A vehicle trajectory prediction method for emergency scene in an autonomous driving system

Abstract

The application provides a vehicle trajectory prediction method for emergency scene in an automatic driving system, comprising: acquiring and managing perception information provided by an automatic driving system on the automatic driving vehicle; extracting road structure information from a high-precision map database, finding and matching road structure information at a current time for each vehicle in the perception range of the automatic driving vehicle, and managing road structure information at the current time and historical time of each vehicle; the automatic driving vehicle judging whether each vehicle in the perception range is in an emergency scene according to road structure information at the current time and historical time, and current perception information and historical perception information; and obtaining a predicted trajectory of each vehicle according to the judgment result by using different motion models. The application has the advantages of strong operability, strong expansibility and high practical value by means of several scenes to evaluate whether the vehicle is in an emergency scene.

Description

TECHNICAL FIELD

[0001] The present application mainly relates to the field of automatic driving system, and particularly relates to a vehicle trajectory prediction method for emergency scene in an automatic driving system. BACKGROUND

[0002] Vehicle trajectory prediction is an important and basic problem in the field of autonomous vehicles. For example, when starting automatic driving on public roads, accurately predicting where other vehicles and pedestrians are expected to be can help autonomous vehicles take safe and effective actions.

[0003] According to the differences in the input of the trajectory prediction model and the intermediate processing steps, the prediction models of the previous studies can be roughly divided into three categories.

[0004] The first category is the prediction model based on physical constraints. The vehicle cannot change the trajectory instantaneously due to inertia, and the vehicle is represented as a dynamic entity subject to Newton's physical laws. The evolution of the control state (such as wheel angle, acceleration, etc.), the characteristics of the car (such as the self-weight of the car body), and the road environment factors (such as the static friction coefficient) are linked to the vehicle state (such as position, heading, speed), and the dynamics model and kinematics model are used to predict the future motion trend of the vehicle.

[0005] The second category is the prediction model based on behavior intention. The prediction process is divided into two parts: identifying intention and predicting trajectory. First, identify the intention of the vehicle (follow the straight line of the front vehicle, turn left at the intersection, turn right, etc.), and then predict the future continuous physical state of the vehicle so as to predict the trajectory corresponding to the possible behavior intention.

[0006] The third category is the prediction model based on learning. The vehicle driving trajectory data obtained by target detection and tracking and high-precision map information are used to train a deep neural network model, and the trained model is used to infer and output the predicted future driving trajectory of the vehicle.

[0007] The deficiencies of the three categories of prediction models are analyzed as follows:

[0008] The first category of physical model only predicts through the motion model of the vehicle, without considering traffic rules and interactions with other vehicles, especially when passing through an intersection, the prediction trajectory deviation value is large.

[0009] The second type of intention model uses prior knowledge such as vehicle state, road structure information such as speed, acceleration, historical position, and traffic speed limit, but ignores the use of posterior knowledge including traffic vehicle interaction, driver driving habits and the like. In actual traffic scenarios, the posterior knowledge in the driving scene (driver driving habits, etc.) will affect the driving trajectory of the vehicle under the same driving intention. If the posterior knowledge in the driving scene is ignored, the accuracy of the obtained motion trajectory cannot be guaranteed.

[0010] The third type of learning-based prediction model, that is, the end-to-end design based on a deep neural network, uses training data including not only trajectory information but also more abundant environmental information including static map road network environment static information and dynamic target (such as vehicles and pedestrians) perception information. The deep neural network model is learned. With the increase of the amount of training data and the increase of the type of data scene, different driving scenes can be adapted, better generality and accuracy can be obtained. However, there are also some limitations, such as strong limitation on the reliability of input perception data, long calculation time leading to poor real-time performance, the need to accumulate several periods of historical trajectory points to perform prediction, generally 9 periods of historical trajectories are needed, and the ability to deal with sudden situations is poor.

[0011] Whether it is a human-driven car or an autonomous car, it is inevitable to encounter unexpected events (such as a vehicle suddenly rushing into its own lane due to an accident) during driving. It is undoubtedly necessary for an autonomous car to have the ability to safely drive in the face of unexpected events. SUMMARY

[0012] It should be understood that the above general description and the following detailed description of the present disclosure are exemplary and illustrative, and are intended to provide further explanation of the present disclosure.

[0013] To deal with the situation that the above-mentioned autonomous driving system encounters unexpected events, it is a necessary ability requirement for an autonomous car. The present application proposes a vehicle trajectory prediction method for unexpected event scenarios, which is a prediction model based on physical constraints.

[0014] The present application discloses a vehicle trajectory prediction method for unexpected event scenarios in an autonomous driving system, applied to an autonomous vehicle, characterized in that the method comprises:

[0015] Step 1: Obtain and manage the perception information provided by the autonomous driving system on the autonomous vehicle;

[0016] Step 2: Extract road structure information from a high-precision map database, find and match the road structure information at the current time for each car within the perception range of the autonomous vehicle, and manage the road structure information at the current time and the historical time of each car.

