A method and apparatus for predicting the cutting intention of a target vehicle

Abstract

This invention relates to a method and apparatus for predicting the entry intention of a target vehicle, comprising: acquiring different motion states of a target vehicle under different scenarios; and predicting the entry intention of the target vehicle by adopting different judgment criteria according to the different motion states, thereby solving the problem that the entry intention of a target vehicle cannot be accurately predicted in the prior art.

Description

Technical Field

[0001] This invention relates to the field of intelligent safe driving technology, specifically to a method and apparatus for predicting the entry intention of a target vehicle. Background Technology

[0002] With the continuous development of autonomous driving and driver assistance technologies, automobiles are gradually entering the era of intelligence. However, how to accurately predict the lane-changing behavior of target vehicles (commonly known as "cutting in") to ensure the safety and comfort of autonomous driving remains a problem worth exploring.

[0003] To address the problem of predicting the cutting intention of target vehicles, current mainstream methods all employ neural network models. These models can predict the cutting intention of target vehicles in typical application scenarios where the target vehicle exhibits a clear movement trend. However, many atypical cutting scenarios exist in real-world roads, such as aggressive driving styles that disregard traffic rules or complex intersection road relationships. Because atypical cutting scenarios are infrequent, training samples are limited, making it difficult for neural network models to accurately predict the cutting intention of target vehicles, thus affecting the safety of autonomous driving. Summary of the Invention

[0004] One objective of this invention is to provide a method for predicting the entry intention of a target vehicle, so as to solve the problem that the entry intention of a target vehicle cannot be accurately predicted in the prior art; another objective is to provide a device for predicting the entry intention of a target vehicle.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A method for predicting the entry intention of a target vehicle includes: acquiring different motion states of a target vehicle under different scenarios; and predicting the entry intention of the target vehicle by using different judgment criteria according to the different motion states.

[0007] Based on the above technical means, different motion states of target vehicles in different scenarios can be obtained. Based on the different motion states of target vehicles, different judgment criteria are used to predict the entry intention of target vehicles. Therefore, without the need to collect training samples, it is possible to predict multiple entry intentions of vehicles more quickly and accurately, thereby improving the safety of autonomous driving.

[0008] Furthermore, the motion state of the target vehicle includes any of the following: the target vehicle deviating from the lane, the target vehicle merging into the lane, the target vehicle crossing the lane line, or the target vehicle moving close to the lane where the autonomous vehicle is located but not entering the lane where the autonomous vehicle is located.

[0009] Based on the aforementioned technical means, the vehicle's motion state can be determined.

[0010] Furthermore, the step of predicting the entry intention of the target vehicle by adopting different judgment criteria according to the different motion states includes:

[0011] The cut-in intention of the target vehicle is predicted based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located and the longitudinal distance between the target vehicle and the autonomous vehicle. The lane includes: lane center line or lane boundary line, or the cut-in intention of the target vehicle is predicted based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located.

[0012] Based on the aforementioned technical means, the cutting intention of the target vehicle can be predicted according to the distance between the target vehicle and the lane where the autonomous vehicle is located. This can avoid the problem of being unable to accurately predict the cutting intention of the target vehicle due to insufficient training samples, thereby improving the safety of autonomous driving.

[0013] Furthermore, the step of predicting the entry intention of the target vehicle based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located and the longitudinal distance between the target vehicle and the autonomous vehicle includes:

[0014] Obtain the first lateral distance P between the target vehicle and the lane. d ;

[0015] Determine whether the first lateral distance is less than the first distance determination parameter D offset ;

[0016] If it is less than the first distance determination parameter, then calculate the first lateral distance P. d With the first lateral velocity V of the target vehicle d Multiply by the product to obtain the first prediction result;

[0017] If the first prediction result is less than zero, then calculate the first distance from the target vehicle to the autonomous vehicle;

[0018] If the first distance is less than or equal to a set first distance threshold, then the first longitudinal distance P between the target vehicle and the autonomous vehicle is obtained. s ;

[0019] If the first longitudinal distance is greater than the set longitudinal distance threshold, it is determined that the target vehicle has the intention to cut in.

[0020] Based on the aforementioned technical means, by judging the lateral and longitudinal distances of the target vehicle, the movement of the target vehicle can be detected in a timely manner, and the timing of the target vehicle's entry can be accurately determined, thereby improving the safety of autonomous driving.

[0021] Furthermore, the step of predicting the entry intention of the target vehicle based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located and the longitudinal distance between the target vehicle and the autonomous vehicle includes:

[0022] Determine whether the lane where the target vehicle is located intersects with the lane where the autonomous vehicle is located;

[0023] If an intersection exists, then calculate the second distance from the target vehicle to the autonomous vehicle;

[0024] If the second distance is less than or equal to a set second distance threshold, then the second longitudinal distance P between the target vehicle and the autonomous vehicle is obtained. s ;

[0025] If the second longitudinal distance is greater than a set longitudinal distance threshold, then the second lateral distance P between the target vehicle and the lane where the autonomous vehicle is located is obtained. d ;

[0026] If the second lateral distance P d Less than the second distance judgment parameter D offset Calculate the second lateral distance P d With the second lateral velocity V of the target vehicle d Multiply by the product to obtain the second prediction result;

[0027] If the second prediction result is less than zero, then the second lateral velocity V of the target vehicle is determined. d Is it greater than the second lateral velocity threshold?

[0028] If so, it is determined that the target vehicle has the intention to cut in.

[0029] Based on the aforementioned technical means, by judging the lateral and longitudinal distances of the target vehicle, the movement of the target vehicle can be detected in a timely manner, and the timing of the target vehicle's entry can be accurately determined, thereby improving the safety of autonomous driving.

[0030] Furthermore, the step of predicting the cutting intention of the target vehicle based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located includes:

[0031] Determine whether the target vehicle has crossed the line;

[0032] If so, calculate the third lateral distance P of the target vehicle. dThe third lateral velocity V of the target vehicle d Multiply by the product to obtain the third prediction result;

[0033] If the third prediction result is less than zero, then it is determined whether the pressure line parameter is greater than the pressure line parameter judgment threshold, wherein the pressure line parameter includes: pressure line amount and pressure line amount change rate;

[0034] If so, obtain the third lateral velocity V of the target vehicle. d ;

[0035] If the third lateral velocity is greater than the velocity determination parameter Vcutin, then it is determined that the target vehicle has the intention to cut in.

[0036] If the third lateral velocity is less than the Vcutin, then the presence of a cut-in intention by the target vehicle is determined based on the line pressure and the heading angle of the target vehicle.

[0037] Based on the aforementioned technical means, by judging the longitudinal distance of the target vehicle, the movement of the target vehicle can be detected in a timely manner, and the timing of the target vehicle's entry can be accurately determined, thereby improving the safety of autonomous driving.

[0038] Furthermore, the distance determination parameter D is calculated in the following manner. offset :

[0039] Calculate the heading angle of the target vehicle's travel direction relative to the travel direction of the lane where the autonomous vehicle is located;

[0040] The vehicle length, vehicle width, lane width of the lane where the target vehicle is located, first lateral speed and lateral time distance of the target vehicle are obtained;

[0041] The distance determination parameters can be calculated based on the heading angle, vehicle length, vehicle width, lane width, first lateral speed, and lateral time distance, or the distance determination parameters can be calculated based on the heading angle, vehicle length, vehicle width, first lateral speed, and lateral time distance.

[0042] Furthermore, the method also includes:

[0043] Obtain the fourth lateral distance P between the target vehicle and the lane. d And the heading angle of the target vehicle's travel direction relative to the travel direction of the lane where the autonomous vehicle is located;

[0044] If the fourth lateral distance and the heading angle have the same sign, then it is determined whether the target vehicle has crossed the line;

[0045] If there is no line pressing, then calculate the fourth lateral distance P. dThe fourth lateral velocity V of the target vehicle d Multiply by the product to obtain the fourth prediction result; or determine the fourth lateral velocity V. d Is it less than or equal to the second lateral velocity threshold?

[0046] If the fourth prediction result is greater than or equal to zero; or if the fourth lateral velocity is less than or equal to the third lateral velocity threshold, then it is determined that the target vehicle does not have the intention to cut in.

[0047] Based on the above technical means, misjudgment of the target vehicle's intention to enter can be avoided.

[0048] Furthermore, the method also includes:

[0049] Obtain the fifth lateral distance P between the target vehicle and the lane. d ;

[0050] Determine whether the fifth lateral distance is greater than the first lane width threshold;

[0051] If it is greater than the lane width threshold, then calculate the fifth lateral distance P. d The fifth lateral velocity V of the target vehicle d Multiply by the product to obtain the fifth prediction result; or determine the fifth lateral velocity V. d Is it less than or equal to the fifth lateral velocity threshold?

[0052] If the fifth prediction result is less than zero, or the fifth lateral velocity is less than the fifth lateral velocity threshold, then the fifth lateral distance P between the target vehicle and the lane is calculated. d The sixth prediction result is obtained by multiplying the heading angle of the target vehicle's travel direction relative to the travel direction of the autonomous vehicle's lane.

