Collision risk intention recognition and avoidance method and device, electronic equipment and storage medium

Abstract

The invention discloses a collision risk intention recognition and avoidance method and device, electronic equipment and a storage medium, and the method comprises the steps: determining a collision detection boundary and a target obstacle, the collision detection boundary being determined according to route information and map lane data; calculating collision information of the target obstacle reaching the collision detection boundary according to the target obstacle; and according to the collision information and the current state of the vehicle, calculating an expected acceleration value of the vehicle so as to enable the vehicle to execute an avoidance action in decision planning. According to the invention, on one hand, the intention of small and stable state quantity change is timely and effectively identified, so that the reasonability of obstacle decision making is ensured, and on the other hand, the collision risk of longitudinal planning is reduced.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a method, device, electronic device, and storage medium for identifying and avoiding collision risk intentions. Background Art

[0002] The current mainstream autonomous driving solutions divide the entire algorithm system into multiple relatively independent modules such as positioning, perception, prediction, decision-making and planning. From the perspective of the decision-making and planning module, it is necessary to inform the prediction module of possible changes in the surrounding environment status in the future and generate a safe vehicle driving trajectory in the decision-making and planning module.

[0003] Specifically, the mainstream prediction algorithm is based on learning, and its output is seriously affected by the training data. The prediction of cutin scenarios has problems such as lag and inaccuracy, which makes the final planned vehicle driving trajectory unsafe. Summary of the invention

[0004] The embodiments of the present application provide a device, electronic device, and storage medium for identifying and avoiding collision risk intentions, and timely and effectively identify intentions with small but stable changes in state quantities, thereby ensuring the rationality of obstacle decisions and reducing the collision risk of longitudinal planning.

[0005] The present application embodiment adopts the following technical solutions:

[0006] In a first aspect, an embodiment of the present application provides a method for identifying and avoiding collision risk intentions, wherein the method includes:

[0007] Determining a collision detection boundary and a target obstacle, wherein the collision detection boundary is determined based on route information and map lane data;

[0008] Calculating collision information of the target obstacle reaching the collision detection boundary according to the target obstacle;

[0009] According to the collision information and the current state of the ego vehicle, the expected acceleration value of the ego vehicle is calculated so as to allow the ego vehicle to perform evasive action in decision planning.

[0010] In some embodiments, calculating, according to the target obstacle, collision information of the target obstacle reaching the collision detection boundary includes:

[0011] According to the motion state information of the target obstacle in the Frenet system, a virtual obstacle trajectory of the target obstacle in a future period of time is obtained;

[0012] Considering the geometric shape of the obstacle, calculating whether the obstacle virtual trajectory intersects with the collision detection boundary;

[0013] If they intersect, the position information of the first intersection point is calculated, where the first intersection point is the position point that first reaches the collision detection boundary;

[0014] According to the position information of the first intersection point, the intersection time T is calculated. c ;

[0015] If they do not intersect, it is considered that the target obstacle poses no risk to the vehicle.

[0016] In some embodiments, the method further comprises:

[0017] Based on the structured road, the Frenet reference system is established with the lane centerline as the X-axis and the vertical lane centerline as the Y-axis;

[0018] According to the Frenet reference system, the positions of the corner points of the obstacles sensed by the upstream module are projected in the Frenet reference system, and the corner points closest to the center line of the lane are retained to calculate the collision information, wherein the upstream module includes a prediction module and / or a fusion perception module;

[0019] The perceived obstacle position and obstacle speed are projected in the Frenet reference system to obtain the position and speed state of the corresponding obstacle in the Frenet reference system.

[0020] In some embodiments, the calculation of the expected acceleration value of the ego vehicle based on the collision information and the current state of the ego vehicle so as to allow the ego vehicle to perform an evasive action in decision planning includes:

[0021] Project the collision point in the collision information onto the lateral route to obtain the lateral route length S corresponding to the collision point c , the transverse route length S c is the length between the ego vehicle and the collision point;

[0022] Based on the geometry of the vehicle and / or the obstacle and the preset safety distance, the overtaking point S is obtained by adding the expected overtaking distance to the lateral route length corresponding to the collision point. o Subtract the expected courtesy distance to get the courtesy point S y .

[0023] In some embodiments, the acceleration value expected by the vehicle includes: an expected acceleration for overtaking decision, and the method further includes:

[0024] The current position of the vehicle is assumed to be an initial value of 0, and the current speed state of the vehicle is V ego As the initial state, the future driving trajectory of the ego vehicle is modeled as a uniform acceleration model;

[0025] The upper limit of acceleration when limiting overtaking decision is a b , and the overtaking decision of the ego vehicle is equivalent to the decision at time T c The vehicle's current position S ego Passing overtaking point S o , the corresponding problem can be modeled as the first vehicle motion problem;

[0026] If the first vehicle motion problem has a solution, there is no need to make a courtesy decision;

[0027] If the first vehicle motion problem has a solution, the expected acceleration for the overtaking decision is determined.