[0017] Step three, the autonomous vehicle determines whether each car in its perception range is in an emergency scene according to the road structure information at the current time and the road structure information at the historical time, and the perception information at the current time and the perception information at the historical time;

[0018] Step four, different motion models are used to obtain the predicted trajectory of each car according to the judgment result.

[0019] Preferably, the application further discloses a vehicle trajectory prediction method for emergency scenes in an autonomous driving system, characterized in that the judgment of whether it is in an emergency scene in step three further comprises:

[0020] Step three one, determine whether the autonomous vehicle is on a straight road or a turning road at the current time?

[0021] Step three two, if it is on a straight road, continue to determine whether the acceleration of the autonomous vehicle at the current time is zero?

[0022] Step three three, if it is on a turning road, continue to determine whether the acceleration of the autonomous vehicle at the current time is zero?

[0023] Step three four, if the autonomous vehicle is currently on a straight road and the acceleration at the current time is not zero, a CV motion model is used to generate a vehicle predicted trajectory;

[0024] Step three five, if the autonomous vehicle is currently on a straight road and the acceleration at the current time is zero, a CA motion model is used to generate a vehicle predicted trajectory;

[0025] Step three six, if the autonomous vehicle is currently on a turning road and the acceleration at the current time is not zero, a CTRV motion model is used to generate a vehicle predicted trajectory;

[0026] Step three seven, if the autonomous vehicle is currently on a turning road and the acceleration at the current time is zero, a CTRA motion model is used to generate a vehicle predicted trajectory.

[0027] Preferably, the application further discloses a vehicle trajectory prediction method for emergency scenes in an autonomous driving system, characterized in that the emergency scene comprises:

[0028] Scene one, at a roundabout, the autonomous vehicle encounters a vehicle driving in the opposite direction;

[0029] Scene two, the autonomous vehicle merges out of the highway / elevated road and is about to enter the exit ramp, the autonomous vehicle encounters a vehicle driving in the opposite direction, and the time when the two vehicles may collide is less than 3 seconds;

[0030] Scenario three, the autonomous vehicle enters the highway / overpass from the ramp, has entered the ramp, and encounters a vehicle driving in the opposite direction, and the time when the two vehicles may collide is less than 3 seconds;

[0031] Scenario four, at the intersection, the autonomous vehicle encounters a vehicle driving in the opposite direction;

[0032] Scenario five, on a straight road, the autonomous vehicle encounters a vehicle driving in the opposite direction occupying the lane;

[0033] Scenario six, on a curved road, the autonomous vehicle encounters a vehicle driving in the opposite direction occupying the lane.

[0034] Preferably, the application further discloses a vehicle trajectory prediction method for sudden event scenarios in an autonomous driving system, characterized in that the CV and CA motion models are applicable to scenarios one, two and five.

[0035] Preferably, the application further discloses a vehicle trajectory prediction method for sudden event scenarios in an autonomous driving system, characterized in that the CTRV motion model is applicable to scenarios one, two and three.

[0036] Preferably, the application further discloses a vehicle trajectory prediction method for sudden event scenarios in an autonomous driving system, characterized in that the CTRV motion model is applicable to scenarios one, two and three.

[0037] Preferably, the application further discloses a vehicle trajectory prediction method for sudden event scenarios in an autonomous driving system, characterized in that the step one acquires perception information including the position, length, width, color, speed, acceleration, yaw angle and yaw angle change rate of the autonomous vehicle.

[0038] The step one manages the acquired perception information, including adding, deleting and modifying the perception information of new vehicles appearing in the perception range, and searching for the perception information of a certain vehicle in the perception range according to the search conditions.

[0039] Preferably, the application further discloses a vehicle trajectory prediction method for sudden event scenarios in an autonomous driving system, characterized in that if it is judged that the vehicle is not in a sudden event scenario in the step three, a vehicle prediction algorithm for non-sudden event scenarios is called to calculate the predicted trajectory.

[0040] The present invention further discloses a vehicle trajectory prediction method for emergency scenarios in an autonomous driving system, characterized in that after step four completes the prediction trajectory for each car within the perception range of the autonomous driving vehicle, the system returns to step one for a new round of prediction.

