Predictive trajectory generation method, apparatus, and vehicle

Abstract

Embodiments of the present application disclose a kind of prediction trajectory generation method, device and vehicle.The method comprises: obtaining the state information and reference travel trajectory of target vehicle current;Based on the starting position, the starting speed, the starting acceleration, the starting heading angle and the reference travel trajectory, first prediction trajectory is obtained;The first prediction trajectory is extended, and second prediction trajectory is obtained;Based on the first prediction trajectory and the second prediction trajectory, the prediction trajectory of the target vehicle is generated.The above-mentioned mode makes it possible to obtain the first prediction trajectory composed of multiple circular arc curves based on the state information and reference travel trajectory of target vehicle current, and the first prediction trajectory is extended, and the second prediction trajectory parallel to the tangent direction of reference travel trajectory is obtained, so that accurate prediction trajectory can be obtained, and driving safety is improved.

Description

Technical Field

[0001] The present application relates to the field of vehicle technology, and more specifically, to a method and device for generating a predicted trajectory, and a vehicle. Background Art

[0002] As research into intelligent driving deepens, in order to navigate complex traffic scenarios more safely and efficiently, intelligent vehicles can predict the trajectory information of surrounding vehicles and plan routes to avoid collisions with other vehicles. In related methods, the vehicle's motion state and road characteristics can be combined, and specific algebraic or geometric methods can be applied to generate a continuous, smooth trajectory curve that can be described by a mathematical formula, from which a sequence of trajectory points can be sampled. The regular curve method can be further divided into the polynomial method and the circular arc method. The industry generally uses the polynomial method to generate the obstacle vehicle's future driving trajectory. However, in these related methods, the predicted driving trajectory is still not accurate enough. Summary of the Invention

[0003] In view of the above problems, the present application proposes a method, device and vehicle for generating a predicted trajectory to improve the above problems.

[0004] In a first aspect, the present application provides a method for generating a predicted trajectory, the method comprising: obtaining current state information and a reference driving trajectory of a target vehicle, the state information comprising a starting position, a starting speed, a starting acceleration, and a starting heading angle of the target vehicle, the reference driving trajectory being a trajectory generated based on a target lane corresponding to the target vehicle; obtaining a first predicted trajectory based on the starting position, the starting speed, the starting acceleration, the starting heading angle, and the reference driving trajectory, the first predicted trajectory being composed of a plurality of circular arc curves, wherein the plurality of circular arc curves are obtained by trajectory fitting based on their respective corresponding preview points, and the preview points are obtained based on the reference driving trajectory; extending the first predicted trajectory to obtain a second predicted trajectory, the tangent direction of the second predicted trajectory being parallel to the tangent direction of the reference driving trajectory and the distance between the second predicted trajectory and the reference driving trajectory being the same as the distance between the end point of the first predicted trajectory and the reference driving trajectory; generating a predicted trajectory of the target vehicle based on the first predicted trajectory and the second predicted trajectory.

[0005] In a second aspect, the present application provides a predicted trajectory generation device, the device comprising: an information acquisition unit, configured to acquire current state information and a reference driving trajectory of a target vehicle, the state information comprising a starting position, a starting speed, a starting acceleration, and a starting heading angle of the target vehicle, the reference driving trajectory being a trajectory generated based on a target lane corresponding to the target vehicle; a first predicted trajectory generation unit, configured to obtain a first predicted trajectory based on the starting position, the starting speed, the starting acceleration, the starting heading angle, and the reference driving trajectory, the first predicted trajectory being composed of a plurality of circular arc curves, wherein the plurality of circular arc curves are obtained by trajectory fitting based on their respective corresponding preview points, and the preview points are obtained based on the reference driving trajectory; a second predicted trajectory generation unit, configured to extend the first predicted trajectory to obtain a second predicted trajectory, the tangent direction of the second predicted trajectory being parallel to the tangent direction of the reference driving trajectory and the distance between the second predicted trajectory and the reference driving trajectory being the same as the distance between the end point of the first predicted trajectory and the reference driving trajectory; a predicted trajectory generation unit, configured to generate a predicted trajectory of the target vehicle based on the first predicted trajectory and the second predicted trajectory.

[0006] In a third aspect, the present application provides a vehicle comprising one or more processors and a memory; one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to execute the above-mentioned method.

[0007] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores program code, wherein the above method is executed when the program code is run.

[0008] The present application provides a predicted trajectory generation method, device, vehicle, and storage medium. After obtaining the current status information of a target vehicle including the starting position, starting speed, starting acceleration, and starting heading angle of the target vehicle and a reference driving trajectory generated based on the target lane corresponding to the target vehicle, a first predicted trajectory consisting of multiple circular arc curves is obtained based on the starting position, the starting speed, the starting acceleration, the starting heading angle, and the reference driving trajectory. The multiple circular arc curves are obtained by trajectory fitting based on their respective corresponding preview points, and the preview points are obtained based on the reference driving trajectory. The first predicted trajectory is extended to obtain a second predicted trajectory whose tangent direction is parallel to the tangent direction of the reference driving trajectory and the distance between the second predicted trajectory and the reference driving trajectory is the same as the distance between the end point of the first predicted trajectory and the reference driving trajectory. The predicted trajectory of the target vehicle is generated based on the first predicted trajectory and the second predicted trajectory. Through the above method, a first predicted trajectory consisting of multiple circular arc curves can be obtained based on the current state information of the target vehicle and the reference driving trajectory. The multiple circular arc curves are obtained by trajectory fitting based on their corresponding preview points determined based on the reference driving trajectory, and the curvature of the circular arc curves themselves can reflect the information state of the target vehicle's lateral and longitudinal motion, thereby obtaining an accurate first predicted trajectory. Furthermore, the first predicted trajectory can be extended to obtain a second predicted trajectory parallel to the tangent direction of the reference driving trajectory, thereby obtaining an accurate predicted trajectory and improving driving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A flowchart of a method for generating a predicted trajectory proposed in an embodiment of the present application is shown;

[0011] Figure 2 A schematic diagram of a vehicle interior module proposed in an embodiment of the present application is shown;

[0012] Figure 3 Shows this application Figure 1 A flowchart of an implementation method proposed in S120;

[0013] Figure 4 A schematic diagram showing a projection point position proposed in an embodiment of the present application is shown;

[0014] Figure 5 Shows this application Figure 3 A flowchart of an implementation method proposed in S122;

[0015] Figure 6 Shows this application Figure 3 A flowchart of an implementation method proposed in S123;

[0016] Figure 7 Shows this application Figure 1 A flowchart of an implementation method proposed in S130;

[0017] Figure 8 A schematic diagram of a predicted trajectory proposed in this application is shown;

[0018] Figure 9 A flowchart of a method for generating a predicted trajectory proposed in another embodiment of the present application is shown;

[0019] Figure 10 A schematic diagram of multiple reference prediction trajectories proposed in this application is shown;

[0020] Figure 11 Shows this application Figure 9 A flowchart of an implementation method proposed in S230;

[0021] Figure 12 The following is a structural block diagram of a prediction trajectory generation device proposed in an embodiment of the present application;

[0022] Figure 13 Shows a structural block diagram of a vehicle proposed in this application;

