Other vehicle behavior mining method and device, electronic equipment and storage medium

Abstract

The invention discloses an other vehicle behavior mining method and device, electronic equipment and a storage medium, and the method comprises the steps: screening out the sequence information of an own vehicle and other vehicles meeting a mining condition, and the sequence information comprises a vehicle position, a YAW angle, a vehicle size and timestamp information; according to the sequence information of the own vehicle and the other vehicle meeting the mining condition, the transverse displacement variation of the own vehicle and the other vehicle is judged; and if the transverse displacement variations of the own vehicle and the other vehicle meet respective corresponding conditions, recording the sample data as the behavior sample data of the other vehicle. According to the method, on one hand, the other vehicle behavior mining result with high accuracy can be obtained, and on the other hand, other vehicle behavior mining marked manually is replaced.

Description

Technical Field

[0001] This application relates to the technical field of vehicle behavior mining, and particularly to a method and device for mining the behavior of other vehicles, an electronic device, and a storage medium. Background Art

[0002] The cutin behavior refers to the behavior that when a vehicle is driving, another vehicle suddenly cuts into the current lane in front. The recognition and prediction of cutin behavior in autonomous driving have always been key challenges for vehicle safe driving. Especially in complex urban traffic environments and on highways, if the cutin behavior of other vehicles occurs frequently, it is extremely risky.

[0003] The current safety of autonomous driving decision-making systems mainly depends on the rapid judgment of the behavior of other vehicles, and the recognition of the cutin behavior of other vehicles is one of the main difficulties. If reliable cutin labels of other vehicles can be mined and the driving intention of the cutin behavior of other vehicles can be recognized, the safety and reliability of the decision-making of the autonomous driving system can be greatly improved. Summary of the Invention

[0004] Embodiments of this application provide a method and device for mining the behavior of other vehicles, an electronic device, and a storage medium to improve the accuracy of mining the behavior of other vehicles.

[0005] Embodiments of this application adopt the following technical solutions:

[0006] In a first aspect, embodiments of this application provide a method for mining the behavior of other vehicles, where the method includes:

[0007] Screen out the sequence information of the host vehicle and other vehicles that meet the mining conditions, where the sequence information includes: vehicle position, YAW angle, vehicle size, and timestamp information;

[0008] According to the sequence information of the host vehicle and other vehicles that meet the mining conditions, judge the lateral displacement change amount between the host vehicle and other vehicles;

[0009] If the lateral displacement change amounts between the host vehicle and other vehicles all meet their respective corresponding conditions, record them as sample data of the behavior of other vehicles.

[0010] In some embodiments, the judging the lateral displacement change amount between the host vehicle and other vehicles according to the sequence information of the host vehicle and other vehicles that meet the mining conditions includes:

[0011] According to the sequence information of the host vehicle and other vehicles that meet the mining conditions, sequentially judge whether any one or more of the lateral displacement changes in the lateral distance, lateral offset, and minimum lateral distance in the lateral displacement change amount between the host vehicle and other vehicles are associated with the preset driving behavior of other vehicles;

[0012] If it is determined that all the lateral displacement changes are associated with the preset driving behaviors of other vehicles, continue to use the sequence information of the host vehicle and other vehicles that meet the mining conditions;

[0013] If it is determined that one of the lateral displacement changes is not associated with the preset driving behaviors of other vehicles, discard the sequence information of the host vehicle and other vehicles that meet the mining conditions.

[0014] In some embodiments, if the lateral displacement changes between the host vehicle and other vehicles both meet their respective corresponding conditions, record them as sample data of other vehicle behaviors, including:

[0015] Judge the lateral distance between the host vehicle and other vehicles;

[0016] If it is determined that the lateral distance between the host vehicle and other vehicles meets the preset threshold, calculate the lateral offset between the host vehicle and other vehicles in the future time period;

[0017] If the lateral offset conforms to the behavior of other vehicles approaching the host vehicle, judge the minimum lateral distance between the host vehicle and other vehicles in the future time period;

[0018] If it is determined that the minimum lateral distance between the host vehicle and other vehicles in the future time period meets the preset threshold, record it as positive sample data of other vehicle behaviors.

[0019] In some embodiments, the preset driving behaviors of other vehicles include the cut-in behavior of other vehicles. The sample data recorded as other vehicle behaviors includes:

[0020] Take the moment of the positive sample data of other vehicle behaviors as the start time of the cut-in of other vehicles;

[0021] Based on the start time and the empirical value, determine the end time of the cut-in of other vehicles;

[0022] If the minimum lateral distance time is less than the empirical value, take the empirical value as the shortest time, and the minimum lateral distance time is determined by the minimum lateral distance;

[0023] If the minimum lateral distance time is greater than the empirical value, take the moment when the minimum lateral distance occurs as the end time of the cut-in of other vehicles.