[0041] This invention employs different motion models to calculate the predicted trajectory of a vehicle in a sudden event scenario. By matching different sudden event scenarios, the method can calculate the predicted trajectory of the vehicle. While maintaining high performance, this method has the advantages of being simple and efficient, consuming few computing resources, having short calculation time, and strong real-time performance. Attached Figure Description

[0042] Embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. Preferred embodiments of the present disclosure will now be described in detail, examples of which are illustrated in the drawings. Wherever possible, the same reference numerals will be used in all the drawings to denote the same or similar parts. Furthermore, although the terminology used in this disclosure is selected from commonly known and used terminology, some terms referenced in this disclosure may have been chosen by the applicant at his or her judgment, and their detailed meanings are explained in the relevant sections of the description herein. Moreover, this disclosure should be understood not only by the actual terms used, but also by the meaning implied by each term.

[0043] The above and other objects, features and advantages of the present invention will become apparent to those skilled in the art from the detailed description thereof, with reference to the accompanying drawings.

[0044] Figure 1 This is a schematic diagram of a sudden event scenario one provided by the present invention;

[0045] Figure 2 This is a schematic diagram of scenario two of the sudden events provided by the present invention;

[0046] Figure 3 This is a schematic diagram of scenario three of the emergency events provided by the present invention;

[0047] Figure 4 This is a schematic diagram of scenario four of the emergency events provided by the present invention;

[0048] Figure 5 This is a schematic diagram of scenario five of the emergency events provided by the present invention;

[0049] Figure 6 This is a schematic diagram of scenario six of the emergency events provided by the present invention;

[0050] Figure 7 This diagram illustrates the overall process of how an autonomous vehicle generates a predicted trajectory during its operation.

[0051] Figure 8 is the flowchart of several burst scenarios of the present application. Figures 1-6 DETAILED DESCRIPTION

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some examples or embodiments of the present application, and for those skilled in the art, the present application can also be applied to other similar scenarios without creative labor on the basis of these drawings. Unless it is clear from the language context or otherwise indicated, the same reference numbers in the drawings represent the same structures or operations.

[0053] As shown in the present application and claims, unless the context clearly indicates otherwise, the words "one", "an", "a", and / or "the" do not specify a singular form, but can also include a plural form. Generally speaking, the terms "comprise" and "include" only indicate the inclusion of the steps and elements explicitly identified, and these steps and elements do not constitute an exclusive list, and the method or device can also include other steps or elements.

[0054] Unless otherwise specifically indicated, the relative arrangement of the components and steps, numerical expressions, and numerical values set forth in these embodiments do not limit the scope of the present application. At the same time, it should be understood that the sizes of the various parts shown in the drawings are not drawn in proportion to the actual proportions. The technology, methods and devices known to those skilled in the relevant art can not be discussed in detail, but under appropriate circumstances, the technology, methods and devices should be considered as part of the authorized description. In all examples shown and discussed here, any specific value should be interpreted as merely exemplary, and not as a limitation. Therefore, other examples of exemplary embodiments can have different values. It should be noted that similar reference numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0055] In the description of the present application, it should be understood that the orientation words such as "front, back, up, down, left, right", "horizontal, vertical, perpendicular, horizontal" and "top, bottom" and the like indicate the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and in the absence of contrary indications, these orientation words do not indicate and imply that the indicated device or element must have a particular orientation or be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the scope of protection of the present application; the orientation words "inner, outer" refer to the inner and outer relative to the contour of the parts themselves.

[0056] ​For purposes of the description hereinafter, spatially relative terms, such as "above", "below", "up", "down", "between", "within", "left", "right", "rear", "front", "upper", "lower", "horizontal", "vertical", "above", "below", "up", "down", "top", "bottom", "under" and the like, can be used herein for ease of description to describe one element or feature's relationship to another element(s) or feature(s) as illustrated in the figures. It will be understood that the spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientations depicted in the figures. For example, if a device in the figures is inverted, elements described as "above" or "up" other elements or features would then be oriented "below" or "down" the other elements or features. Thus, the exemplary term "above" can encompass both an orientation of above and below. The device can be otherwise oriented (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein interpreted accordingly. The terms "first", "second", "third", etc., as used herein do not denote any order, quantity, or importance, but are used to distinguish one element from another.

[0057] In addition, it should be noted that the use of "first", "second", "third", etc. words to qualify parts, is only for the convenience of distinguishing the corresponding parts, and the above words have no special meaning unless otherwise stated, and therefore cannot be understood as limiting the scope of protection of the present application. In addition, although the terms used in the present application are selected from well-known and commonly used terms, some terms mentioned in the specification of the present application can be selected by the applicant according to his or her judgment, and the detailed meaning of each term is described in the relevant part of the description. In addition, the present application is required to be understood not only by the actual terms used, but also by the meaning implied by each term.

[0058] Flowcharts have been used herein to illustrate the operations performed by systems in accordance with embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in sequence. On the contrary, various steps can be processed in reverse order or simultaneously. While, or some steps or operations are added to or removed from these processes.

[0059] Regarding the emergency scenario related to the present application, it is particularly related to the following contents, and the automatic driving vehicle is marked A in the figure:

[0060] Scenario one: at the roundabout, the automatic driving vehicle encounters a vehicle driving in the opposite direction.