[0023] Figure 14 It is a storage unit for storing or carrying program codes for implementing the predicted trajectory generation method according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0025] In order to solve the above technical problems, the inventors proposed a predicted trajectory generation method, device and vehicle in the present application. After obtaining the current status information of the target vehicle including the starting position, starting speed, starting acceleration, starting heading angle of the target vehicle and the reference driving trajectory generated based on the target lane corresponding to the target vehicle, a first predicted trajectory consisting of multiple circular arc curves is obtained based on the starting position, the starting speed, the starting acceleration, the starting heading angle and the reference driving trajectory, wherein the multiple circular arc curves are obtained by trajectory fitting based on their respective corresponding preview points, and the preview points are obtained based on the reference driving trajectory. The first predicted trajectory is extended to obtain a second predicted trajectory whose tangent direction is parallel to the tangent direction of the reference driving trajectory and the distance between the second predicted trajectory and the reference driving trajectory is the same as the distance between the end point of the first predicted trajectory and the reference driving trajectory. The predicted trajectory of the target vehicle is generated based on the first predicted trajectory and the second predicted trajectory. Through the above method, a first predicted trajectory consisting of multiple circular arc curves can be obtained based on the current state information of the target vehicle and the reference driving trajectory. The multiple circular arc curves are obtained by trajectory fitting based on their corresponding preview points determined based on the reference driving trajectory, and the curvature of the circular arc curves themselves can reflect the information state of the target vehicle's lateral and longitudinal motion, thereby obtaining an accurate first predicted trajectory. Furthermore, the first predicted trajectory can be extended to obtain a second predicted trajectory parallel to the tangent direction of the reference driving trajectory, thereby obtaining an accurate predicted trajectory and improving driving safety.

[0026] The embodiments of this application will be described below with reference to the accompanying drawings.

[0027] See also Figure 1 , the present application provides a method for generating a predicted trajectory, the method comprising:

[0028] S110: Obtain current status information and a reference driving trajectory of the target vehicle, wherein the status information includes a starting position, a starting speed, a starting acceleration, and a starting heading angle of the target vehicle, and the reference driving trajectory is a trajectory generated based on a target lane corresponding to the target vehicle.

[0029] The target vehicle may refer to the vehicles surrounding the vehicle implementing the prediction trajectory generation method provided by this application, and the surrounding area may refer to the range detectable by the vehicle implementing the prediction trajectory generation method provided by this application through its own electronic devices (such as: laser radar, camera, millimeter wave radar, etc.). The starting position may refer to the current position information of the target vehicle when the target vehicle is detected, the starting speed and starting acceleration may refer to the current speed and acceleration of the target vehicle when the target vehicle is detected, respectively, and the starting heading angle may refer to the current direction of travel of the target vehicle when the target vehicle is detected.

[0030] The target lane may refer to the lane that the target vehicle is estimated to enter by the vehicle implementing the prediction trajectory generation method provided in this application, and the reference driving trajectory may overlap with the center line of the target lane.

[0031] As a way, Figure 2 As shown, the vehicle can be provided with a perception module, and the perception module can correspond to a variety of information acquisition devices (such as: lidar, camera, millimeter wave radar, etc.), so as to obtain the status information of the target vehicle through the perception module.

[0032] As a way, Figure 2 As shown, the vehicle can be provided with an intention assessment module and a map module. The map module can provide map information of the target vehicle (such as lane information, road information, etc.). The intention assessment module can determine the target lane based on the status information of the target vehicle and the map information of the map module, and generate a reference driving trajectory based on the target lane.

[0033] Optionally, since the lane information of the map module is usually stored in the form of a discrete point sequence, the cubic spline interpolation method can be used to smoothly connect and resample all discrete reference trajectory point sequences to generate a reference driving trajectory with continuous second-order derivatives.

[0034] S120: Based on the starting position, the starting speed, the starting acceleration, the starting heading angle, and the reference driving trajectory, a first predicted trajectory is obtained, wherein the first predicted trajectory is composed of multiple circular arc curves, wherein the multiple circular arc curves are obtained by trajectory fitting based on their respective corresponding preview points, and the preview points are obtained based on the reference driving trajectory.

[0035] As a way, Figure 3 As shown, based on the state information and the reference driving trajectory, a first predicted trajectory is obtained, including:

[0036] S121: During the current prediction process, based on the starting position, the first coordinate system and the second coordinate system, the projection point position corresponding to the current prediction process is obtained, and the projection point position is the projection position of the target vehicle on the reference driving trajectory.

[0037] The first coordinate system may refer to a Cartesian coordinate system, and the second coordinate system may refer to a Frenet coordinate system.

[0038] As a way, Figure 4 As shown, in the current prediction process, the starting position can be determined by the first coordinate system, the reference driving trajectory can be determined by the second coordinate system, and the starting position can be P n (x n ,yn ), by P n (x n ,y n ) is projected vertically onto the reference driving trajectory, and we can get

[0039] '

[0040] The projection point position P corresponding to the current prediction process n (s n ,l n ), where l n =0.

[0041] S122: Based on the starting point position, the projection point position, the starting point speed, the starting point acceleration and the preset preview time, obtain the target curvature corresponding to the current prediction process.

[0042] As a way, Figure 5 As shown, based on the starting point position, the projection point position, the starting point speed, the starting point acceleration and the preset preview time, the target curvature corresponding to the current prediction process is obtained, including:

[0043] S1221: Based on the projection point position, the starting point velocity, the starting point acceleration and the preset preview time, obtain the preview point position of the current prediction process.

[0044] As a method, the reference preview point position of the current prediction process can be obtained based on the projection point position, the starting point speed and the preset preview time. The reference preview point position can be on the reference driving trajectory; the reference preview point position is converted from the second coordinate system to the first coordinate system to obtain the preview point position corresponding to the current prediction process.

[0045] The status information may further include the length of the target vehicle, that is, the distance from the front to the rear of the target vehicle.

[0046] Optionally, the preview distance of the current prediction process can be obtained based on the starting speed, starting acceleration, length and pre-set preview time; the reference preview point position of the current prediction process can be obtained based on the projection point position and the preview distance of the current prediction process.

[0047] Optionally, a reference preview distance of the current prediction process can be obtained based on the starting speed, starting acceleration, length and preset preview time; the reference preview distance of the current prediction process is compared with the preset preview distance; if the reference preview distance is greater than or equal to the preset preview distance, the reference preview distance of the current prediction process is used as the preview distance of the current prediction process; if the reference preview distance is less than the preset preview distance, the preset preview distance is used as the preview distance of the current prediction process.

[0048] Optionally, the calculation formula for the reference preview distance can be:

[0049]

[0050] Among them, v n It can represent the starting speed, a n It can be expressed as the starting acceleration, t aim It can represent the preview time, and L can represent the length of the target vehicle.

[0051] Optionally, to prevent the subsequent curvature from being too large due to the preview distance being too close, the preset preview distance can be a preset minimum preview distance s. min For example, the minimum preview distance can be 5 meters. aim and s min For comparison, if s aim Greater than or equal to s min , will s aim As the preview distance of the current prediction process; if s aim Less than s min , will s min As the preview distance of the current prediction process.

[0052] Optionally, after obtaining the preview distance, the reference preview point position of the current prediction process can be obtained based on the projection point position and the preview distance, and then the reference preview point position is converted from the second coordinate system to the first coordinate system to obtain the preview point position corresponding to the current prediction process.

[0053] For example, Figure 4 As shown, in the second coordinate system, the preview distance can be s aim , the projection point position can be Then the reference preview point position can be T′ n (s n_aim ,l n_aim ), where s n_aim =s n +s aim , l n_ain = 0. Then set the preview point T n ′(s n_aim ,l n_aim ) coordinate system transformation can get the preview point position T n (x n_aim ,y n_aim ).

[0054] Alternatively, the preview time may be obtained based on multiple trials.

[0055] S1222: Based on the starting point position and the preview point position, obtain the target curvature corresponding to the current prediction process.

[0056] As a method, the reference curvature corresponding to the current prediction process can be obtained based on the starting point position and the preview point position; the curvature limit of the target vehicle can be obtained based on the starting speed and length; if the reference curvature is greater than or equal to the curvature limit, the curvature limit is used as the target curvature corresponding to the current prediction process; if the reference curvature is less than the curvature limit, the reference curvature is used as the target curvature corresponding to the current prediction process.