[0024] In some embodiments, the screening of the sequence information of the host vehicle and other vehicles that meet the mining conditions includes:

[0025] Based on the sequence information of other vehicles, determine whether the status of other vehicles meets the first mining requirement, where the first mining requirement includes at least one of the following: the existence time of other vehicles meets the requirements, other vehicles are in a non - stationary state, the horizontal and longitudinal distances between other vehicles and the host vehicle meet all possible conditions for other vehicle behavior mining, other vehicles and the host vehicle are in the same direction, and other vehicles and the host vehicle are in different lanes;

[0026] If it is determined that the status of other vehicles all meets the first mining requirement, then filter out the sequence information of other vehicles that meet the mining conditions.

[0027] In some embodiments, the filtering out of the sequence information of the host vehicle and other vehicles that meet the mining conditions further includes:

[0028] Based on the sequence information of the host vehicle, determine whether the status of the host vehicle meets the second mining requirement, where the second mining requirement includes at least one of the following: the existence time of the host vehicle exceeds a first time, the host vehicle exists for more than a second time in a future time period, and the host vehicle is in a non - stationary state;

[0029] If it is determined that the status of the host vehicle all meets the second mining requirement, then filter out the sequence information of the host vehicle that meets the mining conditions.

[0030] In some embodiments, the method further includes:

[0031] Filter out adjacent obstacles in the horizontal direction of the host vehicle and vehicles in the same lane in the vertical direction of the host vehicle in the current scenario;

[0032] Record the adjacent obstacles in the horizontal direction of the host vehicle and the vehicles in the same lane in the vertical direction of the host vehicle as negative sample data for other vehicle behavior.

[0033] In a second aspect, an embodiment of the present application further provides an other vehicle behavior mining device, where the device includes:

[0034] A filtering module, configured to filter out the sequence information of the host vehicle and other vehicles that meet the mining conditions, where the sequence information includes: vehicle position, YAW angle, vehicle size, and timestamp information;

[0035] A judgment module, configured to judge the lateral displacement change amount between the host vehicle and other vehicles according to the sequence information of the host vehicle and other vehicles that meet the mining conditions;

[0036] A recording module, configured to record as sample data for other vehicle behavior if the lateral displacement change amounts between the host vehicle and other vehicles all meet their respective corresponding conditions.

[0037] In a third aspect, an embodiment of the present application further provides an electronic device, including: a processor; and a memory arranged to store computer - executable instructions, where the executable instructions, when executed, cause the processor to execute the above - mentioned method.

[0038] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, cause the electronic device to execute the above method.

[0039] The above at least one technical solution adopted in the embodiment of the present application can achieve the following beneficial effects: First, filter out the sequence information of the host vehicle and other vehicles that meet the mining conditions, and then judge the lateral displacement change amount between the host vehicle and other vehicles according to the sequence information of the host vehicle and other vehicles that meet the mining conditions. In this way, if the lateral displacement change amounts between the host vehicle and other vehicles all meet their respective corresponding conditions, they are recorded as sample data of other vehicle behaviors. Through the above method, for sudden lane-changing or cutting-in behaviors, an intention data set of other vehicle behaviors can be constructed, thereby improving the decision-making ability and safety of the autonomous driving system. In addition, automatic labeling can be completed before training the algorithm according to the sample data of other vehicle behaviors, reducing the labeling cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings described herein are used to provide a further understanding of the present application, and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application, and do not constitute an improper limitation to the present application. In the drawings:

[0041] FIG. 1(a) is a schematic diagram of the relationship between the host vehicle and other vehicles in an embodiment of the present application;

[0042] FIG. 1(b) is a schematic diagram of the cut-in behavior of other vehicles in an embodiment of the present application;

[0043] Figure 2 is a schematic flowchart of a method for mining other vehicle behaviors in an embodiment of the present application;

[0044] Figure 3 is a schematic structural diagram of a device for mining other vehicle behaviors in an embodiment of the present application;

[0045] Figure 4 is a schematic structural diagram of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] To make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions of the present application will be clearly and completely described below in conjunction with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0047] At present, there are few studies on the cutin intention of other cars, and there are few related mining algorithms. The main difficulties are:

[0048] (1) The suddenness and randomness of the other car’s behavior. The cutin behavior of other cars is often sudden and difficult to accurately predict using traditional historical trajectory prediction models. For example, when other cars suddenly change lanes, the vehicle may not have time to react, thus causing safety hazards.

[0049] (2) The relevant mining algorithms have limitations. There are few studies on cutin mining algorithms, and most of them are done through manual labeling. This method brings two hidden dangers. One is that different labelers have different understandings of cutin behavior. This inconsistency may affect the performance of the model and cause cutin recognition to be too aggressive or conservative. The other is to rely on manual labeling, which is costly and has slow iteration efficiency. If it is not appropriate, it needs to be re-labeled, which has poor versatility and efficiency.