[0053] If the sixth prediction result is less than zero, or the heading angle is less than the first angle threshold, then the third distance dis2line between the target vehicle and the lane where the autonomous vehicle is located is calculated.

[0054] If the third distance dis2line is less than the distance threshold parameter D dis2line Or the fifth lateral distance P d If the distance is less than the second lane width threshold, then the longitudinal time distance of the target vehicle is obtained;

[0055] If the longitudinal time interval is within the set range, then the target vehicle is determined to be in an intermediate state of the intended entry.

[0056] Based on the aforementioned technical means, target vehicles whose entry intentions are not predicted are further screened to determine whether there is a risk of entry intentions, thereby ensuring the safety of autonomous driving.

[0057] A target vehicle cutting-in intention prediction device, comprising:

[0058] The motion state module is used to acquire different motion states of the target vehicle in different scenarios;

[0059] The prediction module is used to predict the entry intention of the target vehicle based on different judgment criteria according to the different motion states.

[0060] Furthermore, the motion state of the target vehicle includes any of the following: the target vehicle deviating from the lane, the target vehicle merging into the lane, the target vehicle crossing the lane line, or the target vehicle moving close to the lane where the autonomous vehicle is located but not entering the lane where the autonomous vehicle is located.

[0061] Furthermore, the prediction module includes:

[0062] The first prediction intent unit is used to predict the entry intent of the target vehicle based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located and the longitudinal distance between the target vehicle and the autonomous vehicle. The lane includes: lane center line or lane boundary line, or the entry intent of the target vehicle is predicted based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located.

[0063] Furthermore, the first predictive intent unit includes:

[0064] The first lateral submodule is used to obtain the first lateral distance P between the target vehicle and the lane. d ;

[0065] The first judgment submodule is used to determine whether the first lateral distance is less than the first distance judgment parameter D. offset ;

[0066] The first prediction submodule is used to determine if the distance is less than the first distance judgment parameter D. offset Then calculate the first lateral distance P. d With the first lateral velocity V of the target vehicle d Multiply by the product to obtain the first prediction result;

[0067] The first calculation submodule is used to calculate the first distance from the target vehicle to the autonomous vehicle if the first prediction result is less than zero.

[0068] The first longitudinal submodule is used to obtain the first longitudinal distance P between the target vehicle and the autonomous vehicle if the first distance is less than or equal to a set first distance threshold. s ;

[0069] The first cutting-in submodule is used to determine that the target vehicle has a cutting-in intention if the first longitudinal distance is greater than a set longitudinal distance threshold.

[0070] Furthermore, the first predictive intent unit includes:

[0071] The second judgment submodule is used to determine whether the lane where the target vehicle is located intersects with the lane where the autonomous vehicle is located.

[0072] The second calculation submodule is used to calculate the second distance from the target vehicle to the autonomous vehicle if there is an intersection.

[0073] The second longitudinal submodule is used to obtain the second longitudinal distance P between the target vehicle and the autonomous vehicle if the second distance is less than or equal to a set second distance threshold. s ;

[0074] The second lateral submodule is used to obtain the second lateral distance P between the target vehicle and the lane where the autonomous vehicle is located if the second longitudinal distance is greater than a set longitudinal distance threshold. d ;

[0075] The second prediction submodule is used to predict if the second lateral distance P d Less than the second distance judgment parameter D offset Then calculate the second lateral distance P. d With the second lateral velocity V of the target vehicle d Multiply by the product to obtain the second prediction result;

[0076] The third judgment submodule is used to determine the second lateral velocity V of the target vehicle if the second prediction result is less than zero. d Is it greater than the second lateral velocity threshold?

[0077] The second cut-in submodule is used to determine that the target vehicle has a cut-in intention if the speed is greater than the second lateral speed threshold.

[0078] Furthermore, the first predictive intent unit includes:

[0079] The fourth judgment submodule is used to determine whether the target vehicle has crossed the line;

[0080] The third calculation submodule is used to calculate the third lateral distance P of the target vehicle if the condition is met. d The third lateral velocity V of the target vehicled Multiply by the product to obtain the third prediction result;

[0081] The fifth judgment submodule is used to determine whether the pressure line parameter is greater than the pressure line parameter judgment threshold if the third prediction result is less than zero. The pressure line parameter includes: pressure line amount and pressure line amount change rate.

[0082] The third lateral submodule is used to, if so, obtain the third lateral velocity V of the target vehicle. d ;

[0083] The third cutting-in submodule is used to determine whether the target vehicle has a cutting-in intention if the third lateral speed is greater than the speed determination parameter Vcutin; and to determine whether the target vehicle has a cutting-in intention based on the line overlap amount and the heading angle of the target vehicle if the third lateral speed is less than Vcutin.

[0084] Furthermore, the distance determination parameter D is calculated in the following manner. offset :

[0085] Calculate the heading angle of the target vehicle's travel direction relative to the travel direction of the lane where the autonomous vehicle is located;

[0086] The vehicle length, vehicle width, lane width of the lane where the target vehicle is located, first lateral speed and lateral time distance of the target vehicle are obtained;

[0087] The distance determination parameters can be calculated based on the heading angle, vehicle length, vehicle width, lane width, first lateral speed, and lateral time distance, or the distance determination parameters can be calculated based on the heading angle, vehicle length, vehicle width, first lateral speed, and lateral time distance.

[0088] Furthermore, the device also includes:

[0089] The first acquisition module is used to acquire the fourth lateral distance P between the target vehicle and the lane. d And the heading angle of the target vehicle's travel direction relative to the travel direction of the lane where the autonomous vehicle is located;

[0090] The first judgment module is used to determine whether the target vehicle has crossed the line if the fourth lateral distance and the heading angle have opposite signs.

[0091] The first calculation module is used to calculate the fourth lateral distance P if there is no line pressing. d The fourth lateral velocity V of the target vehicle d Multiply by the product to obtain the fourth prediction result; or determine the fourth lateral velocity V. d Is it less than or equal to the second lateral velocity threshold?

[0092] The output module is used to determine that the target vehicle does not have the intention to cut in if the fourth prediction result is greater than or equal to zero, or the fourth lateral velocity is less than or equal to the third lateral velocity threshold.

[0093] Furthermore, the device also includes:

[0094] The second acquisition module is used to acquire the fifth lateral distance P between the target vehicle and the lane. d ;

[0095] The second judgment module is used to determine whether the fifth lateral distance is greater than the first lane width threshold.

[0096] The second calculation module is used to calculate the fifth lateral distance P if it is greater than the lane width threshold. d The fifth lateral velocity V of the target vehicle d Multiply by the product to obtain the fifth prediction result; or determine the fifth lateral velocity V. d Is it less than or equal to the fifth lateral velocity threshold?

[0097] The deviation module is used if the fifth prediction result is less than zero, or if the fifth lateral velocity V... d If the speed is less than the fifth lateral velocity threshold, then the fifth lateral distance P between the target vehicle and the lane is calculated. d The sixth prediction result is obtained by multiplying the target vehicle's direction of travel by the heading angle relative to the direction of travel of the autonomous vehicle in the lane.

[0098] The centerline distance module is used to calculate the third distance dis2line between the target vehicle and the lane where the autonomous vehicle is located if the sixth prediction result is less than zero, or if the heading angle is less than the first angle threshold.

[0099] The longitudinal time-distance module is used to determine if the third distance dis2line is less than the distance threshold parameter D. dis2line Or the fifth lateral distance P d If the distance is less than the second lane width threshold, then the longitudinal time distance of the target vehicle is obtained;

[0100] The status module is used to determine that the target vehicle is in an intermediate state of the cutting intention if the longitudinal time distance is within a set range.

[0101] The beneficial effects of this invention are:

[0102] (1) This invention obtains different motion states of target vehicles in different scenarios, and uses different judgment criteria to predict the entry intention of the target vehicle based on the different motion states of the target vehicle. Therefore, without the need to collect training samples, it can more quickly and accurately predict the multiple entry intentions of the vehicle, thereby improving the safety of autonomous driving.

[0103] (2) By judging the lateral and longitudinal distances of the target vehicle, the present invention can detect the movement of the target vehicle in a timely manner and accurately determine the timing of the target vehicle's entry, thereby achieving accurate determination of the entry intention and improving the safety of autonomous driving.

[0104] (3) The present invention further filters the target vehicle that is predicted to cut in, and determines whether there will be a misjudgment of the cutting in intention, thereby achieving accurate determination of the cutting in intention and improving the safety of autonomous driving. Attached Figure Description

[0105] Figure 1 This is an overall design flowchart of a target vehicle cutting intention prediction method according to the present invention;

[0106] Figure 2 This is a flowchart of a method for predicting the entry intention of a target vehicle according to the present invention;

[0107] Figure 3 This is a flowchart of a method for predicting the cutting intention of a target vehicle according to the present invention;

[0108] Figure 4 This is a flowchart of a method for predicting the cutting intention of a target vehicle according to the present invention;

[0109] Figure 5 This is a flowchart of a method for predicting the cutting intention of a target vehicle according to the present invention;

[0110] Figure 6 This is a flowchart of a method for handling misjudgment of a target vehicle's cutting intention according to the present invention;

[0111] Figure 7 This is a flowchart of a method for predicting the entry intention of a target vehicle according to the present invention;

[0112] Figure 8 This is a schematic diagram of the structure of a target vehicle cutting intention prediction device according to the present invention. Detailed Implementation

[0113] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0114] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0115] See Figure 1 The diagram illustrates the overall design flowchart of a target vehicle cutting intention prediction method according to the present invention, which specifically includes:

[0116] Step 101: Obtain road information and status information of all detectable target vehicles in the coordinate system of the autonomous vehicle.