[0057] Optionally, the reference curvature may be calculated as follows:

[0058]

[0059] Among them, (x n_aim ,y n_aim ) can represent the coordinates of the preview point position, (x n ,y n ) can represent the coordinates of the starting position.

[0060] Optionally, the maximum curvature κ allowed during vehicle driving can be max As the curvature limit. Curvature limit κ max The starting speed v n As a function of length κ max =f(v n ,L) indicates that the specific functional relationship can be obtained through the vehicle dynamics formula, or by establishing a database of the maximum curvatures of a large number of different vehicles traveling at different speeds.

[0061] S123: Based on the target curvature, the starting point speed, the starting point acceleration, the starting point heading angle and the preset step time, the prediction point position, prediction point speed, prediction point acceleration and prediction point heading angle corresponding to the current prediction process are obtained.

[0062] As a way, Figure 6 As shown, the prediction point position, prediction point velocity, prediction point acceleration, and prediction point heading angle corresponding to the current prediction process are obtained based on the target curvature, the starting point velocity, the starting point acceleration, the starting point heading angle, and the preset step time, including:

[0063] S1231: Based on the starting point acceleration, the starting point velocity and the step time, the arc length corresponding to the target curvature is obtained, and the arc length represents the distance between the starting point position corresponding to the current prediction process and the prediction point position in the arc corresponding to the target curvature.

[0064] As a way, the arc length can be calculated as:

[0065] Δs=v n *Δt+0.5*a n Δt 2

[0066] Among them, v n It can represent the starting speed, a n It can represent the starting point acceleration, and Δt can represent the step time.

[0067] Optionally, the step time may be obtained based on multiple experiments, for example, the step time may be Δt=0.2 s.

[0068] S1232: Based on the target curvature, the starting point position and the starting point heading angle, obtain the center position corresponding to the target curvature.

[0069] As a way, the calculation formula of the center position of the circle is as follows:

[0070] x 0n =x n -sinθ n / κ n

[0071] y 0n =y n +cosθ n / κ n

[0072] Among them, (x n ,y n ) can represent the coordinates of the starting position, κ n It can represent the target curvature, θ n It can represent the starting heading angle.

[0073] S1233: Based on the starting point heading angle, the target curvature and the arc length, obtain the prediction point heading angle corresponding to the current prediction process.

[0074] As a way, the heading angle of the predicted point is calculated as follows:

[0075] θ n+1 =θ n +κ n *Δs

[0076] Among them, θ n It can represent the starting heading angle, κ n It can represent the target curvature, and Δs can represent the arc length.

[0077] S1234: Obtain a predicted point speed corresponding to the current prediction process based on the start speed, the start acceleration, and the step time.

[0078] As one way, the calculation formula of the predicted point speed is as follows:

[0079] v n+1 = v n + a n * Δt

[0080] Wherein, v n may represent the start speed, a n may represent the start acceleration, and Δt may represent the step time.

[0081] S1235: Obtain a predicted point acceleration corresponding to the current prediction process based on the start speed and the start acceleration.

[0082] As one way, the calculation formula of the predicted point acceleration is as follows:

[0083] a n+1 = a n *(1 / (1+exp(1 / (|v n |+0.001)))

[0084] Wherein, v n may represent the start speed, and a n may represent the start acceleration.

[0085] S1236: Obtain a predicted point position corresponding to the current prediction process based on the predicted point heading angle, the center position, and the target curvature.

[0086] As one way, the calculation formula of the predicted point position is as follows:

[0087] x n+1 = x 0n + sinθ n+1 / κ n

[0088] y n+1 = y 0n -cosθ n+1 / κ n

[0089] Wherein, (x 0n , y 0n ) may represent the coordinates of the center position, θ n+1 may represent the predicted heading angle, and κ n may represent the target curvature.

[0090] S124: Obtaining an arc curve corresponding to the current prediction process based on the starting point position, the prediction point position corresponding to the current prediction process, and the target curvature.

[0091] As a method, a circular arc curve from the starting point position to the prediction point position can be determined based on the target curvature, and the circular arc curve can be used as the circular arc curve corresponding to the current prediction process.

[0092] S125: Obtain the first predicted trajectory based on the arc curve and all arc curves before the current prediction process.

[0093] As a method, if the predicted point position corresponding to the current prediction process meets the target conditions, the first predicted trajectory is obtained based on the circular arc curve corresponding to the current prediction process and all circular arc curves before the current prediction process; if the predicted point position corresponding to the current prediction process does not meet the target conditions, the predicted point position, predicted point speed, predicted point acceleration, and predicted point heading angle corresponding to the current prediction process are used as the starting position, starting speed, starting acceleration, and starting heading angle of the next prediction process, respectively, to perform the next prediction.

[0094] Optionally, the predicted point position corresponding to the current prediction process can be converted from the first coordinate system to the second coordinate system to obtain the predicted projection point position corresponding to the current prediction process, and the predicted projection point position can be on the reference driving trajectory; the first tangential angle corresponding to the current prediction process is obtained, and the first tangential angle can be the tangential angle of the reference driving trajectory at the predicted projection point position; based on the first tangential angle corresponding to the current prediction process and the predicted point heading angle, a first difference is obtained; if the first difference corresponding to the current prediction process is within a preset range, it is determined that the predicted point position corresponding to the current prediction process meets the target condition.

[0095] The first difference can represent the degree of parallelism between the heading angle of the prediction point and the reference driving trajectory. The smaller the first difference, the more parallel the heading angle of the prediction point is to the reference driving trajectory. The preset range can be obtained based on multiple experiments. For example, the preset range can be [-Δθ max , Δθ max ], where Δθ max It can be 1°.

[0096] For example, the predicted point position can be obtained by P n+1 (x n+1 ,y n+1 ) indicates that the predicted projection point position can be obtained by P′ n+1 (s n+1 ,l n+1 ), the first tangential angle can be expressed by θ′ n+1 Indicates that the heading angle of the predicted point can be expressed by θ n+1Indicates that the first difference can be Δθ n+1 =θ′ n+1 -θ n+1 , when |Δθ n+1 |≤Δθ max When , it can indicate that the first difference corresponding to the current prediction process is within a preset range, and then it is determined that the prediction point position corresponding to the current prediction process meets the target condition.

[0097] Optionally, if the first difference corresponding to the current prediction process is not within a preset range, obtain the second tangential angle corresponding to the current prediction process, and the second tangential angle can be the tangential angle of the reference driving trajectory at the projection point position; based on the second tangential angle corresponding to the current prediction process and the starting point heading angle, obtain the second difference; if the product of the first difference corresponding to the current prediction process and the second difference is less than or equal to 0, determine that the prediction point position corresponding to the current prediction process meets the target condition; if the product of the first difference corresponding to the current prediction process and the second difference is greater than 0, determine that the prediction point position corresponding to the current prediction process does not meet the target condition.

[0098] Among them, the product of the first difference and the second difference being less than or equal to 0 can indicate that there is an inflection point between the starting point position and the predicted point position, that is, a point parallel to the reference driving trajectory.

[0099] For example, the predicted projection point position can be obtained by P′ n (s n ,l n ), the second tangential angle can be expressed by θ′ n Indicates that the starting heading angle can be expressed by θ n Indicates that the second difference can be Δθ n =θ′ n -θ n , if Δθ n *Δθ n+1 ≤0, it can be shown that the heading of the prediction point is parallel to the reference driving trajectory, and it is determined that the position of the prediction point corresponding to the current prediction process meets the target condition; if Δθ n *Δθ n+1 >0, it is determined that the prediction point position corresponding to the current prediction process does not meet the target condition.