[0050] In view of the above shortcomings, a method for mining other vehicle behaviors is provided in an embodiment of the present application, which realizes automatic mining of other vehicle behaviors and can mark the start time and end time of other vehicle behaviors.

[0051] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0052] The present application embodiment provides a method for mining other vehicle behaviors, such as Figure 2 As shown, a flow chart of a method for mining other vehicle behaviors in an embodiment of the present application is provided, and the method at least includes the following steps S210 to S230:

[0053] Step S210 , filtering out sequence information of the own vehicle and other vehicles that meet the mining conditions, wherein the sequence information includes: vehicle position, YAW angle, vehicle size, and timestamp information.

[0054] After preprocessing, the data is obtained to obtain the sequence information of the self-vehicle and other vehicles. The sequence information is based on the time sequence. Each sequence information contains the vehicle position, YAW angle, vehicle size, and has corresponding timestamp information. In this way, the self-vehicle and other vehicle information with time sequence can be obtained. The timestamp information can ensure that the sequence information of the self-vehicle and other vehicles is aligned.

[0055] It can be understood that the vehicle position can be determined by the combined positioning method of IMU+GNSS. The YAW angle is a parameter for the vehicle posture transformation. The vehicle size is the external dimensions of the vehicle or other vehicles, usually referring to the length and width information.

[0056] As shown in Figure 1(a), the "other vehicle" and the "own vehicle" are traveling in the same direction. The "other vehicle" may appear behind the "own vehicle" in the same lane, or the "other vehicle" may be parallel to the "own vehicle" and in a different lane.

[0057] As shown in Figure 1(b), during the process of the "own vehicle" traveling in its own lane, there is an "other vehicle" suddenly cutting into the current lane ahead, which is the cut-in behavior of the other vehicle.

[0058] Step S220: According to the sequence information of the own vehicle and the other vehicle that meet the mining conditions, judge the lateral displacement change amount between the own vehicle and the other vehicle.

[0059] Judge whether the state of the own vehicle meets the mining requirements. If not, discard the data of the current frame. If it meets, continue to judge the state of the other vehicle. Judge whether the state of the other vehicle meets the mining requirements. If not, discard the data of the current frame. If both meet, perform the judgment of the other vehicle's behavior. Similarly, if not, discard the data of the current frame.

[0060] Furthermore, through the sequence information of the own vehicle and the other vehicle that meet the mining conditions obtained by screening, the lateral displacement change amount between the own vehicle and the other vehicle is judged. It can be understood that the lateral displacement change amount includes but is not limited to: the current lateral distance between the own vehicle and the other vehicle, the lateral offset of the own vehicle in a future period of time and the other vehicle in a future period of time, etc.

[0061] Step S230: If the lateral displacement change amounts between the own vehicle and the other vehicle both meet their respective corresponding conditions, record them as sample data of the other vehicle's behavior.

[0062] If the lateral displacement change amounts between the own vehicle and the other vehicle both meet their respective corresponding conditions, it can be recorded as the other vehicle's behavior to be mined and used as sample data. It should be noted that if any one of the parameters in the lateral displacement change amount does not meet the corresponding conditions, the data of the current frame should be discarded.

[0063] Through the above method, the accuracy of mining the other vehicle's behavior is improved, ensuring that the annotation results are correct. Thus, the algorithm model can learn relevant features well, and the accuracy rate of the model trained by the algorithm can be higher than 95%.

[0064] Through the above method, compared with manual annotation, the mining of the other vehicle's behavior is more reliable, saving labor costs, ensuring data quality, and solving the problem that different personnel have different judgment criteria during manual annotation.

[0065] Different from the related art, the understandings of other vehicles' behaviors by different annotators are different, and such inconsistencies may affect the performance of the model, resulting in problems such as over-aggressive or conservative recognition of other vehicles' behaviors. Through the above method, the sequence information of the host vehicle and other vehicles that meet the mining conditions is screened out. The sequence information includes: vehicle position, YAW angle, vehicle size, and timestamp information; according to the sequence information of the host vehicle and other vehicles that meet the mining conditions, the lateral displacement change amount between the host vehicle and other vehicles is judged; if the lateral displacement change amounts between the host vehicle and other vehicles all meet their respective corresponding conditions, they are recorded as sample data of other vehicles' behaviors.

[0066] Different from the related art, relying on the method of manual labeling has high costs and slow iteration efficiency. If it is not appropriate, relabeling is required, with poor versatility and efficiency. Through the above method, after two-layer judgment: judging the lateral displacement change amount between the host vehicle and other vehicles and judging whether the lateral displacement change amounts between the host vehicle and other vehicles all meet their respective corresponding conditions. Data that meets the requirements can be screened out as positive sample data of other vehicles' behaviors.