[0061] As Figure 1 shown, the scene parameters are: the number of target vehicles: 1-4, the speed of the target vehicle: 0-120km / h; the diameter of the roundabout: 4-50m; the number of lanes: 1-2; the position of the roundabout exit: 1-4, the time when the two vehicles may collide is less than 3 seconds.

[0062] Scenario two: the automatic driving vehicle merges out of the highway / viaduct, is about to enter the ramp, and encounters a vehicle driving in the opposite direction, and the time when the two vehicles may collide is less than 3 seconds.

[0063] Please refer to Fig. 2, which shows the scenario parameters: the number of target vehicles: 1-4, the speed of the target vehicles: 0-120 km / h, and the number of lanes: 1-2. Figure 2

[0064] Scenario three: the automatic driving vehicle enters the highway / viaduct from the ramp, has entered the ramp, and encounters a vehicle driving in the opposite direction, and the time when the two vehicles may collide is less than 3 seconds.

[0065] Please refer to Fig. 3, which shows the scenario parameters: the number of target vehicles: 1-4, the speed of the target vehicles: 0-120 km / h, and the number of lanes: 1-2. Figure 3

[0066] Scenario four: at an intersection, the automatic driving vehicle encounters a vehicle driving in the opposite direction.

[0067] Please refer to Fig. 4, which shows the scenario parameters: the number of target vehicles: 1-4, the speed of the target vehicles: 0-120 km / h, the number of lanes: 1-6, the position of the intersection exit: 1-4, and the time when the two vehicles may collide is less than 3 seconds. Figure 4

[0068] Scenario five: on a straight road, the automatic driving vehicle encounters a vehicle driving in the opposite direction and occupying the lane.

[0069] Please refer to Fig. 5, which shows the scenario parameters: the number of target vehicles: 1-4, the speed of the target vehicles: 0-120 km / h, the lane overlap rate of the two vehicles: 25%-100%, the number of lanes: 1-6, and the time when the two vehicles may collide is less than 3 seconds. Figure 5

[0070] Scenario six: on a curved road, the automatic driving vehicle encounters a vehicle driving in the opposite direction and occupying the lane.

[0071] Please refer to Fig. 6, which shows the scenario parameters: the number of target vehicles: 1-4, the speed of the target vehicles: 0-120 km / h, the lane overlap rate of the two vehicles: 25%-100%, the number of lanes: 1-6, and the time when the two vehicles may collide is less than 3 seconds. Figure 6

[0072] The aforementioned target vehicle speed refers to the speed range of other vehicles within the perception range of the automatic driving vehicle.

[0073] The present application regards the above six scenarios as emergency scenarios, and all vehicle trajectory prediction methods related to emergency scenarios are based on the six scenarios.

[0074] ​​​​​Figure 7 The overall flow chart of how the vehicle prediction trajectory is generated during the driving process of the autonomous vehicle is shown, and the steps are described as follows in combination with the flow chart:

[0075] Step 71, obtaining the perception information provided by the autonomous driving system;

[0076] In the preferred embodiment, the obtained perception information includes the position, length, width, color, speed, acceleration, yaw angle, and yaw angle change rate of the vehicle.

[0077] Step 72, managing the obtained perception information;

[0078] In the preferred embodiment, the management of the obtained perception information includes adding, deleting, modifying, and searching the perception information. Specifically:

[0079] a) Adding: adding the perception information of a newly appearing vehicle within the perception range;

[0080] b) Deleting: deleting the perception information of a vehicle outside the perception range;

[0081] c) Modifying: updating the perception information of a vehicle already present within the perception range;

[0082] d) Searching: searching for the perception information of a vehicle within the perception range according to a certain search condition;

[0083] Step 73, extracting the road structure information from the high-precision map database, searching and matching the road structure information at the current time for each vehicle, and then managing the road structure information of each vehicle at the current time and in the past history.

[0084] It should be noted that in this step, "each vehicle" means each vehicle within the perception range of the autonomous vehicle. The meaning of "each vehicle" appearing in the text of the present application is consistent with this.

[0085] The road structure information includes the current lane of each vehicle, the front and rear connecting lanes of the current lane, and the left and right adjacent lanes of the current lane.

[0086] Similarly, the meaning of "management" here includes the ability to add, delete, modify, and search the perception information.

[0087] Step 74, determining whether each vehicle is in a sudden event scenario based on the road structure information at the current time and the historical time, and the perception information at the current time and the historical time.

[0088] It needs to be specially pointed out that in this process, the vehicle of the automatic driving function is to evaluate whether all vehicles within a certain distance (for example, 100 meters) around it are in a sudden scene, and after the evaluation is completed, the automatic driving vehicle goes to predict the future trajectory of all vehicles in the range.