[0100] In the embodiments of the present application, stopping the generation of the first predicted trajectory when the predicted point meets the target condition allows the generated predicted trajectory to maintain a certain distance from the reference driving trajectory, rather than fitting the reference driving trajectory perfectly. This allows the generated predicted trajectory to better align with the actual driving conditions of the target vehicle, thereby improving the accuracy of the predicted trajectory. For example, when the target vehicle is actually driving, it is unlikely to always travel along the centerline of the target lane, but rather maintain a certain distance from the centerline of the target lane.

[0101] S130: Extend the first predicted trajectory to obtain a second predicted trajectory, where a tangent direction of the second predicted trajectory is parallel to a tangent direction of the reference driving trajectory and a distance between the second predicted trajectory and the reference driving trajectory is the same as a distance between an end point of the first predicted trajectory and the reference driving trajectory.

[0102] The first predicted trajectory may be composed of multiple predicted points, and each predicted point may correspond to a predicted point position and speed.

[0103] As a way, Figure 7 As shown, extending the first predicted trajectory to obtain a second predicted trajectory includes:

[0104] S131: During the current prediction process, converting the last prediction point in the first prediction trajectory from the first coordinate system to the second coordinate system based on the reference point position of the last prediction point.

[0105] As a way, Figure 8 As shown, the predicted point position of the last predicted point in the first predicted trajectory can be expressed as P N (x N ,y N ), based on the transformation of the last predicted point from the first coordinate system to the second coordinate system, the reference point position of the last predicted point can be expressed as P N (s N ,l N ).

[0106] S132: Based on the predicted point position of the last predicted point and the reference driving trajectory, a lateral offset corresponding to the current prediction process is obtained, where the lateral offset represents the lateral distance between the last predicted point and the reference driving trajectory.

[0107] As a way, since the last predicted point in the first predicted trajectory is parallel to the reference driving trajectory, the lateral offset can be the lateral distance between the last predicted point and the reference driving trajectory, that is, the vertical distance from the last predicted point to the reference driving trajectory.

[0108] Optionally, the calculation formula for the lateral offset can be:

[0109] L N+i =L N

[0110] Among them, the reference point position of the last predicted point can be expressed as P N (s N ,l N ), then the lateral offset can be L N .

[0111] S133: Based on the reference point position, speed and preset step time of the last prediction point, obtain the longitudinal offset corresponding to the current prediction process.

[0112] As a method, the longitudinal offset can be calculated as:

[0113] s N+i =s N +v N Δt

[0114] Among them, the reference point position of the last predicted point can be expressed as P N (s N ,l N ), v N It can represent the speed of the last prediction point, and Δt can represent the step time.

[0115] Among them, v N =v N-1 *Δt+0.5*a N-1 Δt 2 , v N-1 It can represent the speed of the previous prediction point of the last prediction point, a N-1 It can represent the acceleration of the previous prediction point of the last prediction point.

[0116] S134: Based on the lateral offset and longitudinal offset corresponding to the current prediction process, obtain the predicted reference point position corresponding to the current prediction process, and continue to obtain the predicted reference point position corresponding to the next prediction process based on the predicted reference point position corresponding to the current prediction process, until a predicted reference trajectory is generated.

[0117] As a method, after obtaining the horizontal offset and vertical offset corresponding to the current prediction process based on steps S132 and S133, the horizontal offset and vertical offset can be spliced ​​to obtain the prediction reference point position corresponding to the current prediction process. For example, the horizontal offset can be L N+i , the longitudinal offset can be s N+i , then the predicted reference point position can be P M+i (s M+i ,L M+i ).

[0118] S135: Convert the predicted reference trajectory from the second coordinate system to the first coordinate system to obtain the second predicted trajectory.

[0119] As a way, Figure 8 As shown, the predicted reference trajectory {P N+1 (s N+1 ,l N+1 ), PN+2 (s N+2 ,l N+2 ), ..., P N+i (s N+i ,l N+i )} is transformed from the second coordinate system to the first coordinate system to obtain the second predicted trajectory {P N+1 (X N+1 ,Y N+1 ), P N+2 (X N+2 ,Y N+2 ), ..., P N+i (X N+i ,Y N+i )}.

[0120] S140: Generate a predicted trajectory of the target vehicle based on the first predicted trajectory and the second predicted trajectory.

[0121] As a method, the first predicted trajectory and the second predicted trajectory can be spliced ​​in chronological order to obtain the predicted trajectory of the target vehicle.

[0122] For example, the first predicted trajectory can be {P0, P1, ..., P N} and the second predicted trajectory can be {P N+1 ,P N+2 ,…,P N+n}, then the predicted trajectory can be {P0,P1,…,P N , P N+1 ,P N+2 ,…,P N+n}.

[0123] This embodiment provides a method for generating a predicted trajectory. After obtaining current status information of a target vehicle including a starting position, starting speed, starting acceleration, and starting heading angle of the target vehicle and a reference driving trajectory generated based on a target lane corresponding to the target vehicle, a first predicted trajectory consisting of multiple circular arc curves is obtained based on the starting position, starting speed, starting acceleration, starting heading angle, and the reference driving trajectory. The multiple circular arc curves are obtained by trajectory fitting based on their respective corresponding preview points, which are obtained based on the reference driving trajectory. The first predicted trajectory is extended to obtain a second predicted trajectory whose tangent direction is parallel to the tangent direction of the reference driving trajectory and whose distance between the second predicted trajectory and the reference driving trajectory is the same as the distance between the end point of the first predicted trajectory and the reference driving trajectory. The predicted trajectory of the target vehicle is generated based on the first predicted trajectory and the second predicted trajectory. Through the above method, a first predicted trajectory consisting of multiple circular arc curves can be obtained based on the current state information of the target vehicle and the reference driving trajectory. The multiple circular arc curves are obtained by trajectory fitting based on their corresponding preview points determined based on the reference driving trajectory, and the curvature of the circular arc curves themselves can reflect the information state of the target vehicle's lateral and longitudinal motion, thereby obtaining an accurate first predicted trajectory. Furthermore, the first predicted trajectory can be extended to obtain a second predicted trajectory parallel to the tangent direction of the reference driving trajectory, thereby obtaining an accurate predicted trajectory and improving driving safety.

[0124] See also Figure 9 , the present application provides a method for generating a predicted trajectory, the method comprising:

[0125] S210: Obtain the current state information and reference driving trajectory of the target vehicle, where the state information includes the starting position, starting speed, starting acceleration, and starting heading angle of the target vehicle, and the reference driving trajectory is a trajectory generated based on the target lane corresponding to the target vehicle.

[0126] S220: If there are multiple preview times, obtain multiple reference prediction trajectories, wherein each reference prediction trajectory is generated based on the projection point position, the starting point velocity, the starting point acceleration, and the corresponding preview time.

[0127] Each reference prediction trajectory may be composed of multiple prediction points, and each of the multiple prediction points may correspond to a speed and a target curvature.

[0128] In the embodiment of the present application, in order to further improve the accuracy of the predicted trajectory, multiple preview times may be set, thereby generating multiple reference predicted trajectories based on the method of step S120.

[0129] For example, Figure 10 As shown, the multiple preview times can be 1.0s, 1.2s, 1.4s, and 1.6s, respectively, so that four reference prediction trajectories can be generated.

[0130] Optionally, the reference predicted trajectory may be a trajectory formed within a prediction period, which may be 8 seconds. Within the prediction period, a prediction point may be obtained every step time, and the step time may be 0.2 seconds.

[0131] S230: Obtaining the total cost corresponding to each of the plurality of reference predicted trajectories, where the total cost represents the cost of the target vehicle traveling according to the reference driving trajectory.