[0067] In an embodiment of the present application, the judging of the lateral displacement change amount between the host vehicle and other vehicles according to the sequence information of the host vehicle and other vehicles that meet the mining conditions includes: according to the sequence information of the host vehicle and other vehicles that meet the mining conditions, sequentially judging whether any one or more of the lateral distance, lateral offset, and minimum lateral distance in the lateral displacement change amount between the host vehicle and other vehicles is associated with the preset driving behavior of other vehicles; if it is judged that the lateral displacement change amounts are all associated with the preset driving behavior of other vehicles, continue to use the sequence information of the host vehicle and other vehicles that meet the mining conditions; if it is judged that one of the lateral displacement change amounts is not associated with the preset driving behavior of other vehicles, discard the sequence information of the host vehicle and other vehicles that meet the mining conditions.

[0068] When judging the behavior of other vehicles, it is usually necessary to judge whether the lateral distance between the other vehicle and the host vehicle meets the threshold. And judge whether the lateral offsets between the other vehicle and the host vehicle in the past and future are monotonically decreasing. It is also necessary to judge whether the minimum lateral distance between the other vehicle and the host vehicle in the future is less than the threshold. If not, the data of the current frame (camera image and / or lidar) needs to be discarded.

[0069] Case 1, when it is judged that the lateral displacement change amounts are all associated with the preset driving behavior of other vehicles, continue to use the sequence information of the host vehicle and other vehicles that meet the mining conditions. The preset driving behavior includes the cut-in behavior of the host vehicle.

[0070] Case 2, when it is judged that one of the lateral displacement change amounts is not associated with the preset driving behavior of other vehicles, discard the sequence information of the host vehicle and other vehicles that meet the mining conditions.

[0071] In an embodiment of the present application, if the lateral displacement change amounts between the host vehicle and the other vehicle all meet their respective corresponding conditions, they are recorded as sample data of the other vehicle's behavior, including: determining the lateral distance between the host vehicle and the other vehicle; if it is determined that the lateral distance between the host vehicle and the other vehicle meets a preset threshold, calculating the lateral offset between the host vehicle and the other vehicle in a future time period; if the lateral offset conforms to the behavior of the other vehicle approaching the host vehicle, determining the minimum lateral distance between the host vehicle and the other vehicle in a future time period; if it is determined that the minimum lateral distance between the host vehicle and the other vehicle in a future time period meets a preset threshold, it is recorded as positive sample data of the other vehicle's behavior.

[0072] Taking the cut-in behavior of the other vehicle as an example to illustrate the other vehicle's behavior. When judging the cut-in behavior of the other vehicle:

[0073] (1) Determine whether the lateral distance between the current other vehicle and the host vehicle is less than or equal to half of the width of the other vehicle plus half of the width of the host vehicle plus x meters. At this time, cut-in may occur. x is an empirical value, and the optional value of x can be adjusted to 0.4. Calculate the lateral offset based on the obtained lateral distance.

[0074] (2) Calculate whether the lateral offset between the other vehicle in the future and the host vehicle in the future is monotonically decreasing. Monotonically decreasing means that in mathematics, if a function in a certain interval, as the independent variable increases, the function value gradually decreases. It can be understood that if the lateral offset between the other vehicle in the future and the host vehicle in the future is monotonically decreasing, it is considered that the other vehicle is gradually approaching the host vehicle.

[0075] Specifically, after sampling the data sequences of the host vehicle and the other vehicle at 0.2-second intervals, in a time period from the current moment to the future, taking 1.2 seconds as an example (1.2 seconds is the minimum cut-in time obtained through testing and statistics of the test set), there are a total of 6 points. Among them, if 4 points are monotonically decreasing and it is also monotonous within the historical 1.2 s, it can be considered that the other vehicle has a behavior of approaching the host vehicle.

[0076] (3) Determine the minimum lateral distance between the other vehicle and the host vehicle in the future. If it is less than the sum of half of the widths of the host vehicle and the other vehicle, it can be known that cut-in will definitely occur in the future. It can be understood that this is also the start time of the cut-in behavior of the other vehicle. Usually, the end time = start time + 1.2 seconds. Here, 1.2 seconds is only an example, and the specific value can be determined according to statistics.

[0077] In an embodiment of the present application, the preset driving behavior of the other vehicle includes the cut-in behavior of the other vehicle. The sample data recorded as the behavior of the other vehicle includes: taking the moment of the positive sample data recording the behavior of the other vehicle as the start time when the other vehicle cuts in; determining the end time when the other vehicle cuts in based on the start time and the empirical value; if the minimum lateral distance time is less than the empirical value, taking the empirical value as the shortest time, where the minimum lateral distance time is determined by the minimum lateral distance; if the minimum lateral distance time is greater than the empirical value, taking the moment when the minimum lateral distance occurs as the end time when the other vehicle cuts in.