[0089] The predicted all vehicles in the range include various types of vehicles, which can be unmanned or manned vehicles.

[0090] Specifically, the judgment of step 74 is carried out one by one, that is, the first calculation period evaluates the first vehicle, and then predicts the future trajectory of the first vehicle, the second calculation period evaluates the second vehicle, and then predicts the future trajectory of the second vehicle, and so on until all vehicles within the set range are evaluated and predicted. After this round of evaluation and prediction is completed, it continues to start from the beginning, evaluates and predicts the first vehicle, and then evaluates and predicts the second vehicle. Until the automatic driving vehicle is turned off and stopped, the evaluation and prediction work will stop.

[0091] Step 75, if the judgment result of step 74 is yes, that is, it indicates that a certain vehicle is in a sudden event scene, then the vehicle prediction algorithm for sudden event scene is called to calculate and predict the trajectory.

[0092] Step 76, if the judgment result of step 74 is no, that is, it indicates that a certain vehicle is not in a sudden event scene, then the vehicle prediction algorithm for non-sudden event scene is called to calculate and predict the trajectory.

[0093] Step 77, receiving the prediction trajectories generated in steps 75 and 76, for each vehicle, generating a vehicle prediction trajectory in the future several seconds.

[0094] In the overall process of the vehicle trajectory prediction method described above, steps 74 and 75 are related to sudden time scenes, and the application of three motion models in the related technical solutions thereof is introduced as follows:

[0095] (1) CV and CA motion model.

[0096] Among them, the constant velocity motion model (Constant Velocity motion model, abbreviated as CV)

[0097] Constant acceleration motion model (Constant Acceleration motion model, abbreviated as CA)

[0098] The above two models are physical models that assume that the speed of the automatic driving vehicle remains unchanged or the acceleration remains unchanged, which are suitable for the scene of vehicles driving on straight roads, for example, scenes one, two, and five.

[0099] (2)CTRV motion model.

[0100] Constant Turn Rate and Velocity motion model (CTRV)

[0101] The yaw angle and yaw rate of change are introduced into the vehicle state vector to reflect the change of the vehicle around the Z axis, and it is assumed that the yaw rate of change and the speed remain unchanged, wherein the yaw angle is referred to as the yaw angle, i.e., the orientation of the vehicle head.

[0102] This motion model is suitable for the scenario in which the autonomous vehicle drives on a curve, for example, scenarios one, two, and three.

[0103] (3)CTRA motion model.

[0104] Constant Turn Rate and Acceleration motion model (CTRA)

[0105] The yaw angle and yaw rate of change are introduced into the vehicle state vector to reflect the change of the vehicle around the Z axis, and it is assumed that the yaw rate of change and the acceleration remain unchanged.

[0106] This motion model is also suitable for the scenario in which the autonomous vehicle drives on a curve, for example, scenarios four and six.

[0107] Please refer to Figure 8 , which is an expanded flowchart of steps 74-75 in Figure 7 , and the following is explained in conjunction with the chart:

[0108] Step 81, extract several parameters of the autonomous vehicle at the current time, such as position, speed, acceleration, yaw, yaw rate of change, etc., and extract the road structure information around the autonomous vehicle at the current time from the high-precision map database;

[0109] Step 82, according to step 81, determine whether the autonomous vehicle at the current time is on a straight road or a turning road?

[0110] Step 83, if currently on a straight road, continue to determine whether the acceleration of the autonomous vehicle at the current time is zero?

[0111] Step 84, if currently on a turning road, continue to determine whether the acceleration of the autonomous vehicle at the current time is zero?

[0112] Step 85, if the vehicle is currently driving straight and the acceleration of the autonomous vehicle at the current time is not zero, then the CV motion model is used;

[0113] Step 86, if the current vehicle is straight and the acceleration of the autonomous vehicle at the current time is zero, then the CA motion model is adopted;

[0114] Step 87, if the current vehicle is on a turning road and the acceleration of the autonomous vehicle at the current time is not zero, then the CTRV motion model is adopted;

[0115] Step 88, if the current vehicle is on a turning road and the acceleration of the autonomous vehicle at the current time is zero, then the CTRA motion model is adopted.

[0116] Step 89, generate the vehicle prediction trajectory for each vehicle for several seconds in the future.

[0117] In step 89, according to the different motion models selected, the vehicle prediction trajectory is generated, and the process is described as follows:

[0118] When steps 85 and 86 select the CV and CA motion models, under this motion model, the vehicle speed remains unchanged or the acceleration remains unchanged.