[0132] As a way, Figure 11 As shown, obtaining the total cost corresponding to each of the plurality of reference prediction trajectories includes:

[0133] S231: Based on the speeds and target curvatures of the multiple prediction points corresponding to each of the reference prediction trajectories, a centripetal acceleration cost corresponding to each of the multiple reference prediction trajectories is obtained, where the centripetal acceleration cost represents the magnitude of the centrifugal force generated during the driving process of the target vehicle.

[0134] As a method, the centripetal acceleration cost of each reference driving trajectory is calculated as follows:

[0135]

[0136] Among them, N can represent the total number of prediction points of a reference driving trajectory, i can represent the number of prediction points of a reference driving trajectory, and v i It can represent the speed of the i-th prediction point, κ i It can represent the target curvature of the i-th prediction point.

[0137] S232: Based on the number of prediction points corresponding to each reference prediction trajectory and the step time, obtain a lane change time cost corresponding to each of the multiple reference prediction trajectories, where the lane change time cost represents the time required for the target vehicle to change lanes.

[0138] As a method, the lane-changing time cost of each reference driving trajectory is calculated as follows:

[0139] C time =NΔt

[0140] Among them, N can represent the total number of prediction points of a reference driving trajectory, and Δt can represent the step time.

[0141] S233: Based on the target curvatures of the multiple prediction points corresponding to each of the reference prediction trajectories, obtain the average curvature change rate cost corresponding to each of the multiple reference prediction trajectories, where the average curvature change rate cost represents the smoothness of the target vehicle's driving.

[0142] As a method, the average curvature change cost of each reference driving trajectory is calculated as follows:

[0143]

[0144] Among them, N can represent the total number of prediction points of a reference driving trajectory, i can represent the number of prediction points of a reference driving trajectory, κ i+1 It can represent the target curvature of the i+1th prediction point, κ i It can represent the target curvature of the i-th prediction point.

[0145] S234: Based on the target lateral distance corresponding to each of the reference predicted trajectories, a lateral deviation cost corresponding to each of the multiple reference predicted trajectories is obtained, where the target lateral distance represents the lateral distance between the last predicted point of each of the reference predicted trajectories and the reference driving trajectory, and the lateral deviation cost is used to measure the size of the lateral distance.

[0146] As a method, the lateral deviation cost of each reference driving trajectory is calculated as follows:

[0147]

[0148] Among them, L N It can refer to the vertical distance from the last predicted point of each reference driving trajectory to the reference driving trajectory.

[0149] S235: Obtaining a total cost corresponding to each of the plurality of reference predicted trajectories based on the centripetal acceleration cost, lane change time cost, average curvature change rate cost, and lateral deviation cost corresponding to each of the plurality of reference predicted trajectories.

[0150] As a method, the total cost of each reference driving trajectory is calculated as follows:

[0151] C total =w1C acc_lat +w2C acc_lat +w3C Δκ +w4C lat

[0152] Among them, w1, w2, w3, and w4 are weight coefficients of corresponding costs respectively. The value of each weight coefficient can be obtained based on multiple experiments, w1+w2+w3+w4=1.

[0153] S240: Using the reference prediction trajectory with the lowest total cost among the multiple reference prediction trajectories as the first prediction trajectory.

[0154] As one approach, after obtaining the total costs corresponding to the multiple reference prediction trajectories, the reference prediction trajectory with the lowest total cost among the multiple reference prediction trajectories may be used as the first prediction trajectory.

[0155] S250: Extend the first predicted trajectory to obtain a second predicted trajectory, where a tangent direction of the second predicted trajectory is parallel to a tangent direction of the reference driving trajectory and a distance between the second predicted trajectory and the reference driving trajectory is the same as a distance between an end point of the first predicted trajectory and the reference driving trajectory.

[0156] S260: Generate a predicted trajectory of the target vehicle based on the first predicted trajectory and the second predicted trajectory.

[0157] This embodiment provides a method for generating a predicted trajectory. Through the above-mentioned method, a first predicted trajectory consisting of multiple circular arc curves can be obtained based on the current state information of the target vehicle and the reference driving trajectory. The multiple circular arc curves are obtained by trajectory fitting based on their respective corresponding preview points determined based on the reference driving trajectory, and the curvature of the circular arc curves themselves can reflect the information state of the lateral and longitudinal movement of the target vehicle, thereby obtaining an accurate first predicted trajectory. Furthermore, the first predicted trajectory can be extended to obtain a second predicted trajectory parallel to the tangent direction of the reference driving trajectory, thereby obtaining an accurate predicted trajectory and improving driving safety. Furthermore, in this embodiment, multiple reference predicted trajectories are generated by using multiple preview times, and the first predicted trajectory is obtained based on the total cost corresponding to each of the multiple reference predicted trajectories, thereby improving the accuracy of the predicted trajectory.

[0158] See also Figure 12 The present application provides a predicted trajectory generation device 400, the device 400 comprising:

[0159] The information acquisition unit 410 is used to obtain the current state information and reference driving trajectory of the target vehicle. The state information includes the starting position, starting speed, starting acceleration, and starting heading angle of the target vehicle. The reference driving trajectory is a trajectory generated based on the target lane corresponding to the target vehicle.

[0160] The first predicted trajectory generating unit 420 is configured to obtain a first predicted trajectory based on the starting position, the starting speed, the starting acceleration, the starting heading angle, and the reference driving trajectory. The first predicted trajectory is composed of multiple circular arc curves, wherein the multiple circular arc curves are obtained by trajectory fitting based on their respective corresponding preview points, and the preview points are obtained based on the reference driving trajectory.

[0161] The second predicted trajectory generating unit 430 is configured to extend the first predicted trajectory to obtain a second predicted trajectory, wherein a tangent direction of the second predicted trajectory is parallel to a tangent direction of the reference driving trajectory and a distance between the second predicted trajectory and the reference driving trajectory is the same as a distance between an end point of the first predicted trajectory and the reference driving trajectory.

[0162] The predicted trajectory generating unit 440 is configured to generate a predicted trajectory of the target vehicle based on the first predicted trajectory and the second predicted trajectory.

[0163] As a method, the first predicted trajectory generation unit 420 is specifically used to obtain the projection point position corresponding to the current prediction process based on the starting position, the first coordinate system and the second coordinate system, and the projection point position is the projection position of the target vehicle on the reference driving trajectory; based on the starting position, the projection point position, the starting speed, the starting acceleration and the preset preview time, obtain the target curvature corresponding to the current prediction process; based on the target curvature, the starting speed, the starting acceleration, the starting heading angle and the preset step time, obtain the predicted point position, predicted point speed, predicted point acceleration and predicted point heading angle corresponding to the current prediction process; based on the starting position, the predicted point position corresponding to the current prediction process and the target curvature, obtain the circular arc curve corresponding to the current prediction process; based on the circular arc curve and all circular arc curves before the current prediction process, obtain the first predicted trajectory.

[0164] Optionally, the first predicted trajectory generation unit 420 is specifically used to obtain the preview point position of the current prediction process based on the projection point position, the starting point speed, the starting point acceleration and the pre-set preview time; and obtain the target curvature corresponding to the current prediction process based on the starting point position and the preview point position.

[0165] Optionally, the first predicted trajectory generation unit 420 is specifically used to obtain a reference preview point position of the current prediction process based on the projection point position, the starting point speed, the starting point acceleration and the pre-set preview time, and the reference preview point position is on the reference driving trajectory; convert the reference preview point position from the second coordinate system to the first coordinate system to obtain the preview point position corresponding to the current prediction process.