[0078] When the other vehicle performs a cut-in behavior, there will be continuous lateral displacement changes and finally it will cut in front of the host vehicle. If the time when the current minimum lateral distance occurs is less than 1.2 seconds, the default shortest time is 1.2 seconds. Otherwise, if the time when the minimum lateral distance occurs is greater than 1.2 seconds, the corresponding moment when the minimum lateral distance occurs in (3) is calculated as the cut-in end moment. Here, 1.2 seconds is only an example, and the specific value can be determined according to statistics.

[0079] In an embodiment of the present application, the screening of the sequence information of the host vehicle and the other vehicle that meets the mining conditions includes: judging whether the state of the other vehicle meets the first mining requirement according to the sequence information of the other vehicle, where the first mining requirement includes at least one of the following: the existence time of the other vehicle meets the requirement, the other vehicle is in a non-stationary state, the horizontal and vertical distances between the other vehicle and the host vehicle meet all possible conditions for other vehicle behavior mining, the other vehicle and the host vehicle are in the same direction, and the other vehicle and the host vehicle are in different lanes; if it is judged that the state of the other vehicle meets all the first mining requirements, the sequence information of the other vehicle that meets the mining conditions is screened out.

[0080] Judging whether the state of the other vehicle meets the mining requirements mainly includes 5 conditions: that is, the existence time of the other vehicle meets the requirement, the non-stationary state, the horizontal and vertical distances from the host vehicle meet the possible conditions for cut-in, being basically in the same direction as the host vehicle, and whether it is in the same lane as the host vehicle. It can be understood that only when the above conditions are met, the other vehicle may cut in front of the host vehicle. It can be understood that the existence time of the other vehicle is related to the actual scenario and can be selected according to the corresponding scenario.

[0081] In an embodiment of the present application, the screening of the sequence information of the host vehicle and the other vehicle that meets the mining conditions further includes: judging whether the state of the host vehicle meets the second mining requirement according to the sequence information of the host vehicle, where the second mining requirement includes at least one of the following: the existence time of the host vehicle exceeds the first time, the host vehicle exists for more than the second time in the future time period, the host vehicle is in a non-stationary state; if it is judged that the state of the host vehicle meets all the second mining requirements, the sequence information of the host vehicle that meets the mining conditions is screened out.

[0082] Determine whether the state of the host vehicle meets the mining requirements. For example, the existence time of the host vehicle should exceed 2 seconds, the future existence should exceed 1 second, and the host vehicle cannot be in a stationary state, etc. At this time, if the host vehicle is driving abnormally, the cut-in behavior of other vehicles has no impact on the host vehicle. It can be understood that there are various options for whether the state of the host vehicle meets the mining requirements, and specific limitations are not imposed in the embodiments of the present application.

[0083] In an embodiment of the present application, the method further includes: screening out adjacent obstacles in the horizontal direction of the host vehicle and vehicles in the same lane in the vertical direction of the host vehicle in the current scenario; recording the adjacent obstacles in the horizontal direction of the host vehicle and the vehicles in the same lane in the vertical direction of the host vehicle as negative sample data of the behavior of other vehicles.

[0084] In addition to mining positive samples of the cut-in behavior of other vehicles, negative samples of the cut-in behavior of other vehicles also need to be mined during algorithm training. That is, for vehicles (negative samples) that are relatively easy to be misjudged as cut-in, which are adjacent obstacles in the horizontal direction and vehicles in the same lane in the vertical direction respectively, two types of normally driving vehicles can be screened out as negative samples.

[0085] In an embodiment of the present application, a model training method is also provided, using the above-mentioned method for mining the behavior of other vehicles to obtain sample data for mining the behavior of other vehicles, which is used for training related models.

[0086] An embodiment of the present application also provides a device 300 for mining the behavior of other vehicles, as Figure 3 shown, a structural schematic diagram of the device for mining the behavior of other vehicles in an embodiment of the present application is provided. The device 300 for mining the behavior of other vehicles at least includes: a screening module 310, a judgment module 320, and a recording module 330, where:

[0087] In an embodiment of the present application, the screening module 310 is specifically configured to: screen out sequence information of the host vehicle and other vehicles that meet the mining conditions, and the sequence information includes: vehicle position, YAW angle, vehicle size, and timestamp information.

[0088] The data is preprocessed to obtain sequence information of the host vehicle and other vehicles. The sequence information is in chronological order. Each sequence information contains vehicle position, YAW angle, vehicle size, and has corresponding timestamp information. In this way, chronological information of the host vehicle and other vehicles can be obtained. The timestamp information can ensure the alignment of the sequence information of the host vehicle and other vehicles.

[0089] It can be understood that the vehicle position can be obtained through a combined positioning method of IMU + GNSS. The YAW angle is a parameter for the transformation of the vehicle body posture. The vehicle size is the external dimension of the host vehicle or other vehicles, usually referring to the length and width information.