[0119] At the current time k, assume the state vector X k of the autonomous vehicle is:

[0120]

[0121] wherein,

[0122] x(k) represents the vehicle position coordinate x at time k;

[0123] y(k) represents the vehicle position coordinate y at time k;

[0124] vx(k) represents the vehicle x-direction speed value at time k;

[0125] vy(k) represents the vehicle y-direction speed value at time k;

[0126] ax(k) represents the vehicle x-direction acceleration value at time k;

[0127] ay(k) represents the vehicle y-direction acceleration value at time k;

[0128] t represents the update calculation period value, which is usually 1 second, and the unit is second.

[0129] The state vector X k+1 of the vehicle at time k+1 is:

[0130]

[0131] The vehicle predicted trajectory is calculated according to the following steps:

[0132] Step one, the state vector X of the vehicle at time k+1 is calculated by using formula (2) k+1 .

[0133] Step two, the state vector X of the vehicle at time k+1 is calculated k+1 Then, the state vector X of the vehicle at time k+2 is further calculated by using formula (2) k+2 .

[0134] Step three, the predicted trajectory points at future time are obtained by such a cycle, and the predicted trajectory points at future time constitute the vehicle predicted trajectory, so that the vehicle predicted trajectory can be calculated.

[0135] When step 87 is selected as the CTRV motion model, the model introduces yaw angle and yaw angle change rate in the vehicle state vector to reflect the change of the vehicle around the Z axis, and it is assumed that the yaw angle change rate and the speed remain unchanged.

[0136] At the current time k, it is assumed that the state vector X of the autonomous vehicle is k According to the formula:

[0137]

[0138] Among them,

[0139] x represents the vehicle position coordinate x at time k;

[0140] y represents the vehicle position coordinate y at time k;

[0141] v k , represents the vehicle speed value at time k;

[0142] yaw k , represents the vehicle yaw value at time k;

[0143] yawd k , represents the vehicle yaw change rate value at time k;

[0144] acc k , represents the vehicle acceleration value at time k;

[0145] t, represents the update calculation period value, which is usually 1 second, and the unit is second.

[0146] The state vector X of the vehicle at time k+1 is k+1 respectively:

[0147]

[0148] and

[0149]

[0150] The vehicle predicted trajectory is calculated according to the following steps:

[0151] Step 1, the state vector X of the vehicle at time k+1 is calculated using formula (4) or formula (5) k+1 , wherein:

[0152] If |yawd k |≤0.01, the state vector X of the vehicle at time k+1 is calculated using formula (4) k+1 ;

[0153] If |yawd k |>0.01, the state vector X of the vehicle at time k+1 is calculated using formula (5) k+1 .

[0154] Step 2, the state vector X of the vehicle at time k+1 is calculated k+1 , then the state vector X of the vehicle at time k+2 is calculated using formula (4) or formula (5) k+2 .

[0155] Step 3, the predicted trajectory points at future time points are obtained by repeating the above steps, and the predicted trajectory points at future time points constitute the vehicle predicted trajectory, so that the vehicle predicted trajectory can be calculated.

[0156] When step 88 is selected as the CTRA motion model, the yaw angle and the yaw angle change rate are introduced into the vehicle state vector to reflect the change of the vehicle around the Z axis, and it is assumed that the yaw angle change rate and the acceleration remain unchanged.

[0157] At the current time k, it is assumed that the state vector X of the autonomous vehicle is k , according to the formula:

[0158]

[0159] The state vector X of the vehicle at time k+1 is k+1 :

[0160]

[0161] , wherein

[0162] x represents the position coordinate x of the vehicle at time k;

[0163] This represents the vehicle's position coordinate y at time k;

[0164] θ k , represents the vehicle's yaw value at time k;

[0165] v k , represents the vehicle speed at time k;

[0166] w k , represents the rate of change of vehicle yaw at time k;

[0167] a k , represents the vehicle acceleration value at time k;

[0168] t represents the update calculation period value, which is usually 1 second, and the unit is seconds.

[0169] The predicted vehicle trajectory is calculated using the following steps:

[0170] Step 1: Using formula (7), the state vector X of the vehicle at time k+1 can be calculated. k+1 ;

[0171] Step 2: Calculate the vehicle's state vector X at time k+1. k+1 Then, the state vector X of the vehicle at time k+2 can be calculated using formula (7). k+2 .

[0172] Step 3: By repeating this process, we can obtain the predicted trajectory points for several future moments. These predicted trajectory points constitute the vehicle's predicted trajectory, which can then be calculated.

[0173] In summary, this invention is the first in the industry to propose a method for assessing whether a car is in a sudden event scenario by utilizing several specific scenarios. By generalizing scenario parameters, the method becomes more comprehensive, offering advantages such as strong operability, strong scalability, and high practical value. Furthermore, this invention employs different motion models to calculate the predicted trajectory of a vehicle in a sudden event scenario. By matching different sudden event scenarios to calculate the predicted vehicle trajectory, this method maintains high performance while offering advantages such as simplicity, efficiency, low computational resource consumption, short computation time, and strong real-time performance.