[0117] Specifically, the coordinate system of an autonomous vehicle generally refers to the vehicle coordinate system, or VCS coordinate system. An autonomous vehicle can obtain road information within a certain range from a high-precision map at the current moment. This road information includes the boundary coordinates and centerline coordinates of all lanes within the effective range. It also obtains the state information of all target vehicles that can be detected by sensors within a certain range from the perception fusion unit at the current moment. This target vehicle state information includes the target vehicle's motion state, direction of travel, speed, acceleration, etc.

[0118] Step 102: Obtain the lane information of the autonomous vehicle at the current moment.

[0119] Specifically, the system obtains the centerline coordinates and lane width of the lane the autonomous vehicle is currently traveling in. It then projects the vehicle's current position onto all lane centerlines and calculates the perpendicular distance from the vehicle to each lane centerline. If the perpendicular distance from the vehicle to each lane centerline is less than or equal to the lane width, the lane with the closest perpendicular distance is selected as the vehicle's current lane. If there are two or more lanes where the perpendicular distance from the vehicle to each lane centerline is less than or equal to the lane width, the lane whose direction of travel is closest to the vehicle's direction of travel is selected as the vehicle's current lane.

[0120] Step 103: Calculate the position and velocity information of the target vehicle in different coordinate systems.

[0121] The coordinate system can be the Frenet coordinate system or the VCS coordinate system, or other coordinate systems. This invention does not impose any specific restrictions on this.

[0122] The following example uses the Frenet coordinate system to illustrate how to calculate the position and speed information of a target vehicle in different coordinate systems. The Frenet coordinate system is constructed using the center line of the lane where the vehicle is located as the reference line, and the position and speed information of the target vehicle in the Frenet coordinate system are calculated.

[0123] Specifically, using the centerline of the lane where the vehicle is currently located as the reference line, a Frenet coordinate system is established using the tangent and normal vectors of the reference line. The initial point is the projection of the vehicle itself onto the reference line. This Frenet coordinate system includes S-direction and D-direction. The S-direction refers to the distance along the reference line, i.e., along the road direction, usually called the longitudinal direction. The D-direction refers to the distance deviating from the reference line (perpendicular to the normal vector of the reference line), usually called the lateral direction. The coordinate transformation calculation is from the VCS coordinate system (vehicle coordinate system) to the Frenet coordinate system, which makes it easier to obtain the tangential and normal positions of the target vehicle on the reference path (p). s ,p d ), velocity (v) s ,v d ) and acceleration (a s ,a d )information.

[0124] For the specific calculation process of converting from the VCS coordinate system (vehicle coordinate system) to the Frenet coordinate system, please refer to the coordinate transformation methods in related technologies. This invention will not elaborate on it here.

[0125] It should be noted that, in order to meet subsequent needs, the Frenet coordinate system is constructed in this invention. However, the solution of this invention can still be used for calculations even if the Frenet coordinate system is not used.

[0126] Step 104: Estimate the position of all target vehicles after a certain period of time, and preliminarily determine whether the target vehicles have the intention to cut in based on whether they enter the vehicle lane at that position.

[0127] Specifically, the position of each target vehicle in the Frenet coordinate system or other coordinate systems is calculated after a certain prediction time. If the target vehicle is in the Frenet coordinate system or other coordinate systems, then p dIf the absolute value is less than or equal to 1 / 2 of the lane width, it indicates that the target vehicle will enter the lane after a certain predicted time, and has the intention to cut in. In typical scenarios, for target vehicles with obvious movement trends, this step can basically lead to a preliminary conclusion about their cutting intention; that is, whether the target vehicle is cutting in or not.

[0128] Step 105: Optimize the cut-in intent prediction results for target vehicles in various application scenarios.

[0129] In complex real-world application scenarios, target vehicles with minimal lateral displacement or speed can exert significant pressure on the vehicle itself. However, step 104 above cannot predict the merging intention in time, leading to collision risks or a poor driving experience. Therefore, this invention proposes classifying the target vehicle's state into various extreme scenarios based on information such as its heading, lane marking, degree of movement towards the vehicle's lane, time to collision (TTC), and whether it is a merging target. Different judgment criteria are then used to predict the merging intention based on these extreme scenarios.

[0130] Special Scenario 1: The target vehicle (truck, trailer, bus, etc.) is driving close to the lane line and slowly encroaching into the lane of the vehicle. During this process, there are no significant lateral speed or frontal orientation motion characteristics, making it difficult or untimely to predict the intention, thus creating a collision risk.

[0131] Special Scenario 2: A target vehicle that is continuously driving over the lane line suddenly veers into the lane of the driver. This type of scenario requires a rapid response from the driver, placing high demands on the accuracy and timeliness of the prediction algorithm.

[0132] Special Scenario 3: The target vehicle quickly cuts into this lane at close range. This type of scenario usually occurs in traffic congestion or with target vehicles with aggressive driving styles, and requires high timeliness of the prediction algorithm.

[0133] Special Scenario 4: In merging scenarios, most target vehicles will intrude into the driver's lane before the merging point. The closer to the merging point, the narrower the target vehicle's driving space, and the greater the threat to the driver. This type of scenario requires the prediction algorithm to predict the intrusion intention as far in advance as possible, allowing time for the driver's decision-making and system response.

[0134] Special Scenario 5: A target vehicle rapidly approaches the adjacent lane from the lane next to it. Although it does not encroach on the driver's lane, it creates a strong sense of urgency for the driver. This type of scenario requires predicting the target vehicle's potential intentions to provide prior knowledge for the driver's safety decisions.

[0135] Special Scenario 6: A target vehicle attempts to encroach on the driver's lane from an adjacent lane. The intent behind this process may be discontinuous or ambiguous. Although the target vehicle ultimately does not enter the driver's lane, the interaction may create pressure and threat to the driver. This type of scenario also requires predicting the target vehicle's potential intent to provide prior knowledge for the smoothness and comfort of the driver's decision-making.

[0136] Based on the special scenarios 1-6 above, the entry intent of the target vehicle is predicted. For special scenarios 1-4, the following can be used: Figure 2-6 The method predicts the entry intention of the target vehicle. Special scenarios 5-6 employ... Figure 7 The method is used to predict the entry intention of the target vehicle.

[0137] Step 106: Optimize for misjudgments that indicate an intention to enter the market.

[0138] Specifically, after steps 104 and 105, some target vehicles may be judged as having the intention to cut in, but in reality, some target vehicles may simply be continuously crossing the line without ultimately cutting into the lane. Therefore, in step 106, targets that are continuously crossing the line but do not show any further tendency to cut in will be processed to avoid misjudging the cutting intention of the target vehicles. For details, please refer to [link to relevant documentation]. Figure 6 The following embodiments will provide a detailed explanation of the misjudgment of the cutting intent.

[0139] Step 107: Predict the target vehicles that may attempt to cut in.

[0140] Specifically, because the target vehicle engages in a game with the chariot vehicle during the process of generating the intention to cut in, accurately predicting the timing of the target vehicle's cutting-in intention presents a significant challenge. If the above steps fail to predict the cutting-in intention of a target vehicle, but there is a potential risk of it cutting into the chariot vehicle's lane, this invention will predict the target vehicle that may be attempting to cut in and output an intermediate state of the cutting-in intention. This intermediate state is the attempt to cut in; that is, the intermediate state is between cutting in and not cutting in.

[0141] The above steps detail the overall concept of the present invention for predicting the entry intention of a target vehicle. The following embodiments will mainly focus on steps 105-107 to describe in detail the process of predicting the entry intention of a target vehicle.

[0142] See Figure 2 The flowchart below illustrates a method for predicting the cutting intention of a target vehicle, as provided in an embodiment of the present invention. The method includes:

[0143] Step 201: Obtain the different motion states of the target vehicle under different scenarios.

[0144] Different scenarios include special scenarios 1-6 in the aforementioned embodiments, which will not be described again here.

[0145] The motion state of the target vehicle includes any of the following: the target vehicle deviates from the lane, the target vehicle merges into the lane, the target vehicle crosses the line, or the target vehicle is driving close to the lane where the autonomous vehicle is located but has not entered the lane where the autonomous vehicle is located.