[0090] As shown in Figure 1(a), the "other vehicle" and the "self vehicle" are traveling in the same direction. The "other vehicle" may appear behind the "self vehicle" in the same lane, or the "other vehicle" may be parallel to the "self vehicle" and in a different lane.

[0091] As shown in Figure 1(b), during the process of the "self vehicle" traveling in its own lane, there is an "other vehicle" suddenly cutting into the current lane in front, which is the cutin behavior of the other vehicle.

[0092] In an embodiment of the present application, the judgment module 320 is specifically configured to: judge the lateral displacement change amount between the self vehicle and the other vehicle according to the sequence information of the self vehicle and the other vehicle that meets the mining conditions.

[0093] Judge whether the state of the self vehicle meets the mining requirements. If not, discard the data of the current frame. If it meets, continue to judge the state of the other vehicle. Judge whether the state of the other vehicle meets the mining requirements. If not, discard the data of the current frame. If both meet, perform the judgment of the other vehicle's behavior. Similarly, if not, discard the data of the current frame.

[0094] Furthermore, the lateral displacement change amount between the self vehicle and the other vehicle is judged through the sequence information of the self vehicle and the other vehicle that meets the mining conditions obtained by screening. It can be understood that the lateral displacement change amount includes but is not limited to: the current lateral distance between the self vehicle and the other vehicle, the lateral offset of the self vehicle in a future period of time and the other vehicle in a future period of time, etc.

[0095] In an embodiment of the present application, the recording module 330 is specifically configured to: if the lateral displacement change amounts between the self vehicle and the other vehicle all meet their respective corresponding conditions, record them as sample data of the other vehicle's behavior.

[0096] If the lateral displacement change amounts between the self vehicle and the other vehicle all meet their respective corresponding conditions, it can be recorded as the other vehicle's behavior that needs to be mined and used as sample data. It should be noted that if any one of the parameters in the lateral displacement change amount does not meet the corresponding conditions, the data of the current frame is discarded.

[0097] In an embodiment of the present application, the judgment module 320 is further configured to

[0098] Judge whether any one or more of the lateral displacement change amounts of the lateral distance, lateral offset, and minimum lateral distance in the lateral displacement change amount between the self vehicle and the other vehicle are associated with the preset driving behavior of the other vehicle according to the sequence information of the self vehicle and the other vehicle that meets the mining conditions;

[0099] If it is judged that the lateral displacement change amounts are all associated with the preset driving behavior of the other vehicle, continue to use the sequence information of the self vehicle and the other vehicle that meets the mining conditions;

[0100] If it is determined that one of the lateral displacement changes is not associated with the preset driving behavior of the other vehicle, the sequence information of the host vehicle and the other vehicle that meet the mining conditions is discarded.

[0101] In one embodiment of the present application, the recording module 330 is further configured to

[0102] Determine the lateral distance between the host vehicle and the other vehicle;

[0103] If it is determined that the lateral distance between the host vehicle and the other vehicle meets the preset threshold, calculate the lateral offset between the host vehicle and the other vehicle within the future time period;

[0104] If the lateral offset conforms to the behavior of the other vehicle approaching the host vehicle, determine the minimum lateral distance between the host vehicle and the other vehicle within the future time period;

[0105] If it is determined that the minimum lateral distance between the host vehicle and the other vehicle within the future time period meets the preset threshold, record it as the positive sample data of the other vehicle's behavior.

[0106] In one embodiment of the present application, the preset driving behavior of the other vehicle includes the cut-in behavior of the other vehicle. The recording module 330 is further configured to

[0107] Take the moment when the positive sample data of the other vehicle's behavior is recorded as the start time of the other vehicle's cut-in;

[0108] Based on the start time and the empirical value, determine the end time of the other vehicle's cut-in;

[0109] If the minimum lateral distance time is less than the empirical value, use the empirical value as the shortest time. The minimum lateral distance time is determined by the minimum lateral distance;

[0110] If the minimum lateral distance time is greater than the empirical value, take the moment when the minimum lateral distance occurs as the end time of the other vehicle's cut-in.

[0111] In one embodiment of the present application, the screening module 310 is further configured to

[0112] According to the sequence information of the other vehicle, determine whether the state of the other vehicle meets the first mining requirement. The first mining requirement includes at least one of the following: the existence time of the other vehicle meets the requirement, the other vehicle is in a non-stationary state, the lateral and longitudinal distances between the other vehicle and the host vehicle meet all possible conditions for other vehicle behavior mining, the other vehicle and the host vehicle are in the same direction, and the other vehicle and the host vehicle are in different lanes;

[0113] If it is determined that all the states of the other vehicle meet the first mining requirement, screen out the sequence information of the other vehicle that meets the mining conditions.