[0166] Optionally, the state information also includes the length of the target vehicle. The first predicted trajectory generation unit 420 is specifically used to obtain the preview distance of the current prediction process based on the starting speed, the starting acceleration, the length and the preset preview time; and obtain the reference preview point position of the current prediction process based on the projection point position and the preview distance of the current prediction process.

[0167] Optionally, the first predicted trajectory generation unit 420 is specifically used to obtain a reference preview distance of the current prediction process based on the starting point speed, the starting point acceleration, the length and the preset preview time; compare the reference preview distance of the current prediction process with the preset preview distance, if the reference preview distance is greater than or equal to the preset preview distance, use the reference preview distance of the current prediction process as the preview distance of the current prediction process; if the reference preview distance is less than the preset preview distance, use the preset preview distance as the preview distance of the current prediction process.

[0168] Optionally, the first predicted trajectory generation unit 420 is specifically used to obtain a reference curvature corresponding to the current prediction process based on the starting point position and the preview point position; obtain a curvature limit of the target vehicle based on the starting speed and the length; if the reference curvature is greater than or equal to the curvature limit, use the curvature limit as the target curvature corresponding to the current prediction process; if the reference curvature is less than the curvature limit, use the reference curvature as the target curvature corresponding to the current prediction process.

[0169] Optionally, the first predicted trajectory generation unit 420 is specifically used to obtain the arc length corresponding to the target curvature based on the starting point acceleration, the starting point speed and the step time, where the arc length represents the distance between the starting point position and the predicted point position corresponding to the current prediction process in the arc corresponding to the target curvature; obtain the center position of the circle corresponding to the target curvature based on the target curvature, the starting point position and the starting point heading angle; obtain the predicted point heading angle corresponding to the current prediction process based on the starting point heading angle, the target curvature and the arc length; obtain the predicted point speed corresponding to the current prediction process based on the starting point speed, the starting point acceleration and the step time; obtain the predicted point acceleration corresponding to the current prediction process based on the starting point speed and the starting point acceleration; obtain the predicted point position corresponding to the current prediction process based on the predicted point heading angle, the center position and the target curvature.

[0170] Optionally, the first predicted trajectory generating unit 420 is specifically used to obtain the first predicted trajectory based on the circular arc curve corresponding to the current prediction process and all circular arc curves before the current prediction process if the predicted point position corresponding to the current prediction process meets the target condition; if the predicted point position corresponding to the current prediction process does not meet the target condition, the predicted point position, predicted point speed, predicted point acceleration, and predicted point heading angle corresponding to the current prediction process are used as the starting position, starting speed, starting acceleration, and starting heading angle of the next prediction process, respectively, to perform the next prediction.

[0171] Optionally, the first predicted trajectory generation unit 420 is specifically used to convert the predicted point position corresponding to the current prediction process from the first coordinate system to the second coordinate system to obtain the predicted projection point position corresponding to the current prediction process, and the predicted projection point position is on the reference driving trajectory; obtain the first tangential angle corresponding to the current prediction process, and the first tangential angle is the tangential angle of the reference driving trajectory at the predicted projection point position; obtain a first difference based on the first tangential angle corresponding to the current prediction process and the heading angle of the predicted point; if the first difference corresponding to the current prediction process is within a preset range, determine that the predicted point position corresponding to the current prediction process meets the target condition.

[0172] Optionally, the first predicted trajectory generation unit 420 is specifically used to obtain a second tangential angle corresponding to the current prediction process if the first difference corresponding to the current prediction process is not within the preset range, the second tangential angle being the tangential angle of the reference driving trajectory at the projection point position; obtain a second difference based on the second tangential angle corresponding to the current prediction process and the starting heading angle; if the product of the first difference and the second difference corresponding to the current prediction process is less than or equal to 0, determine that the predicted point position corresponding to the current prediction process meets the target condition; if the product of the first difference and the second difference corresponding to the current prediction process is greater than 0, determine that the predicted point position corresponding to the current prediction process does not meet the target condition.

[0173] Optionally, the first predicted trajectory generation unit 420 is specifically used to obtain multiple reference predicted trajectories if there are multiple preview times, wherein each reference predicted trajectory is generated based on the projection point position, the starting point speed, the starting point acceleration and the corresponding preview time; obtain the total cost corresponding to each of the multiple reference predicted trajectories, the total cost representing the cost of the target vehicle traveling along the reference driving trajectory; and use the reference predicted trajectory with the lowest total cost among the multiple reference predicted trajectories as the first predicted trajectory.

[0174] Optionally, each of the reference predicted trajectories is composed of multiple predicted points, and the first predicted trajectory generating unit 420 is specifically used to obtain the centripetal acceleration cost corresponding to each of the multiple reference predicted trajectories based on the speed and target curvature of the multiple predicted points corresponding to each of the reference predicted trajectories, and the centripetal acceleration cost represents the magnitude of the centrifugal force generated during the driving process of the target vehicle; based on the number of prediction points and the step time corresponding to each of the reference predicted trajectories, obtain the lane changing time cost corresponding to each of the multiple reference predicted trajectories, and the lane changing time cost represents the time required for the target vehicle to change lanes; based on the target curvature of the multiple predicted points corresponding to each of the reference predicted trajectories, obtain The average curvature change rate cost corresponding to each of the multiple reference predicted trajectories is calculated, and the average curvature change rate cost represents the smoothness of the target vehicle's driving; based on the target lateral distance corresponding to each of the reference predicted trajectories, the lateral deviation cost corresponding to each of the multiple reference predicted trajectories is obtained, and the target lateral distance represents the lateral distance between the last prediction point of each of the reference predicted trajectories and the reference driving trajectory, and the lateral deviation cost is used to measure the size of the lateral distance; based on the centripetal acceleration cost, lane change time cost, average curvature change rate cost, and lateral deviation cost corresponding to each of the multiple reference predicted trajectories, the total cost corresponding to each of the multiple reference predicted trajectories is obtained.

[0175] As a method, the first predicted trajectory is composed of multiple predicted points, each of which corresponds to a predicted point position and speed. The second predicted trajectory generating unit 430 is specifically used to convert the last predicted point in the first predicted trajectory from the first coordinate system to the second coordinate system during the current prediction process, based on the reference point position of the last predicted point; based on the predicted point position of the last predicted point and the reference driving trajectory, obtain the lateral offset corresponding to the current prediction process, and the lateral offset represents the lateral distance between the last predicted point and the reference driving trajectory; based on the reference point position, speed and pre-set step time of the last predicted point, obtain the longitudinal offset corresponding to the current prediction process; based on the lateral offset and longitudinal offset corresponding to the current prediction process, obtain the predicted reference point position corresponding to the current prediction process, and continue to obtain the predicted reference point position corresponding to the next prediction process based on the predicted reference point position corresponding to the current prediction process until a predicted reference trajectory is generated; convert the predicted reference trajectory from the second coordinate system to the first coordinate system to obtain the second predicted trajectory.

[0176] The following will be combined Figure 13 A vehicle provided in this application is described.

[0177] See also Figure 13Based on the above-mentioned predicted trajectory generation method and apparatus, embodiments of the present application also provide another vehicle 100 that can execute the above-mentioned predicted trajectory generation method. Vehicle 100 includes one or more (only one is shown in the figure) processors 102, a memory 104, and an information acquisition module 106 that are coupled to each other. The memory 104 stores a program that can execute the content of the above-mentioned embodiments, and the processor 102 can execute the program stored in the memory 104.

[0178] The processor 102 may include one or more processing cores. The processor 102 utilizes various interfaces and circuits to connect various components within the vehicle 100. It executes instructions, programs, code sets, or instruction sets stored in the memory 104, as well as accesses data stored in the memory 104, to perform various functions and process data within the vehicle 100. Optionally, the processor 102 may be implemented using at least one of the following hardware forms: a digital signal processing (DSP), a field-programmable gate array (FPGA), or a programmable logic array (PLA). The processor 102 may integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing display content; and the modem handles wireless communications. It is understood that the modem may not be integrated into the processor 102 and may instead be implemented via a separate communications chip.