[0146] The target vehicle deviating from its lane generally refers to the target vehicle deviating from its center line. The target vehicle crossing the line generally refers to the target vehicle traveling along the boundary line of the autonomous vehicle's lane; that is, the target vehicle has crossed the line and its front end or the entire vehicle has entered the autonomous vehicle's lane. The target vehicle approaching the autonomous vehicle's lane but not entering the autonomous vehicle's lane generally refers to the target vehicle crossing the line but not entering the autonomous vehicle's lane, or the target vehicle being very close to the autonomous vehicle but not entering its lane. Different methods will be used to predict the target vehicle's input intent for different movement states. In subsequent embodiments, the prediction of the target vehicle's input intent for different movement states will be described in detail.

[0147] Step 202: Based on different judgment criteria for different motion states, predict the cutting intention of the target vehicle.

[0148] In practical applications, the lateral distance between the target vehicle and the lane where the autonomous vehicle is located, and the longitudinal distance between the target vehicle and the autonomous vehicle, are used to predict the entry intention of the target vehicle. The lane includes the lane centerline or lane boundary line. That is, the entry intention of the target vehicle can be predicted based on the lateral distance between the lane centerline and the longitudinal distance between the target vehicle and the autonomous vehicle, or it can be predicted based on the lateral distance between the lane boundary lines and the longitudinal distance between the target vehicle and the autonomous vehicle. Furthermore, the entry intention of the target vehicle can be predicted based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located. This embodiment acquires different motion states of the target vehicle in different scenarios and uses different judgment criteria based on these different motion states to predict the entry intention of the target vehicle. Therefore, without the need to collect training samples, it can more quickly and accurately predict multiple entry intentions of the vehicle, thereby improving the safety of autonomous driving.

[0149] See Figure 3This is a flowchart of a method for predicting the cutting intention of a target vehicle according to an embodiment of the present invention. This embodiment mainly describes the method for predicting the cutting intention of a target vehicle when the target vehicle's motion state is that the target vehicle is deviating from its lane. Specifically, it includes:

[0150] Step 301: Obtain the motion state of the target vehicle.

[0151] The movement state of the target vehicle includes: the target vehicle deviating from the lane, that is, the target vehicle deviating from the center line where the target vehicle is located. In other words, it is the process of cutting into the target vehicle with a clear frontal orientation. That is, the target vehicle deviates to the left or right of the center line. Generally, when deviating to the left, the angle of the target vehicle's orientation is positive, and when deviating to the right, the angle of the target vehicle's orientation is negative.

[0152] Step 302: Obtain the first lateral distance P between the target vehicle and the lane. d .

[0153] In practical applications, the first lateral distance P between the target vehicle and the centerline of the driver's lane is obtained. d Alternatively, obtain the first lateral distance P between the target vehicle and the lane boundary line of the vehicle. d .

[0154] For example, obtain the first lateral distance P between the target vehicle and the lane in the Frenet coordinate system. d This means obtaining the normal vector perpendicular to the reference line.

[0155] Step 303: Determine whether the first lateral distance is less than the first distance determination parameter D. offset .

[0156] In practical applications, it is determined whether the absolute value of the first lateral distance is less than the first distance determination parameter D. offset If it is less than, then calculate the first lateral distance P. d With the first lateral velocity V of the target vehicle d Multiply by the product to obtain the first prediction result; if it is greater than the first prediction result, the process ends.

[0157] In practical applications, when the lane centerline is used as a reference line, the distance determination parameter D is calculated in the following way. offset :

[0158] First, calculate the heading angle delta_heading of the target vehicle's travel direction relative to the travel direction of the vehicle in the lane. Specifically, the target vehicle's travel direction is composed of a series of coordinate points, from which the angle of the target vehicle's travel direction can be obtained. The travel direction of the vehicle in the lane is also composed of a series of coordinate points, from which the angle of the vehicle in the lane can be obtained. Then, the difference between the two angles is used as the heading angle delta_heading.

[0159] Determine the first lateral distance P between delta_heading and the target vehicle in the Frenet coordinate system. d Check if the coordinate values ​​have different signs. If they do, it means the target vehicle intends to approach the center line of the lane. Different signs generally refer to different symbols, i.e., one positive and one negative. If the symbols are the same, they are not different.

[0160] Second, obtain the length and width of the target vehicle, the lane width of the lane where the target vehicle is located, the first lateral speed of the target vehicle, and the lateral time distance.

[0161] Third, the distance determination parameters are calculated based on the heading angle, vehicle length, vehicle width, lane width, first lateral speed, and lateral time distance.

[0162] In practical applications, different angle thresholds are set for the heading angle, which are then divided into different piecewise functions. Different distance determination parameters D are used for different heading angles. offset and velocity determination parameter V cutin For target vehicles that are small in size, such as sedans, the distance determination parameter D can be calculated in the following way. offset .

[0163] Specifically, D can be calculated using formulas 1 and 2 below. offset :

[0164]

[0165]

[0166] The angle thresholds include a second angle threshold, a third angle threshold, and a fourth angle threshold, wherein the second angle threshold is less than the third angle threshold, and the third angle threshold is less than the fourth angle threshold. L is the length of the target vehicle, θ is the orientation of the target vehicle, D is the width of the target vehicle, and the lane width is R. dThe first lateral velocity of the target vehicle is given by delta_heading, the heading angle is given by delta_heading, and T is the lateral time distance, which can be 2.5, 2.0, 1.5, or 0.5. If delta_heading is less than or equal to the second angle threshold, T is 0.5; if delta_heading is greater than the second angle threshold but less than the third angle threshold, T is 1.5; if delta_heading is greater than the third angle threshold but less than the fourth angle threshold, T is 2.0; and if delta_heading is greater than or equal to the fourth angle threshold, T is 2.5.

[0167] Furthermore, if the target vehicle is a large type, such as a truck, bus, or trailer, and the target vehicle intends to approach the center line of the vehicle's lane, then the first distance determination parameter and speed determination parameter V need to be recalculated. cutin Specifically, D can be calculated using formulas 3 and 4 below. offset and velocity determination parameter V cutin

[0168]

[0169]

[0170] The angle thresholds include a first angle threshold and a second angle threshold, wherein the first angle threshold is less than the second angle threshold. L is the length of the target vehicle, θ is the orientation of the target vehicle, D is the vehicle's angle, R is the lane width, and v... d The first lateral velocity of the target vehicle is given by delta_heading, the heading angle is given by T, and the lateral time distance can be 1.5 or 0.5. If delta_heading is less than or equal to the first angle threshold, T is 0.5; if delta_heading is greater than the first angle threshold but less than the second angle threshold, T is 1.5; if delta_heading is greater than or equal to the second angle threshold, T is 1.5.

[0171] In practical applications, when the lane boundary line is used as a reference line, the distance determination parameter D is calculated in the following way. offset :

[0172] First, calculate the heading angle delta_heading of the target vehicle's travel direction relative to the travel direction of the vehicle in the lane. Specifically, the target vehicle's travel direction is composed of a series of coordinate points, from which the angle of the target vehicle's travel direction can be obtained. The travel direction of the vehicle in the lane is also composed of a series of coordinate points, from which the angle of the vehicle in the lane can be obtained. Then, the difference between the two angles is used as the heading angle delta_heading.

[0173] Determine the first lateral distance P between delta_heading and the target vehicle in the Frenet coordinate system. d Check if the coordinate values ​​have different signs. If they do, it means the target vehicle intends to approach the boundary line of the vehicle's lane. Different signs generally refer to different symbols, i.e., one positive and one negative. If the symbols are the same, they are not different.

[0174] Second, obtain the length, width, first lateral speed, and lateral time distance of the target vehicle.

[0175] Third, the distance determination parameters are calculated based on the heading angle, vehicle length, vehicle width, first lateral speed, and lateral time distance.

[0176] In practical applications, different angle thresholds are set for the heading angle, which are then divided into different piecewise functions. Different distance determination parameters D are used for different heading angles. offset and velocity determination parameter V cutin For target vehicles that are small in size, such as sedans, the distance determination parameter D can be calculated in the following way. offset .

[0177] Specifically, D can be calculated using formulas 1 and 2 below. offset :

[0178]

[0179]

[0180] The angle thresholds include a second angle threshold, a third angle threshold, and a fourth angle threshold, wherein the second angle threshold is less than the third angle threshold, and the third angle threshold is less than the fourth angle threshold. L represents the length of the target vehicle, θ represents the orientation of the target vehicle, D represents the width of the target vehicle, and v... d The first lateral velocity of the target vehicle is given by delta_heading, the heading angle is given by delta_heading, and T is the lateral time distance, which can be 2.5, 2.0, 1.5, or 0.5. If delta_heading is less than or equal to the second angle threshold, T is 0.5; if delta_heading is greater than the second angle threshold but less than the third angle threshold, T is 1.5; if delta_heading is greater than the third angle threshold but less than the fourth angle threshold, T is 2.0; and if delta_heading is greater than or equal to the fourth angle threshold, T is 2.5.