[0114] In one embodiment of the present application, the screening module 310 is further configured to

[0115] According to the sequence information of the host vehicle, determine whether the state of the host vehicle meets the second mining requirement, where the second mining requirement includes at least one of the following: the existence time of the host vehicle exceeds the first time, the host vehicle exists for more than the second time within a future time period, and the host vehicle is in a non-stationary state;

[0116] If it is determined that the states of the host vehicle all meet the second mining requirement, then screen out the sequence information of the host vehicle that meets the mining conditions.

[0117] In one embodiment of the present application, it further includes a negative sample module, which is used to

[0118] Screen out adjacent obstacles in the horizontal direction of the host vehicle and vehicles in the same lane in the vertical direction of the host vehicle in the current scenario;

[0119] Record the adjacent obstacles in the horizontal direction of the host vehicle and the vehicles in the same lane in the vertical direction of the host vehicle as negative sample data of other vehicle behaviors.

[0120] It can be understood that the above-mentioned other vehicle behavior mining device can implement each step of the other vehicle behavior mining method provided in the foregoing embodiment. The relevant explanations regarding the other vehicle behavior mining method are applicable to the other vehicle behavior mining device and will not be elaborated here.

[0121] Figure 4 It is a schematic structural diagram of an electronic device according to an embodiment of the present application. Please refer to Figure 4 , at the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. Among them, the memory may include a memory, such as a high-speed random access memory (Random-Access Memory, RAM), and may also include a non-volatile memory, such as at least one disk memory, etc. Of course, the electronic device may also include other hardware required for other services.

[0122] The processor, network interface, and memory can be interconnected through an internal bus, and the internal bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, Figure 4It is represented by only one bidirectional arrow, but it does not mean that there is only one bus or one type of bus.

[0123] A memory for storing programs. Specifically, the program may include program code, and the program code includes computer operation instructions. The memory may include a memory and a non-volatile memory, and provide instructions and data to the processor.

[0124] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a behavior mining device of other vehicles at the logical level. The processor executes the program stored in the memory and is specifically used to perform the following operations:

[0125] Filter out the sequence information of the host vehicle and other vehicles that meet the mining conditions, and the sequence information includes: vehicle position, YAW angle, vehicle size, and timestamp information;

[0126] According to the sequence information of the host vehicle and other vehicles that meet the mining conditions, judge the lateral displacement change amount between the host vehicle and other vehicles;

[0127] If the lateral displacement change amounts between the host vehicle and other vehicles all meet their respective corresponding conditions, record them as sample data of the behavior of other vehicles.

[0128] The above is as in the present application Figure 2The method executed by the other vehicle behavior mining device disclosed in the illustrated embodiment can be applied to or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor or the instructions in the form of software. The above-mentioned processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as being executed by the hardware decoding processor, or executed by the combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory, and the processor reads the information in the memory and combines its hardware to complete the steps of the above method.

[0129] The electronic device can also execute Figure 2 the method executed by the other vehicle behavior mining device in Figure 2 the illustrated embodiment and implement the functions of the other vehicle behavior mining device in

[0130] Embodiments of the present application also propose a computer-readable storage medium that stores one or more programs. The one or more programs include instructions that, when executed by an electronic device including a plurality of application programs, can enable the electronic device to execute Figure 2 the method executed by the other vehicle behavior mining device in the illustrated embodiment, and specifically used to execute:

[0131] Filter out the sequence information of the host vehicle and other vehicles that meet the mining conditions. The sequence information includes: vehicle position, YAW angle, vehicle size, and timestamp information;

[0132] According to the sequence information of the host vehicle and other vehicles that meet the mining conditions, judge the lateral displacement change amount between the host vehicle and other vehicles;

[0133] If the lateral displacement change amounts between the host vehicle and other vehicles all meet their respective corresponding conditions, they are recorded as sample data of the behaviors of other vehicles.

[0134] Those skilled in the art should understand that the embodiments of the present invention may be provided as a method, a system, or a computer program product. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) that contain computer-usable program code.

[0135] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be realized by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for realizing the functions specified in one or more flows of the flowchart and / or one or more blocks of the block diagram.

[0136] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including an instruction device that realizes the functions specified in one or more flows of the flowchart and / or one or more blocks of the block diagram.

[0137] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for realizing the functions specified in one or more flows of the flowchart and / or one or more blocks of the block diagram.

[0138] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and a memory.

[0139] The memory may include non-permanent memory in the computer-readable medium, in the form of random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0140] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer 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 technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0141] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0142] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0143] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for mining other vehicle behaviors, wherein: The method comprises: Filter out the sequence information of the own vehicle and other vehicles that meet the mining conditions, wherein the sequence information includes: vehicle position, YAW angle, vehicle size and timestamp information; According to the sequence information of the own vehicle and other vehicles that meet the excavation conditions, determining the lateral displacement changes of the own vehicle and other vehicles; If the lateral displacement changes of the self-vehicle and the other vehicle both meet their corresponding conditions, they are recorded as sample data of the other vehicle's behavior.

