle ahead based on the target of radar echo energy when driving on a curve according to claim 1, characterized in that: In step 1, the front vehicle includes the front vehicle located in the current lane of the vehicle, the left lane of the vehicle, and the right lane of the vehicle.

3. The method for distinguishing the lane of the vehicle ahead based on the target of radar echo energy when driving on a curve according to claim 1, characterized in that: In step 2, the lane-level relative position relationship between the front vehicle and the vehicle includes the front vehicle being located in the current lane of the vehicle, the front vehicle being located in the left lane of the vehicle, and the front vehicle being located in the right lane of the vehicle.

4. The method for distinguishing the lane of the vehicle ahead based on the target of radar echo energy when driving on a curve according to claim 3, characterized in that: In step 3, the process of obtaining the second learning data set is as follows: Corresponding to the three relative lane position relationships of the front vehicle in the current lane of the vehicle, the front vehicle in the lane on the left of the vehicle, and the front vehicle in the lane on the right of the vehicle, the collected driving sample data are divided into three groups to obtain three groups of second learning data sets, which are expressed as Among them, i is the relative position relationship of the three lane levels.

5. The method for distinguishing the lane of the vehicle ahead based on the target of radar echo energy when driving on a curve according to claim 4, characterized in that: In step 4, the process of obtaining the third learning data set is as follows: First, the relative speed of the vehicle and the preceding vehicle in the driving sample data is divided into intervals, and then the three groups of second learning data sets are classified according to the divided intervals to obtain multiple groups of third learning data sets, which are expressed as Among them, j is the number of divided relative speed intervals x; Wherein, the relative velocity is given by v re =v s -v f Indicates that v s is the vehicle speed, v f is the speed of the preceding vehicle, and v s Greater than v f .

6. The method for distinguishing the lane of the vehicle ahead based on the target of radar echo energy when driving on a curve according to claim 5, characterized in that: In step 5, the discrete degree of the millimeter wave radar echo energy value is calculated by the following formula: Formula (1) describes the radar echo energy collected within the time window t A total of N times, is the radar echo energy variance, that is, the discrete degree of millimeter wave radar echo energy value; is the mean echo energy; The fourth learning data set is represented as 7. The method for distinguishing the lane of the vehicle ahead based on the target of radar echo energy when driving on a curve according to claim 1, characterized in that: The specific process of the detection phase is as follows: (1) judging whether the vehicle is currently in a curve based on the gyroscope data built into the forward collision warning system; if the vehicle is currently in a curve, proceeding to step 2; otherwise, not executing step 2; (2) Determine the position distribution characteristics of the two vehicles based on the relative angle between the vehicle and the preceding vehicle. If the preceding vehicle and the vehicle meet the preset conditions, proceed to step (3); otherwise, do not execute step (3). (3) The discrete degree of the millimeter-wave radar echo energy value of the current vehicle and Compare and determine the lane-level relative position relationship between the vehicle and the vehicle in front. The comparison conditions are as follows:

8. The method for distinguishing the lane of the vehicle ahead based on the target of radar echo energy when driving on a curve according to claim 7, characterized in that: In step (1), whether the vehicle is currently in a curve driving state is determined based on the gyroscope z-axis angular velocity w z Make a judgment, the judgment conditions are as follows: in, is the angular velocity threshold, and the counterclockwise direction is the positive direction of the angular velocity.

9. The method for distinguishing the lane of the vehicle ahead based on the target of radar echo energy when driving on a curve according to claim 7, characterized in that: In step (2), the preset condition is: when the vehicle turns right, the vehicle in front is on the left side of the vehicle; or when the vehicle turns left, the vehicle in front is on the right side of the vehicle.

Images