[0181] Furthermore, if the target vehicle is a large type, such as a truck, bus, or trailer, and the target vehicle intends to approach the boundary line of the vehicle's lane, then it is necessary to recalculate the first distance determination parameter and the speed determination parameter V.cutin Specifically, D can be calculated using formulas 3 and 4 below. offset and velocity determination parameter V cutin

[0182]

[0183]

[0184] The angle thresholds include a first angle threshold and a second angle threshold, wherein the first angle threshold is less than the second angle threshold. L is the length of the target vehicle, θ is the orientation of the target vehicle, D is the vehicle's angle, and v... d The first lateral velocity of the target vehicle is given by delta_heading, the heading angle is given by T, and the lateral time distance can be 1.5 or 0.5. If delta_heading is less than or equal to the first angle threshold, T is 0.5; if delta_heading is greater than the first angle threshold but less than the second angle threshold, T is 1.5; if delta_heading is greater than or equal to the second angle threshold, T is 1.5.

[0185] Step 304: If the first lateral distance is less than the first distance determination parameter, then calculate the first lateral distance P. d With the first lateral velocity V of the target vehicle d Multiply by the product to obtain the first prediction result.

[0186] Step 305: If the first prediction result is less than zero, calculate the first distance between the target vehicle and the autonomous vehicle.

[0187] The first prediction result can be used to determine whether the target vehicle is close to the autonomous vehicle lane. If the first prediction result is greater than zero, it means that the target vehicle is on the left side of the lane and is not close to the autonomous vehicle lane line. If it is less than zero, it means that the target vehicle is on the right side of the lane and is close to the autonomous vehicle lane, and there is a possibility that it intends to cut in. Then, the first distance between the target vehicle and the autonomous vehicle is calculated, that is, the distance from the rear of the target vehicle to the front of the autonomous vehicle is calculated.

[0188] Step 306: If the first distance is less than or equal to a set first distance threshold, then obtain the first longitudinal distance Ps between the target vehicle and the autonomous vehicle.

[0189] In practical applications, if the first distance is less than or equal to the set first distance threshold, it means that the target vehicle needs to cut into the vehicle's lane. At this time, it is necessary to determine whether the longitudinal distance is acceptable. Based on the judgment result, it is finally determined whether the target vehicle intends to cut in. The first distance threshold can be calculated based on the lateral time distance and the vehicle speed. The lateral time distance is a fixed value, such as 2 or 3.5. In this invention, the fixed value is preferably set to 2.

[0190] Step 307: If the first longitudinal distance is greater than the set longitudinal distance threshold, it is determined that the target vehicle has the intention to cut in.

[0191] In practical applications, the longitudinal distance threshold is a fixed value, typically negative 3 meters.

[0192] In this embodiment, by judging the lateral and longitudinal distances of the target vehicle, the movement of the target vehicle can be detected in a timely manner, and the timing of the target vehicle's entry can be accurately determined, thereby improving the safety of autonomous driving.

[0193] See Figure 4 This is a flowchart of a method for predicting the cutting intention of a target vehicle according to an embodiment of the present invention. This embodiment mainly illustrates that when the target vehicle's motion state is that of a target vehicle merging into a lane, i.e., in a merging scenario, most targets will intrude into the lane before the merging point. The closer to the merging point, the narrower the target's driving space and the greater the threat to the vehicle. The method for predicting the cutting intention of the target vehicle specifically includes:

[0194] Step 401: Obtain the motion state of the target vehicle.

[0195] The motion state of the target vehicle includes: the state of the target vehicle merging into the lane.

[0196] Step 402: Determine whether the lane where the target vehicle is located intersects with the lane where the autonomous vehicle is located. If there is an intersection, proceed to step 403.

[0197] In practical applications, if the lane center line is used as a reference line, it is determined whether the center line of the lane where the target vehicle is located intersects with the center line of the lane where the autonomous vehicle is located. If the center line of the lane where the target vehicle is located intersects with the center line of the lane where the autonomous vehicle is located, it indicates that they are merging.

[0198] In practical applications, if the lane boundary line is used as a reference line, it is determined whether the lane boundary line where the target vehicle is located intersects with the lane boundary line where the autonomous vehicle is located. If the lane boundary line where the target vehicle is located intersects with the lane boundary line where the autonomous vehicle is located, it indicates that they are merging.

[0199] Step 403: Calculate the second distance from the target vehicle to the autonomous vehicle.

[0200] In practical applications, the second distance from the rear of the target vehicle to the front of the vehicle is calculated, and this distance can be used to determine whether it is possible to cut in.

[0201] Step 404: If the second distance is less than or equal to a set second distance threshold, then obtain the second longitudinal distance P between the target vehicle and the autonomous vehicle. s .

[0202] If the second distance is less than or equal to the set second distance threshold, it indicates that there is a trend of the target vehicle cutting in, and the second longitudinal distance between the target vehicle and the autonomous vehicle is obtained.

[0203] In practical applications, the second distance threshold can be calculated based on the lateral time distance and the vehicle speed. The lateral time distance is a fixed value, such as 2 or 3.5, while the fixed value of this invention is set to 3.5.

[0204] Step 405: If the second longitudinal distance is greater than a set longitudinal distance threshold, then obtain the second lateral distance P between the target vehicle and the lane where the autonomous vehicle is located. d .

[0205] In practical applications, the longitudinal distance threshold is a fixed value, typically negative 3 meters.

[0206] Step 406: If the second lateral distance P d Less than the second distance judgment parameter D offset Calculate the second lateral distance P d With the second lateral velocity V of the target vehicle d Multiply by the product to obtain the second prediction result.

[0207] Step 407: If the second prediction result is less than zero, then determine the second lateral velocity V of the target vehicle. d If the lateral velocity is greater than the second lateral velocity threshold, then proceed to step 408.

[0208] Specifically, if the absolute value of the second lateral distance is less than the second distance determination parameter, then the second lateral distance P is calculated. d With the second lateral velocity V of the target vehicle d The product is used to obtain a second prediction result. This second prediction result can be used to determine whether the target vehicle is close to the driving lane. If the second prediction result is greater than zero, it means that the target vehicle is on the left side of the lane and is not close to the driving lane line. If it is less than zero, it means that the target vehicle is on the right side of the lane and is close to the driving lane. Then, the second lateral velocity V of the target vehicle is determined. d Whether it is greater than the second lateral speed threshold, that is, whether the absolute value of the second lateral speed of the target vehicle is greater than the second lateral speed threshold. If so, it is determined that the target vehicle has the intention to cut in.

[0209] The threshold for the second distance determination parameter can be set by anyone skilled in the art in any appropriate manner. For example, the threshold can be set manually based on experience, or a threshold can be set based on the difference in historical data. Preferably, the second distance determination parameter can be 1.3 * the first distance determination parameter D. offset , where D offset The calculation method is described in the example. Figure 3 The method described in the previous article is sufficient, and will not be elaborated further in this invention.

[0210] Step 408: Determine that the target vehicle has the intention to cut in.

[0211] In this embodiment, by judging the lateral and longitudinal distances of the target vehicle, the movement of the target vehicle can be detected in a timely manner, and the timing of the target vehicle's entry can be accurately determined, thereby improving the safety of autonomous driving.

[0212] See Figure 5 This is a flowchart of a method for predicting the cutting intention of a target vehicle according to an embodiment of the present invention. This embodiment mainly describes the method for predicting the cutting intention of a target vehicle when the target vehicle is in a position where it is crossing the line. Specifically, it includes:

[0213] Step 501: Obtain the motion state of the target vehicle.

[0214] The motion state of the target vehicle includes: the target vehicle crossing the line.

[0215] Step 502: Determine whether the target vehicle has crossed the line. If so, proceed to step 503.

[0216] "Trapping the line" generally refers to a vehicle driving on the boundary line of a lane.

[0217] Step 503: Calculate the third lateral distance P of the target vehicle. d The third lateral velocity V of the target vehicle d Multiply by the product to obtain the third prediction result.

[0218] Specifically, if the target vehicle crosses the line, the third lateral distance P is calculated. d The third lateral velocity V of the target vehicle d The product is used to obtain the third prediction result. The third prediction result can be used to determine whether the target vehicle is close to the lane of the vehicle. If the third prediction result is greater than zero, it means that the target vehicle is on the left side of the lane and is not close to the lane of the vehicle. If it is less than zero, it means that the target vehicle is on the right side of the lane and is close to the lane of the vehicle.

[0219] Step 504: If the third prediction result is less than zero, then determine whether the pressure line parameter is greater than the pressure line parameter determination threshold.

[0220] The pressure parameters include: pressure amount and pressure amount change rate.

[0221] The threshold values ​​for determining the pressure line parameters include: the pressure line amount threshold (CrossValue) and the pressure line amount change rate threshold (CrossRate).

[0222] Determining whether the creasing parameter is greater than the creasing parameter determination threshold includes:

[0223] Obtain the line crossing amount of the target vehicle. If the line crossing amount is greater than the line crossing amount threshold CrossValue, determine whether the line crossing amount change rate is greater than the line crossing amount change rate threshold CrossRate. If so, proceed to step 505.