2. The method of claim 1, wherein: The step of determining the lateral displacement changes of the own vehicle and the other vehicle according to the sequence information of the own vehicle and the other vehicle that meet the excavation condition comprises: According to the sequence information of the own vehicle and the other vehicle that meets the mining condition, it is determined in sequence whether any one or more of the lateral displacement changes of the own vehicle and the other vehicle, including the lateral distance, lateral offset, and minimum lateral distance, are associated with the preset driving behavior of the other vehicle; If it is determined that the lateral displacement changes are all associated with the preset driving behaviors of other vehicles, then the sequence information of the own vehicle and other vehicles that meet the mining conditions continues to be used; If it is determined that one of the lateral displacement changes is not associated with the preset driving behavior of the other vehicle, the sequence information of the own vehicle and the other vehicle that meets the mining condition is discarded.

3. The method of claim 2, wherein: If the lateral displacement changes of the self-vehicle and the other vehicle both meet their corresponding conditions, they are recorded as sample data of the other vehicle's behavior, including: Determine the lateral distance between the vehicle and the other vehicle; If it is determined that the lateral distance between the vehicle and the other vehicle meets the preset threshold, the lateral offset between the vehicle and the other vehicle in the future time period is calculated; If the lateral offset is consistent with the behavior of the other vehicle approaching the own vehicle, determining the minimum lateral distance between the own vehicle and the other vehicle in the future time period; If it is determined that the minimum lateral distance between the vehicle and the other vehicle in the future time period meets the preset threshold, it is recorded as positive sample data of the other vehicle's behavior.

4. The method of claim 3, wherein: The preset driving behavior of the other vehicle includes the cutin behavior of the other vehicle, and the record is sample data of the behavior of the other vehicle, including: The time when the positive sample data of the other car's behavior is recorded is regarded as the start time of the other car's cutin; Determine the end time of the cutin of the other vehicle based on the start time and the experience value; If the minimum lateral distance time is less than the empirical value, the empirical value is used as the shortest time, and the minimum lateral distance time is determined by the minimum lateral distance; If the minimum lateral distance time is greater than the empirical value, the time when the minimum lateral distance occurs is used as the end time of the cutin of the other vehicle.

5. The method of claim 1, wherein: The sequence information of the self-vehicle and other-vehicles that meet the mining conditions is screened out, including: According to the sequence information of the other vehicle, whether the state of the other vehicle meets the first mining requirement is judged, and the first mining requirement includes at least one of the following: the existence time of the other vehicle meets the requirement, the other vehicle is in a non-stationary state, the lateral and longitudinal distances between the other vehicle and the own vehicle meet all possible conditions for mining the behavior of the other vehicle, the other vehicle and the own vehicle are in the same direction, and the other vehicle and the own vehicle are in different lanes; If it is determined that the states of the other vehicles all meet the first mining requirement, the sequence information of the other vehicles that meet the mining condition is screened out.

6. The method of claim 5, wherein: The step of screening out the sequence information of the vehicle and other vehicles that meet the mining conditions also includes: According to the sequence information of the ego vehicle, determining whether the state of the ego vehicle meets the second mining requirement, wherein the second mining requirement includes at least one of the following: the existence time of the ego vehicle exceeds the first time, the existence of the ego vehicle in the future time period exceeds the second time, and the ego vehicle is in a non-stationary state; If it is determined that the states of the own vehicles all meet the second mining requirement, the sequence information of the own vehicles that meet the mining condition is screened out.

7. The method according to any one of claims 1 to 6, wherein: The method further comprises: Filter out the adjacent obstacles in the horizontal direction of the vehicle and the vehicles in the same lane in the vertical direction of the vehicle in the current scene; The adjacent obstacles in the horizontal direction of the vehicle and the vehicles in the same lane in the vertical direction of the vehicle are recorded as negative sample data of the other vehicle's behavior.

8. A vehicle behavior excavation device, wherein: The device comprises: A screening module is used to screen out the sequence information of the self-vehicle and other vehicles that meet the mining conditions, and the sequence information includes: vehicle position, YAW angle, vehicle size and timestamp information; A judgment module, used for judging the lateral displacement change of the own vehicle and the other vehicle according to the sequence information of the own vehicle and the other vehicle that meet the excavation condition; The recording module is used to record the sample data of the behavior of the other vehicle if the lateral displacement changes of the own vehicle and the other vehicle both meet their corresponding conditions.

9. An electronic device, comprising: processor; as well as A memory arranged to store computer executable instructions, which when executed cause the processor to perform the method of any one of claims 1 to 7.

10. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, causes the electronic device to execute any one of the methods of claims 1 to 7.

Images