[0179] The memory 104 may include a random access memory (RAM) or a read-only memory (ROM). The memory 104 may be used to store instructions, programs, codes, code sets, or instruction sets. The memory 104 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the following various method embodiments, etc. The data storage area may also store data created by the terminal 100 during use (such as a phone book, audio and video data, chat history data), etc.

[0180] The information acquisition module 106 is used to acquire the status information of the target vehicle. The information acquisition module 106 may include an image information acquisition device (such as a camera, etc.), various sensors, radar, etc.

[0181] Please refer to Figure 14 , which shows a block diagram of a computer-readable storage medium provided in an embodiment of the present application. The computer-readable storage medium 800 stores program code, which can be called by a processor to execute the method described in the above method embodiment.

[0182] The computer-readable storage medium 800 can be an electronic memory such as a flash memory, an EEPROM (Electrically Erasable Programmable Read-Only Memory), an EPROM, a hard disk, or a ROM. Alternatively, the computer-readable storage medium 800 includes a non-transitory computer-readable storage medium. The computer-readable storage medium 800 has storage space for program code 810 for executing any of the method steps described above. These program codes can be read from or written to one or more computer program products. The program code 810 can be compressed, for example, in a suitable form.

[0183] In summary, the present application provides a method, device, and vehicle for generating a predicted trajectory. After obtaining the current status information of the target vehicle including the starting position, starting speed, starting acceleration, and starting heading angle of the target vehicle and the reference driving trajectory generated based on the target lane corresponding to the target vehicle, a first predicted trajectory consisting of multiple circular arc curves is obtained based on the starting position, the starting speed, the starting acceleration, the starting heading angle, and the reference driving trajectory. The multiple circular arc curves are obtained by trajectory fitting based on their respective corresponding preview points, and the preview points are obtained based on the reference driving trajectory. The first predicted trajectory is extended to obtain a second predicted trajectory whose tangent direction is parallel to the tangent direction of the reference driving trajectory and the distance between the second predicted trajectory and the reference driving trajectory is the same as the distance between the end point of the first predicted trajectory and the reference driving trajectory. The predicted trajectory of the target vehicle is generated based on the first predicted trajectory and the second predicted trajectory. Through the above method, a first predicted trajectory consisting of multiple circular arc curves can be obtained based on the current state information of the target vehicle and the reference driving trajectory. The multiple circular arc curves are obtained by trajectory fitting based on their corresponding preview points determined based on the reference driving trajectory, and the curvature of the circular arc curves themselves can reflect the information state of the target vehicle's lateral and longitudinal motion, thereby obtaining an accurate first predicted trajectory. Furthermore, the first predicted trajectory can be extended to obtain a second predicted trajectory parallel to the tangent direction of the reference driving trajectory, thereby obtaining an accurate predicted trajectory and improving driving safety.

[0184] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for generating a predicted trajectory, characterized in that: The method comprises: Obtaining current state information and a reference driving trajectory of the target vehicle, wherein the state information includes a starting position, a starting speed, a starting acceleration, and a starting heading angle of the target vehicle, and the reference driving trajectory is a trajectory generated based on a target lane corresponding to the target vehicle; Obtaining a first predicted trajectory based on the starting position, the starting speed, the starting acceleration, the starting heading angle, and the reference driving trajectory, wherein the first predicted trajectory is composed of a plurality of circular arc curves, wherein the plurality of circular arc curves are obtained by trajectory fitting based on respective corresponding preview points, and the preview points are obtained based on the reference driving trajectory; Extending the first predicted trajectory to obtain a second predicted trajectory, wherein a tangent direction of the second predicted trajectory is parallel to a tangent direction of the reference driving trajectory and a distance between the second predicted trajectory and the reference driving trajectory is the same as a distance between an end point of the first predicted trajectory and the reference driving trajectory; A predicted trajectory of the target vehicle is generated based on the first predicted trajectory and the second predicted trajectory.

2. The method according to claim 1, characterized in that The obtaining of a first predicted trajectory based on the starting position, the starting speed, the starting acceleration, the starting heading angle, and the reference driving trajectory includes: During the current prediction process, a projection point position corresponding to the current prediction process is obtained based on the starting position, the first coordinate system, and the second coordinate system, where the projection point position is the projection position of the target vehicle on the reference driving trajectory; Obtaining a target curvature corresponding to the current prediction process based on the starting point position, the projection point position, the starting point speed, the starting point acceleration, and a preset preview time; Based on the target curvature, the starting point speed, the starting point acceleration, the starting point heading angle and the preset step time, the prediction point position, the prediction point speed, the prediction point acceleration and the prediction point heading angle corresponding to the current prediction process are obtained; Obtaining an arc curve corresponding to the current prediction process based on the starting point position, the prediction point position corresponding to the current prediction process, and the target curvature; The first predicted trajectory is obtained based on the circular arc curve and all circular arc curves before the current prediction process.

3. The method according to claim 2, characterized in that The obtaining of the target curvature corresponding to the current prediction process based on the starting point position, the projection point position, the starting point velocity, the starting point acceleration, and the preset preview time includes: Obtaining a preview point position for the current prediction process based on the projection point position, the starting point velocity, the starting point acceleration, and a preset preview time; Based on the starting point position and the preview point position, the target curvature corresponding to the current prediction process is obtained.

4. The method according to claim 3, characterized in that The step of obtaining the preview point position of the current prediction process based on the projection point position, the starting point velocity, the starting point acceleration, and a preset preview time includes: Based on the projection point position, the starting point speed, the starting point acceleration and the preset preview time, a reference preview point position of the current prediction process is obtained, wherein the reference preview point position is on the reference driving trajectory; The reference preview point position is converted from the second coordinate system to the first coordinate system to obtain the preview point position corresponding to the current prediction process.

5. The method according to claim 4, characterized in that The state information also includes the length of the target vehicle. The reference preview point position of the current prediction process is obtained based on the projection point position, the starting point speed, the starting point acceleration, and the preset preview time, including: Obtaining a preview distance for the current prediction process based on the starting point speed, the starting point acceleration, the length, and a preset preview time; Based on the projection point position and the preview distance of the current prediction process, a reference preview point position of the current prediction process is obtained.

6. The method according to claim 5, characterized in that The obtaining of the preview distance of the current prediction process based on the starting point speed, the starting point acceleration, the length, and the preset preview time includes: Obtaining a reference preview distance for the current prediction process based on the starting point speed, the starting point acceleration, the length, and a preset preview time; Comparing the reference preview distance of the current prediction process with a preset preview distance, and if the reference preview distance is greater than or equal to the preset preview distance, using the reference preview distance of the current prediction process as the preview distance of the current prediction process; If the reference preview distance is smaller than the preset preview distance, the preset preview distance is used as the preview distance for the current prediction process.

7. The method according to claim 3, characterized in that The state information also includes the length of the target vehicle. The target curvature corresponding to the current prediction process is obtained based on the starting point position and the preview point position, including: Based on the starting point position and the preview point position, a reference curvature corresponding to the current prediction process is obtained; Obtaining a curvature limit value of the target vehicle based on the starting speed and the length; If the reference curvature is greater than or equal to the curvature limit, the curvature limit is used as the target curvature corresponding to the current prediction process; If the reference curvature is less than the curvature limit, the reference curvature is used as the target curvature corresponding to the current prediction process.