[0174] The basic concepts have been described above. Obviously, for those skilled in the art, the above disclosure is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore remain within the spirit and scope of the exemplary embodiments of this application.

[0175] Furthermore, this application uses specific terms to describe embodiments of the application. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic related to at least one embodiment of the application. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different locations in this specification do not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of the application can be appropriately combined.

[0176] Some aspects of this application can be executed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The aforementioned hardware or software may be referred to as a "data block," "module," "engine," "unit," "component," or "system." The processor may be one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DAPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, or combinations thereof. Furthermore, aspects of this application may manifest as computer products residing in one or more computer-readable media, including computer-readable program code. For example, computer-readable media may include, but are not limited to, magnetic storage devices (e.g., hard disks, floppy disks, magnetic tapes, etc.), optical discs (e.g., compressed CDs, digital multifunction DVDs, etc.), smart cards, and flash memory devices (e.g., cards, sticks, key drives, etc.).

[0177] A computer-readable medium may contain a propagated data signal containing computer program code, for example, on baseband or as part of a carrier wave. This propagated signal may take various forms, including electromagnetic, optical, and so on, or suitable combinations thereof. A computer-readable medium can be any computer-readable medium other than a computer-readable storage medium, which can be connected to an instruction execution system, apparatus, or device to enable communication, propagation, or transmission of a program for use. The program code located on the computer-readable medium can be propagated through any suitable medium, including radio, cable, fiber optic cable, radio frequency signals, or similar media, or any combination of the above media.

[0178] The basic concepts have been described above. Obviously, for those skilled in the art, the above disclosure is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore remain within the spirit and scope of the exemplary embodiments of this application.

[0179] Also, the use of "a" or "an" to describe an element of the application is merely taken to mean "one or more" in some embodiments, unless otherwise indicated. The singular forms "a" or "an" are used herein to mean one or one or more than one, unless indicated otherwise. The term "another" is used herein to mean one or more than one, unless indicated otherwise. The term "about" is used herein to describe approximately the value that precedes such term. The term "about" is used herein to describe an approximation that is within 20% of the value that precedes such term. Accordingly, the numerical parameters used in this specification and claims are approximations that can vary depending upon the desired properties sought to be obtained by the particular embodiment. Unless otherwise indicated, the numerical values set forth in the detailed description and claims are approximations that can vary depending upon the desired properties sought to be obtained by the particular embodiment. Notwithstanding that the numerical ranges and parameters setting forth the broadest scope of the application are approximations, the numerical values set forth in the specific examples are reported as precisely as possible. Any numerical value, however, can contain certain errors necessarily resulting from the standard deviation found in their respective testing measurements.

[0180] Also, it is noted that the embodiments can sometimes describe components as related to one another. For the purposes of this disclosure, the relationship can be referred to in terms of being "coupled" and / or "directly coupled" and / or "resistantly coupled". These terms are not intended as synonyms for each other. Rather, these terms might be used to describe different types of relationships. For example, a component might be directly coupled to another component if the components are in direct physical contact with one another. On the other hand, a component might be resistantly coupled to another component if the components are not in direct contact with one another, but yet are electrically or otherwise communicatively connected. It will be appreciated that the existence of these relationships between the components is represented by the fact that the components are included in the same figure or description.

[0181] Some embodiments use numerical ranges to describe quantities of ingredients, attributes, etc. It should be understood that such numerical ranges described in the embodiments are, in some examples, modified by the word "about". Unless otherwise indicated, the word "about" means ±20% of the value being described. Accordingly, any numerical value inherently contains certain errors necessarily resulting from the standard deviation found in their respective testing measurements. In some embodiments, the numerical values used in the specification and claims have been reported as precisely as possible. The numerical values, however, are approximations because some testing measurements have a standard deviation necessarily resulting from the respective testing measurement technique.

[0182] Although the application has been described with reference to the current embodiments, it will be understood that the application is to cover any changes and modifications to the embodiments that are within the spirit and scope of the application.