[0224]

[0225]

[0226] Wherein, TTC is the longitudinal time interval, which can be calculated as the quotient of the distance from the rear of the target vehicle to the front of the own vehicle and the speed of the target vehicle relative to the own vehicle. Different thresholds are set for TTC, and TTC is divided into different piecewise functions through the thresholds. For different thresholds of TTC, different CrossValues ​​and CrossRates are obtained respectively. The first threshold is less than the second threshold, and the second threshold is less than the third threshold.

[0227] TTC = Distance from the rear of the target vehicle to the front of the vehicle / Speed ​​of the target vehicle relative to the vehicle

[0228] Step 505: Obtain the third lateral velocity V of the target vehicle. d .

[0229] Step 506: Determine whether the third lateral velocity is greater than the velocity determination parameter Vcutin. If the third lateral velocity is greater than the velocity determination parameter Vcutin, proceed to step 507. If it is less than Vcutin, proceed to step 508.

[0230]

[0231] Wherein, TTC is the longitudinal time distance, which can be calculated as the quotient of the distance from the rear of the target vehicle to the front of the own vehicle and the speed of the target vehicle relative to the own vehicle. Different thresholds are set for TTC, and TTC is divided into different piecewise functions through the thresholds. For different thresholds of TTC, different speed determination parameters Vcutin are obtained respectively. The first threshold is less than the second threshold, and the second threshold is less than the third threshold.

[0232] Step 507: Determine that the target vehicle has the intention to cut in.

[0233] Step 508: Determine whether the target vehicle has the intention to cut in based on the amount of line pressure and the heading angle of the target vehicle.

[0234] In practical applications, determining whether a target vehicle has an intention to cut in based on the amount of line pressure and the heading angle of the target vehicle includes: determining whether the amount of line pressure is greater than 0.5; if so, and the heading angle delta_heading is greater than or equal to a first angle threshold, then it is determined that the target vehicle has an intention to cut in.

[0235] In this embodiment, by judging the lateral and longitudinal distances of the target vehicle, the movement of the target vehicle can be detected in a timely manner, and the timing of the target vehicle's entry can be accurately determined, thereby improving the safety of autonomous driving.

[0236] In passing Figure 3 , 4 5. If a target vehicle is deemed to have the intention to cut in, some target vehicles may be identified as having this intention, but in reality, some may simply be continuously crossing the line without actually cutting into the lane. Therefore, to prevent misjudgment of a target vehicle's intention to cut in, after individually determining that a target vehicle has the intention to cut in, further steps can be taken... Figure 3 , 4 The judgment process for step 5 is corrected to prevent misjudgment; see [link / reference]. Figure 6 This is a flowchart of a method for handling misjudgment of a target vehicle's cutting intention according to the present invention, specifically including:

[0237] Step 601: Obtain the fourth lateral distance P between the target vehicle and the lane. d And the heading angle of the target vehicle's travel direction relative to the travel direction of the lane where the autonomous vehicle is located.

[0238] Step 602: If the fourth lateral distance and the heading angle have the same sign, then determine whether the target vehicle has crossed the line.

[0239] The heading angle can be positive or negative. A positive heading angle indicates that the target vehicle has not turned its head, while a negative heading angle indicates that the target vehicle has turned its head, and further judgment is needed to determine whether it intends to cut in.

[0240] Step 603: If there is no line overlap, calculate the fourth lateral distance P. d The fourth lateral velocity V of the target vehicle d Multiply by the product to obtain the fourth prediction result; or determine the fourth lateral velocity V. d Is it less than or equal to the second lateral velocity threshold?

[0241] Step 604: If the fourth prediction result is greater than or equal to zero; or the fourth lateral velocity is less than or equal to the third lateral velocity threshold, then it is determined that the target vehicle does not have the intention to cut in.

[0242] In this embodiment, a secondary judgment of the target vehicle's intention to enter is made to prevent misjudgment.

[0243] For target vehicles that do not exhibit any intention to engage, to ensure the safety of autonomous driving, this invention proposes further determining the intermediate states of the target vehicle, thereby monitoring its state in real time and predicting the intentions of potential engage vehicles. (See [link to relevant documentation]). Figure 7 This invention illustrates a method for predicting the entry intention of a target vehicle, specifically including:

[0244] Step 701: Obtain the fifth lateral distance P between the target vehicle and the lane. d .

[0245] Step 702: Determine whether the fifth lateral distance is greater than the first lane width threshold.

[0246] Step 703: If it is greater than the lane width threshold, then calculate the fifth lateral distance P. d The fifth lateral velocity V of the target vehicle d Multiply by the product to obtain the fifth prediction result; or determine the fifth lateral velocity V. d Is it less than or equal to the fifth lateral velocity threshold?

[0247] Step 704: If the fifth prediction result is less than zero or the fifth lateral velocity V d If the speed is less than the fifth lateral velocity threshold, then the fifth lateral distance P between the target vehicle and the lane is calculated. d The sixth prediction result is obtained by multiplying the heading angle of the target vehicle's travel direction relative to the travel direction of the autonomous vehicle in the lane.

[0248] Specifically, determine whether the fifth lateral distance and the heading angle have opposite signs. If they have opposite signs, the sixth prediction result is less than zero; if they do not have opposite signs, the sixth prediction result is greater than zero.

[0249] Step 705: If the sixth prediction result is less than zero, or the heading angle is less than the first angle threshold, then calculate the third distance dis2line between the target vehicle and the lane where the autonomous vehicle is located.

[0250] Specifically, calculate the third distance between the target vehicle and the center line of the lane where the autonomous vehicle is located, or calculate the third distance between the target vehicle and the boundary line of the lane where the autonomous vehicle is located.

[0251] dis2line = |p d |-|0.5×D×cos(delta_heading)+0.5×L×sin(delta_heading)|

[0252] Where D is the target vehicle width, L is the target vehicle length, and delta_heading is the heading angle.

[0253] Step 706: If the third distance dis2line is less than the distance threshold parameter D dis2line Or the fifth lateral distance P d If the distance is less than the second lane width threshold, then the longitudinal time distance of the target vehicle is obtained.

[0254] If the center line of the lane is used as the reference line, then D dis2line The following formula is used for calculation:

[0255]

[0256] Where R is the lane width.

[0257] If the lane boundary line is used as the reference line, then D dis2line The following formula is used for calculation:

[0258]

[0259] Step 707: If the longitudinal time distance is within the set range, then the target vehicle is determined to be in an intermediate state of the cutting intention.

[0260] In this embodiment, for target vehicles that do not have the intention to enter, the intermediate state of the target vehicle will be further determined, and target vehicles that may have the intention to enter will be screened out, thereby ensuring the safety of autonomous driving.

[0261] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to this application.

[0262] Based on the description of the above method embodiments, the present invention also provides corresponding device embodiments to implement the content described in the above method embodiments.

[0263] Reference Figure 8The diagram illustrates a structural schematic of a target vehicle cutting intention prediction device according to an embodiment of the present invention, the device comprising:

[0264] The motion state module 801 is used to acquire different motion states of the target vehicle under different scenarios.

[0265] The prediction module 802 is used to predict the cutting intention of the target vehicle based on different judgment criteria according to the different motion states.

[0266] Furthermore, the motion state of the target vehicle includes any of the following: the target vehicle deviating from the lane, the target vehicle merging into the lane, the target vehicle crossing the lane line, or the target vehicle approaching the lane where the autonomous vehicle is located but not entering the lane where the autonomous vehicle is located.

[0267] Furthermore, the prediction module includes:

[0268] The first prediction intent unit is used to predict the entry intent of the target vehicle based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located and the longitudinal distance between the target vehicle and the autonomous vehicle. The lane includes: lane center line or lane boundary line, or the entry intent of the target vehicle is predicted based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located.

[0269] Furthermore, the first predictive intent unit includes:

[0270] The first lateral submodule is used to obtain the first lateral distance P between the target vehicle and the lane. d ;

[0271] The first judgment submodule is used to determine whether the first lateral distance is less than the first distance judgment parameter D. offset ;

[0272] The first prediction submodule is used to determine if the distance is less than the first distance judgment parameter D. offset Then calculate the first lateral distance P. d With the first lateral velocity V of the target vehicle d Multiply by the product to obtain the first prediction result;

[0273] The first calculation submodule is used to calculate the first distance from the target vehicle to the autonomous vehicle if the first prediction result is less than zero.

[0274] The first longitudinal submodule is used to obtain the first longitudinal distance P between the target vehicle and the autonomous vehicle if the first distance is less than or equal to a set first distance threshold. s ;

[0275] The first cutting-in submodule is used to determine that the target vehicle has a cutting-in intention if the first longitudinal distance is greater than a set longitudinal distance.

[0276] Furthermore, the first predictive intent unit includes:

[0277] The second judgment submodule is used to determine whether the lane where the target vehicle is located intersects with the lane where the autonomous vehicle is located.

[0278] The second calculation submodule is used to calculate the second distance from the target vehicle to the autonomous vehicle if there is an intersection.