8. The method according to claim 2, characterized in that The step of obtaining the predicted point position, predicted point speed, and predicted point heading angle corresponding to the current prediction process based on the target curvature, the starting point speed, the starting point heading angle, and the preset step time includes: Based on the starting point acceleration, the starting point velocity, and the step time, an arc length corresponding to the target curvature is obtained, where the arc length represents the distance between the starting point position corresponding to the current prediction process and the prediction point position in the arc corresponding to the target curvature; Obtaining a center position corresponding to the target curvature based on the target curvature, the starting point position, and the starting point heading angle; Based on the starting heading angle, the target curvature and the arc length, obtaining the heading angle of the prediction point corresponding to the current prediction process; Based on the starting point velocity, the starting point acceleration and the step time, obtaining the prediction point velocity corresponding to the current prediction process; Based on the starting point velocity and the starting point acceleration, obtaining the prediction point acceleration corresponding to the current prediction process; Based on the predicted point heading angle, the circle center position and the target curvature, the predicted point position corresponding to the current prediction process is obtained.

9. The method according to claim 2, characterized in that The obtaining of the first predicted trajectory based on the arc curve and all arc curves before the current prediction process includes: If the predicted point position corresponding to the current prediction process meets the target condition, the first predicted trajectory is obtained based on the arc curve corresponding to the current prediction process and all arc curves before the current prediction process; If the predicted point position corresponding to the current prediction process does not meet the target conditions, the predicted point position, predicted point speed, predicted point acceleration, and predicted point heading angle corresponding to the current prediction process will be used as the starting position, starting speed, starting acceleration, and starting heading angle of the next prediction process, respectively, for the next prediction.

10. The method according to claim 9, characterized in that After obtaining the predicted point position, predicted point velocity, predicted point acceleration, and predicted point heading angle corresponding to the current prediction process based on the target curvature, the starting point velocity, the starting point acceleration, the starting point heading angle, and the preset step time, the method further includes: Converting the predicted point position corresponding to the current prediction process from the first coordinate system to the second coordinate system to obtain the predicted projection point position corresponding to the current prediction process, wherein the predicted projection point position is on the reference driving trajectory; Obtaining a first tangential angle corresponding to the current prediction process, where the first tangential angle is the tangential angle of the reference driving trajectory at the predicted projection point; Obtaining a first difference based on the first tangential angle corresponding to the current prediction process and the heading angle of the prediction point; If the first difference corresponding to the current prediction process is within a preset range, it is determined that the prediction point position corresponding to the current prediction process meets the target condition.

11. The method according to claim 10, characterized in that The method further comprises: If the first difference corresponding to the current prediction process is not within the preset range, obtaining a second tangential angle corresponding to the current prediction process, where the second tangential angle is the tangential angle of the reference driving trajectory at the projection point; Obtaining a second difference based on the second tangential angle corresponding to the current prediction process and the starting heading angle; If the product of the first difference and the second difference corresponding to the current prediction process is less than or equal to 0, it is determined that the prediction point position corresponding to the current prediction process meets the target condition; If the product of the first difference and the second difference corresponding to the current prediction process is greater than 0, it is determined that the prediction point position corresponding to the current prediction process does not meet the target condition.

12. The method according to claim 2, characterized in that The method further comprises: If there are multiple preview times, obtain multiple reference prediction trajectories, wherein each reference prediction trajectory is generated based on the projection point position, the starting point velocity, the starting point acceleration, and the corresponding preview time; Obtaining a total cost corresponding to each of the plurality of reference predicted trajectories, where the total cost represents a cost for the target vehicle to travel along the reference travel trajectory; The reference prediction trajectory with the lowest total cost among the multiple reference prediction trajectories is used as the first prediction trajectory.

13. The method according to claim 12, characterized in that Each of the reference predicted trajectories is composed of a plurality of predicted points, each of the plurality of predicted points corresponds to a speed and a target curvature, and obtaining the total cost corresponding to each of the plurality of reference predicted trajectories includes: Based on the speeds and target curvatures of the multiple prediction points corresponding to each of the reference prediction trajectories, obtaining a centripetal acceleration cost corresponding to each of the multiple reference prediction trajectories, wherein the centripetal acceleration cost represents the magnitude of the centrifugal force generated during the driving process of the target vehicle; Based on the number of prediction points and the step time corresponding to each of the reference predicted trajectories, obtaining a lane change time cost corresponding to each of the plurality of reference predicted trajectories, wherein the lane change time cost represents the time required for the target vehicle to change lanes; Based on the target curvatures of the plurality of prediction points corresponding to each of the reference prediction trajectories, obtaining an average curvature change rate cost corresponding to each of the plurality of reference prediction trajectories, wherein the average curvature change rate cost represents a degree of smoothness of the target vehicle's travel; Obtaining a lateral deviation cost corresponding to each of the plurality of reference predicted trajectories based on a target lateral distance corresponding to each of the reference predicted trajectories, wherein the target lateral distance represents a lateral distance between the last predicted point of each reference predicted trajectory and the reference driving trajectory, and the lateral deviation cost is used to measure the size of the lateral distance; Based on the centripetal acceleration cost, lane change time cost, average curvature change rate cost, and lateral deviation cost corresponding to each of the multiple reference predicted trajectories, a total cost corresponding to each of the multiple reference predicted trajectories is obtained.

14. The method according to claim 1, wherein The first predicted trajectory is composed of a plurality of predicted points, each of which corresponds to a predicted point position and a speed. Extending the first predicted trajectory to obtain a second predicted trajectory includes: During the current prediction process, the last predicted point in the first predicted trajectory is converted from the first coordinate system to the second coordinate system, based on the reference point position of the last predicted point; Based on the predicted point position of the last predicted point and the reference driving trajectory, obtaining a lateral offset corresponding to the current prediction process, where the lateral offset represents the lateral distance between the last predicted point and the reference driving trajectory; Based on the reference point position, velocity and preset step time of the last prediction point, a longitudinal offset corresponding to the current prediction process is obtained; Based on the lateral offset and the longitudinal offset corresponding to the current prediction process, a prediction reference point position corresponding to the current prediction process is obtained, and the prediction reference point position corresponding to the next prediction process is obtained based on the prediction reference point position corresponding to the current prediction process, until a prediction reference trajectory is generated; The predicted reference trajectory is converted from the second coordinate system to the first coordinate system to obtain the second predicted trajectory.

15. A predicted trajectory generating device, characterized in that: The device comprises: an information acquisition unit, configured to acquire current state information and a reference driving trajectory of a target vehicle, wherein the state information includes a starting position, a starting speed, a starting acceleration, and a starting heading angle of the target vehicle, and the reference driving trajectory is a trajectory generated based on a target lane corresponding to the target vehicle; a first predicted trajectory generating unit, configured to obtain a first predicted trajectory based on the starting position, the starting speed, the starting acceleration, the starting heading angle, and the reference driving trajectory, wherein the first predicted trajectory is composed of a plurality of circular arc curves, wherein the plurality of circular arc curves are obtained by trajectory fitting based on corresponding preview points, and the preview points are obtained based on the reference driving trajectory; a second predicted trajectory generating unit, configured to extend the first predicted trajectory to obtain a second predicted trajectory, wherein a tangent direction of the second predicted trajectory is parallel to a tangent direction of the reference driving trajectory and a distance between the second predicted trajectory and the reference driving trajectory is the same as a distance between an end point of the first predicted trajectory and the reference driving trajectory; A predicted trajectory generating unit is configured to generate a predicted trajectory of the target vehicle based on the first predicted trajectory and the second predicted trajectory.

16. A vehicle, characterized in that: including one or more processors and memory; One or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to execute the method according to any one of claims 1 to 14.

17. A computer-readable storage medium, characterized in that The computer-readable storage medium stores program code, wherein when the program code is run, the method according to any one of claims 1 to 14 is executed.

Images