Claims

1. A vehicle trajectory prediction method for emergency scenarios in an autonomous driving system, applied to autonomous vehicles, characterized in that, The method includes: Step 1: Acquire and manage the perception information provided by the autonomous driving system on the autonomous vehicle; the perception information includes: the position, length, width, color, speed, acceleration, yaw angle and yaw angle change rate of the autonomous vehicle; managing the acquired perception information includes: adding, deleting and modifying and updating the perception information of newly appearing vehicles within the perception range, and searching for the perception information of a vehicle within the perception range according to search conditions; Step 2: Extract road structure information from the high-precision map database, find and match the current road structure information for each car within the perception range of the autonomous vehicle, and manage the current and historical road structure information for each car. Step three: The autonomous vehicle, based on the current road structure information and historical road structure information, as well as the current perception information and historical perception information, determines whether each vehicle within its perception range is in a sudden event scenario; including: At the current time k, assume the state vector X of the autonomous vehicle is... k According to the formula: in, This represents the vehicle's position coordinate x at time k; This represents the vehicle's position coordinate y at time k; v k , representing the vehicle speed at time k; yaw k , representing the vehicle's yaw value at time k; yawd k , representing the rate of change of the vehicle's yaw at time k; acc k , representing the vehicle acceleration value at time k; t represents the update calculation cycle value, which is usually 1 second, and the unit is seconds; The state vector X of the vehicle at time k+1 k+1 They are respectively: as well as The predicted vehicle trajectory is calculated using the following steps: The first step is to calculate the vehicle's state vector X at time k+1 using formula (2) or formula (3). k+1 ,in: If |yawd k If |≤0.01, then the state vector X of the vehicle at time k+1 can be calculated using formula (2). k+1 ; If |yawd k If |>0.01, then the state vector X of the vehicle at time k+1 can be calculated using formula (3). k+1 ; The second step is to calculate the vehicle's state vector X at time k+1. k+1 Then, the vehicle's state vector X at time k+2 can be calculated using formula (2) or formula (3). k+2 ; The third step is to repeat this process to obtain the predicted trajectory points at several future times. These predicted trajectory points at several future times constitute the vehicle's predicted trajectory, which can then be calculated. Step 4: Based on the judgment results, different motion models are used to obtain the predicted trajectory of each car.

2. The vehicle trajectory prediction method for emergency scenarios in an autonomous driving system according to claim 1, characterized in that, Step three, determining whether a sudden event scenario is in progress, further includes: Step 31: Determine whether the autonomous vehicle is currently on a straight road or a turning road; Step 32: If the vehicle is on a straight road, continue to determine whether the acceleration of the autonomous vehicle is zero at the current moment; Step 33: If the vehicle is on a turning road, continue to determine whether the acceleration of the autonomous vehicle is zero at the current moment; Steps 3 and 4: If the autonomous vehicle is currently traveling straight and its acceleration is not zero at the current moment, then the CV motion model is used to generate the vehicle's predicted trajectory. Step 35: The autonomous vehicle is currently traveling straight and its acceleration is zero at the current moment. Then, the CA motion model is used to generate the vehicle's predicted trajectory. Step 36: If the autonomous vehicle is currently turning and its acceleration is not zero at the current moment, then the CTRV motion model is used to generate the vehicle's predicted trajectory. Step 37: The autonomous vehicle is currently turning and its acceleration is zero at the current moment. The CTRA motion model is then used to generate the vehicle's predicted trajectory.

3. The vehicle trajectory prediction method for emergency scenarios in an autonomous driving system according to claim 2, characterized in that, The emergency scenarios include: Scenario 1: At a roundabout, the autonomous vehicle encounters a vehicle traveling in the opposite direction. Scenario 2: The autonomous vehicle is changing lanes to exit the highway / elevated road and is about to enter the exit ramp. The autonomous vehicle encounters a vehicle traveling in the opposite direction. The time between the two vehicles colliding is less than 3 seconds. Scenario 3: The autonomous vehicle enters the highway / elevated road from the ramp and has already entered the ramp. The autonomous vehicle encounters a vehicle traveling in the opposite direction, and the time between the two vehicles' potential collision is less than 3 seconds. Scenario 4: At an intersection, the autonomous vehicle encounters a vehicle traveling in the opposite direction. Scenario 5: On a straight road, the autonomous vehicle encounters an oncoming vehicle that occupies its lane and is traveling in the wrong direction. Scenario 6: On a curve, the autonomous vehicle encounters an oncoming vehicle that is driving in the wrong direction and occupying its lane.

4. The vehicle trajectory prediction method for emergency scenarios in an autonomous driving system according to claim 3, characterized in that, The CV and CA motion models are applicable to Scene 1, Scene 2, and Scene 5.

5. The vehicle trajectory prediction method for emergency scenarios in an autonomous driving system according to claim 3, characterized in that, The CTRV motion model is applicable to Scene 1, Scene 2, and Scene 3.

6. The vehicle trajectory prediction method for emergency scenarios in an autonomous driving system according to claim 3, characterized in that, The CTRA motion model is applicable to scenarios four and six.

7. The vehicle trajectory prediction method for emergency scenarios in an autonomous driving system according to claim 1, characterized in that, In step three, if it is determined that the vehicle is not in a sudden event scenario, the vehicle prediction algorithm for non-sudden event scenarios is invoked to calculate the predicted trajectory.

8. The vehicle trajectory prediction method for emergency scenarios in an autonomous driving system according to claim 1, characterized in that, After step four completes the prediction trajectory for each car within the perception range of the autonomous vehicle, it returns to step one for a new round of prediction.

Images