[0279] The second longitudinal submodule is used to obtain the second longitudinal distance P between the target vehicle and the autonomous vehicle if the second distance is less than or equal to a set second distance threshold. s ;

[0280] The second lateral submodule is used to obtain the second lateral distance P between the target vehicle and the lane where the autonomous vehicle is located if the second longitudinal distance is greater than a set longitudinal distance. d ;

[0281] The second prediction submodule is used to determine if the second lateral distance is less than the second distance determination parameter D. offset Then calculate the second lateral distance P. d With the second lateral velocity V of the target vehicle d Multiply by the product to obtain the second prediction result;

[0282] The third judgment submodule is used to determine the second lateral velocity V of the target vehicle if the second prediction result is less than zero. d Is it greater than the second lateral velocity threshold?

[0283] The second cut-in submodule is used to determine that the target vehicle has a cut-in intention if the speed is greater than the second lateral speed threshold.

[0284] Furthermore, the first predictive intent unit includes:

[0285] The fourth judgment submodule is used to determine whether the target vehicle has crossed the line;

[0286] The third calculation submodule is used to calculate the third lateral distance P of the target vehicle if the condition is met. d The third lateral velocity V of the target vehicle d Multiply by the product to obtain the third prediction result;

[0287] The fifth judgment submodule is used to determine whether the pressure line parameter is greater than the pressure line parameter judgment threshold if the third prediction result is less than zero. The pressure line parameter includes: pressure line amount and pressure line amount change rate.

[0288] The third lateral submodule is used to, if so, obtain the third lateral velocity V of the target vehicle. d ;

[0289] The third cutting-in submodule is used to determine whether the target vehicle has a cutting-in intention if the third lateral speed is greater than the speed determination parameter Vcutin; and to determine whether the target vehicle has a cutting-in intention based on the line overlap amount and the heading angle of the target vehicle if the third lateral speed is less than Vcutin.

[0290] Furthermore, the distance determination parameter D is calculated in the following manner. offset :

[0291] Calculate the heading angle of the target vehicle's travel direction relative to the travel direction of the lane where the autonomous vehicle is located;

[0292] The vehicle length, vehicle width, lane width of the lane where the target vehicle is located, first lateral speed and lateral time distance of the target vehicle are obtained;

[0293] The distance determination parameters can be calculated based on the heading angle, vehicle length, vehicle width, lane width, first lateral speed, and lateral time distance, or the distance determination parameters can be calculated based on the heading angle, vehicle length, vehicle width, first lateral speed, and lateral time distance.

[0294] Furthermore, the device also includes:

[0295] The first acquisition module is used to acquire the fourth lateral distance P between the target vehicle and the lane. d And the heading angle of the target vehicle's travel direction relative to the travel direction of the lane where the autonomous vehicle is located;

[0296] The first judgment module is used to determine whether the target vehicle has crossed the line if the fourth lateral distance and the heading angle have the same sign.

[0297] The first calculation module is used to calculate the fourth lateral distance P if there is no line pressing. d The fourth lateral velocity V of the target vehicle d Multiply by the product to obtain the fourth prediction result; or determine the fourth lateral velocity V. d Is it less than or equal to the second lateral velocity threshold?

[0298] The output module is used to determine that the target vehicle does not have the intention to cut in if the fourth prediction result is greater than or equal to zero, or the fourth lateral velocity is less than or equal to the third lateral velocity threshold.

[0299] Furthermore, the device also includes:

[0300] The second acquisition module is used to acquire the fifth lateral distance P between the target vehicle and the lane. d ;

[0301] The second judgment module is used to determine whether the fifth lateral distance is greater than the first lane width threshold.

[0302] The second calculation module is used to calculate the fifth lateral distance P if it is greater than the lane width threshold. d The fifth lateral velocity V of the target vehicle d Multiply by the product to obtain the fifth prediction result; or determine the fifth lateral velocity V. d Is it less than or equal to the fifth lateral velocity threshold?

[0303] The deviation module is used if the fifth prediction result is less than zero, or if the fifth lateral velocity V... d If the speed is less than the fifth lateral velocity threshold, then the fifth lateral distance P between the target vehicle and the lane is calculated. d The sixth prediction result is obtained by multiplying the target vehicle's travel direction by the heading angle relative to the autonomous vehicle's lane travel direction.

[0304] The centerline distance module is used to calculate the third distance dis2line between the target vehicle and the lane where the autonomous vehicle is located if the sixth prediction result is less than zero, or if the heading angle is less than the first angle threshold.

[0305] The longitudinal time-distance module is used to determine if the third distance dis2line is less than the distance threshold parameter Ddis2line, or if the fifth lateral distance P... d If the distance is less than the second lane width threshold, then the longitudinal time distance of the target vehicle is obtained;

[0306] The status module is used to determine that the target vehicle is in an intermediate state of the cutting intention if the longitudinal time distance is within a set range.

[0307] In this embodiment, firstly, the different motion states of the target vehicle in different scenarios are obtained. Based on the different motion states of the target vehicle, different judgment criteria are used to predict the entry intention of the target vehicle. Thus, without the need to collect training samples, it is possible to predict the multiple entry intentions of the vehicle more quickly and accurately, thereby improving the safety of autonomous driving.

[0308] Secondly, by judging the lateral and longitudinal distances of the target vehicle, the movement of the target vehicle can be detected in a timely manner, and the timing of the target vehicle's entry can be accurately determined, thereby achieving accurate judgment of the entry intention and improving the safety of autonomous driving.

[0309] Furthermore, the target vehicles whose entry intentions are predicted are further screened to determine whether there will be any misjudgment of the entry intentions, thereby achieving accurate determination of the entry intentions and improving the safety of autonomous driving.

[0310] The above-described apparatus embodiments are basically similar to the method embodiments, so the description is relatively simple. For relevant details, please refer to the description of the method embodiments shown.

[0311] It will be readily apparent to those skilled in the art that any combination of the above embodiments is feasible, and therefore any combination of the above embodiments is an implementation scheme of the present invention. However, due to space limitations, this specification will not describe them in detail here.

[0312] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention.

[0313] The above embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention.

Claims

1. A method for predicting the entry intention of a target vehicle, characterized in that, include: Acquire different motion states of the target vehicle in different scenarios; Different judgment criteria are used to predict the entry intention of the target vehicle based on the different motion states; The step of predicting the entry intention of the target vehicle by using different judgment criteria according to different motion states includes: The cut-in intention of the target vehicle is predicted based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located and the longitudinal distance between the target vehicle and the autonomous vehicle. The lane includes: lane center line or lane boundary line, or the cut-in intention of the target vehicle is predicted based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located. The step of predicting the cutting intention of the target vehicle based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located and the longitudinal distance between the target vehicle and the autonomous vehicle includes: Obtain the first lateral distance between the target vehicle and the lane; Determine whether the first lateral distance is less than the first distance determination parameter; If it is less than the first distance determination parameter, then the product of the first lateral distance and the first lateral velocity of the target vehicle is calculated to obtain the first prediction result; If the first prediction result is less than zero, then calculate the first distance from the target vehicle to the autonomous vehicle; If the first distance is less than or equal to a set first distance threshold, then the first longitudinal distance between the target vehicle and the autonomous vehicle is obtained; If the first longitudinal distance is greater than the set longitudinal distance threshold, it is determined that the target vehicle has the intention to cut in.

2. The prediction method according to claim 1, characterized in that, The different motion states of the target vehicle include any of the following: the target vehicle deviating from the lane, the target vehicle merging into the lane, the target vehicle crossing the line, and the target vehicle moving close to the lane where the autonomous vehicle is located but not entering the lane where the autonomous vehicle is located.

3. The prediction method according to claim 1, characterized in that, Alternatively, the step of predicting the cut-in intention of the target vehicle based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located and the longitudinal distance between the target vehicle and the autonomous vehicle includes: Determine whether the lane where the target vehicle is located intersects with the lane where the autonomous vehicle is located; If an intersection exists, then calculate the second distance from the target vehicle to the autonomous vehicle; If the second distance is less than or equal to the set second distance threshold, then the second longitudinal distance between the target vehicle and the autonomous vehicle is obtained; If the second longitudinal distance is greater than the set longitudinal distance threshold, then the second lateral distance between the target vehicle and the lane where the autonomous vehicle is located is obtained; If the second lateral distance is less than the second distance determination parameter, calculate the product of the second lateral distance and the second lateral velocity of the target vehicle to obtain the second prediction result; If the second prediction result is less than zero, then it is determined whether the second lateral velocity of the target vehicle is greater than the set second lateral velocity threshold. If so, it is determined that the target vehicle has the intention to cut in.

4. The prediction method according to claim 1, characterized in that, Alternatively, the step of predicting the cut-in intention of the target vehicle based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located includes: Determine whether the target vehicle has crossed the line; If so, calculate the product of the third lateral distance of the target vehicle and the third lateral velocity of the target vehicle to obtain the third prediction result; If the third prediction result is less than zero, then it is determined whether the pressure line parameter is greater than the pressure line parameter judgment threshold, wherein the pressure line parameter includes: pressure line amount and pressure line amount change rate; If so, obtain the third lateral velocity of the target vehicle; If the third lateral velocity is greater than the velocity determination parameter, then it is determined that the target vehicle has the intention to cut in. If the third lateral speed is less than the speed determination parameter, then the presence or absence of the target vehicle's cutting intention is determined based on the line overlap and the target vehicle's heading angle.

5. The prediction method according to claim 1 or 3, characterized in that, The distance determination parameters are calculated using the following method: Calculate the heading angle of the target vehicle's travel direction relative to the travel direction of the lane where the autonomous vehicle is located; The vehicle length, vehicle width, lane width of the lane where the target vehicle is located, first lateral speed and lateral time distance of the target vehicle are obtained; The distance determination parameters can be calculated based on the heading angle, vehicle length, vehicle width, lane width, first lateral speed, and lateral time distance, or the distance determination parameters can be calculated based on the heading angle, vehicle length, vehicle width, first lateral speed, and lateral time distance.

6. The prediction method according to claim 1, 3, or 4, characterized in that, The method further includes: Obtain the fourth lateral distance between the target vehicle and the lane, and the heading angle of the target vehicle's driving direction relative to the driving direction of the lane where the autonomous vehicle is located; If the fourth lateral distance and the heading angle have the same sign, then it is determined whether the target vehicle has crossed the line; If there is no line crossing, calculate the product of the fourth lateral distance and the fourth lateral velocity of the target vehicle to obtain the fourth prediction result; or determine whether the fourth lateral velocity is less than or equal to the second lateral velocity threshold. If the fourth prediction result is greater than or equal to zero; or if the fourth lateral velocity is less than or equal to the third lateral velocity threshold, then it is determined that the target vehicle does not have the intention to cut in.

7. The prediction method according to claim 6, characterized in that, The method further includes: Obtain the fifth lateral distance between the target vehicle and the lane; Determine whether the fifth lateral distance is greater than the first lane width threshold; If the distance is greater than the lane width threshold, the product of the fifth lateral distance and the fifth lateral velocity of the target vehicle is calculated to obtain the fifth prediction result; or it is determined whether the fifth lateral velocity is less than or equal to the fifth lateral velocity threshold. If the fifth prediction result is less than zero, or the fifth lateral velocity is less than the fifth lateral velocity threshold, then the product of the fifth lateral distance between the target vehicle and the lane and the heading angle of the target vehicle's travel direction relative to the travel direction of the lane where the autonomous vehicle is located is calculated to obtain the sixth prediction result. If the sixth prediction result is less than zero, or the heading angle is less than the first angle threshold, then the third distance between the target vehicle and the lane where the autonomous vehicle is located is calculated. If the third distance is less than the distance threshold parameter, or the fifth lateral distance is less than the second lane width threshold, then the longitudinal time distance of the target vehicle is obtained; If the longitudinal time interval is within the set range, then the target vehicle is determined to be in an intermediate state of the intended entry.

8. A target vehicle cutting-in intention prediction device, characterized in that, include: The motion state module is used to acquire different motion states of the target vehicle in different scenarios; The prediction module is used to predict the cutting intention of the target vehicle based on different judgment criteria according to the different motion states. The prediction module includes: The first prediction intent unit is used to predict the cutting intent of the target vehicle based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located and the longitudinal distance between the target vehicle and the autonomous vehicle. The lane includes: lane center line or lane boundary line, or the cutting intent of the target vehicle is predicted based on the lateral distance between the target vehicle and the lane where the autonomous vehicle is located. The first predictive intent unit includes: The first lateral submodule is used to obtain the first lateral distance between the target vehicle and the lane; The first judgment submodule is used to determine whether the first horizontal distance is less than the first distance judgment parameter; The first prediction submodule is used to calculate the product of the first lateral distance and the first lateral velocity of the target vehicle if the distance is less than the first distance determination parameter, and obtain the first prediction result. The first calculation submodule is used to calculate the first distance from the target vehicle to the autonomous vehicle if the first prediction result is less than zero. The first longitudinal submodule is used to obtain the first longitudinal distance between the target vehicle and the autonomous vehicle if the first distance is less than or equal to a set first distance threshold. The first cutting-in submodule is used to determine that the target vehicle has a cutting-in intention if the first longitudinal distance is greater than a set longitudinal distance threshold.

9. The prediction device according to claim 8, characterized in that, The motion state of the target vehicle includes any of the following: the target vehicle deviates from the lane, the target vehicle merges into the lane, the target vehicle crosses the line, or the target vehicle is driving close to the lane where the autonomous vehicle is located but has not entered the lane where the autonomous vehicle is located.

10. The prediction device according to claim 8, characterized in that, Alternatively, the first predictive intent unit includes: The second judgment submodule is used to determine whether the lane where the target vehicle is located intersects with the lane where the autonomous vehicle is located. The second calculation submodule is used to calculate the second distance from the target vehicle to the autonomous vehicle if there is an intersection. The second longitudinal submodule is used to obtain the second longitudinal distance between the target vehicle and the autonomous vehicle if the second distance is less than or equal to a set second distance threshold. The second lateral submodule is used to obtain the second lateral distance between the target vehicle and the lane where the autonomous vehicle is located if the second longitudinal distance is greater than a set longitudinal distance threshold. The second prediction submodule is used to calculate the product of the second lateral distance and the second lateral velocity of the target vehicle if the second lateral distance is less than the second distance determination parameter, and obtain the second prediction result. The third judgment submodule is used to determine whether the second lateral velocity of the target vehicle is greater than the set second lateral velocity threshold if the second prediction result is less than zero. The second cut-in submodule is used to determine that the target vehicle has a cut-in intention if the speed is greater than the second lateral speed threshold.

11. The prediction device according to claim 8, characterized in that, Alternatively, the first predictive intent unit includes: The fourth judgment submodule is used to determine whether the target vehicle has crossed the line; The third calculation submodule is used to calculate the product of the third lateral distance of the target vehicle and the third lateral velocity of the target vehicle if the condition is met, and obtain the third prediction result. The fifth judgment submodule is used to determine whether the pressure line parameter is greater than the pressure line parameter judgment threshold if the third prediction result is less than zero. The pressure line parameter includes: pressure line amount and pressure line amount change rate. The third lateral submodule is used to obtain the third lateral velocity of the target vehicle if the condition is met. The third entry submodule is used to determine whether the target vehicle has an entry intention if the third lateral speed is greater than the speed determination parameter; and to determine whether the target vehicle has an entry intention based on the line overlap amount and the heading angle of the target vehicle if the third lateral speed is less than the speed determination parameter.

12. The prediction device according to claim 8 or 10, characterized in that, The distance determination parameters are calculated using the following method: Calculate the heading angle of the target vehicle's travel direction relative to the travel direction of the lane where the autonomous vehicle is located; The vehicle length, vehicle width, lane width of the lane where the target vehicle is located, first lateral speed and lateral time distance of the target vehicle are obtained; The distance determination parameters can be calculated based on the heading angle, vehicle length, vehicle width, lane width, first lateral speed, and lateral time distance, or the distance determination parameters can be calculated based on the heading angle, vehicle length, vehicle width, first lateral speed, and lateral time distance.

13. The prediction device according to claim 8, 10 or 11, characterized in that, The device further includes: The first acquisition module is used to acquire the fourth lateral distance between the target vehicle and the lane and the heading angle of the target vehicle's driving direction relative to the driving direction of the lane where the autonomous vehicle is located; The first judgment module is used to determine whether the target vehicle has crossed the line if the fourth lateral distance and the heading angle have opposite signs. The first calculation module is used to obtain a fourth prediction result by calculating the product of the fourth lateral distance and the fourth lateral speed of the target vehicle if there is no line crossing; or to determine whether the fourth lateral speed is less than or equal to the second lateral speed threshold. The output module is used to determine that the target vehicle does not have the intention to cut in if the fourth prediction result is greater than or equal to zero, or the fourth lateral velocity is less than or equal to the third lateral velocity threshold.

14. The prediction device according to claim 13, characterized in that, The device further includes: The second acquisition module is used to acquire the fifth lateral distance between the target vehicle and the lane; The second judgment module is used to determine whether the fifth lateral distance is greater than the first lane width threshold. The second calculation module is used to calculate the product of the fifth lateral distance and the fifth lateral speed of the target vehicle if the distance is greater than the lane width threshold, and obtain the fifth prediction result; or to determine whether the fifth lateral speed is less than or equal to the fifth lateral speed threshold. The deviation module is used to calculate the product of the fifth lateral distance between the target vehicle and the lane and the heading angle of the target vehicle's driving direction relative to the driving direction of the autonomous vehicle in the lane if the fifth prediction result is less than zero or the fifth lateral velocity is less than the fifth lateral velocity threshold, and obtain the sixth prediction result. The centerline distance module is used to calculate the third distance between the target vehicle and the lane where the autonomous vehicle is located if the sixth prediction result is less than zero, or if the heading angle is less than the first angle threshold. The longitudinal time distance module is used to obtain the longitudinal time distance of the target vehicle if the third distance is less than a distance threshold parameter or the fifth lateral distance is less than a second lane width threshold. The status module is used to determine that the target vehicle is in an intermediate state of the cutting intention if the longitudinal time distance is within a set range.

Images