[0028] In some embodiments, the acceleration value expected by the vehicle includes: a courtesy decision expected acceleration, and the method further includes:

[0029] The current position of the vehicle is assumed to be an initial value of 0, and the current speed state of the vehicle is V ego As the initial state, the future driving trajectory of the ego vehicle is modeled as a uniform acceleration model;

[0030] At time T c The position of the vehicle is S y After that, the courtesy decision of the ego vehicle must satisfy the courtesy speed that is always greater than 0, and the corresponding problem can be modeled as the second vehicle motion problem;

[0031] If the second vehicle motion problem has a solution, the expected acceleration of the courtesy decision is determined and the speed of the vehicle does not need to be uniformly decelerated to 0;

[0032] If the second vehicle motion problem has no solution, the vehicle needs to first decelerate uniformly to 0, then stop and yield, and the resulting problem can be modeled as the third vehicle motion problem.

[0033] In some embodiments, the method further comprises:

[0034] The desired acceleration value of the ego vehicle is used as a penalty term in a dynamic programming search process based on the ST graph.

[0035] In a second aspect, an embodiment of the present application further provides an avoidance device based on collision risk intention recognition, wherein the device comprises:

[0036] A determination module, used to determine a collision detection boundary and a target obstacle, wherein the collision detection boundary is determined based on route information and map lane data;

[0037] A collision detection module, used for calculating collision information of the target obstacle reaching the collision detection boundary according to the target obstacle;

[0038] The expected acceleration calculation module is used to calculate the expected acceleration value of the ego vehicle according to the collision information and the current state of the ego vehicle, so as to allow the ego vehicle to perform evasive actions in decision planning.

[0039] In a third aspect, an embodiment of the present application further provides an electronic device, comprising: a processor; and a memory arranged to store computer executable instructions, wherein the executable instructions, when executed, cause the processor to perform the above method.

[0040] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple application programs, the electronic device executes the above method.

[0041] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects: first, determine the collision detection boundary and the target obstacle, and the collision detection boundary is determined based on the route information and the map lane data. Then, calculate the collision information of the target obstacle reaching the collision detection boundary based on the target obstacle, and finally calculate the expected acceleration value of the vehicle based on the collision information and the current state of the vehicle, so that the vehicle can perform evasive actions in decision-making planning. The above method is a solution based on collision risk intention recognition and application in longitudinal speed planning, and is mainly for the optimization of the longitudinal planning module under the horizontal and longitudinal separation planning framework. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0043] Figure 1 This is a schematic diagram of the collision detection boundary in an embodiment of the present application;

[0044] Figure 2 This is a schematic diagram of collision information in an embodiment of the present application;

[0045] Figure 3 This is a schematic diagram of decision-making planning in the embodiment of this application;

[0046] Figure 4 A flowchart of a method for identifying and avoiding collision risk intention in an embodiment of the present application;

[0047] Figure 5 This is a schematic diagram of the structure of an avoidance device based on collision risk intention recognition in an embodiment of the present application;

[0048] Figure 6 This is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0050] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0051] The present application embodiment provides a method for identifying and avoiding collision risk intentions, such as Figure 4 As shown, a schematic flow chart of a method for identifying and avoiding collision risk intention in an embodiment of the present application is provided, and the method at least includes the following steps S410 to S430:

[0052] Step S410, determining a collision detection boundary and a target obstacle, wherein the collision detection boundary is determined based on route information and map lane data.

[0053] "Obstacles" are obtained through the upstream perception module, and are screened out based on the loaded map data, perception data, and the vehicle's driving path. "Collision detection boundaries" are generated by coupling the vehicle's lateral path (route) with the lane boundary. It can be understood that the lane boundary is the adjacent lane line as the boundary.

[0054] For "obstacles", considering the high uncertainty of future changes in the surrounding environment, the target obstacle cannot be too far away from the collision detection boundary. At the same time, considering that traffic regulations do not allow continuous lane changes, the position range is limited to within a lane width from the collision detection boundary in the embodiment of the present application, and static obstacles are filtered out, such as Figure 1 In addition, some other filtering conditions can be added to avoid interference from irrelevant obstacles. For example, filtering conditions can be added according to the characteristics of the vehicle or engineering experience, so as to filter out unnecessary obstacles for a specific scene.

[0055] like Figure 1 As shown, the left and right collision detection boundaries can be generated according to the path information of the upstream module and the map lane data. It should be noted that the map lane data can preferably use an HD map.

[0056] Please continue to refer to Figure 1Considering the driving habits of drivers, they often use the lane line of their own vehicle as an important judgment sign for whether the vehicles in the adjacent lanes will invade their own lanes. However, it is unsafe to rely solely on the lane line as the collision detection boundary for the active line-crossing behavior of the own vehicle due to other purposes such as the need to bypass obstacles. Therefore, it is necessary to couple the lateral path (route) of the own vehicle with the lane boundary to generate the final collision detection boundary. Specifically, taking the right collision detection boundary as an example to illustrate this process, first calculate the intersection of the right boundary of the lateral path (route) and the right boundary of the current lane. These intersections divide the right boundary of the lateral path (route) and the right boundary of the current lane into multiple intervals. The intervals that form the rightmost boundary are spliced ​​together to form the final right collision detection boundary. The left boundary is processed in the same way to obtain Figure 1 The collision detection boundaries are shown.

[0057] Step S420: calculating collision information of the target obstacle reaching the collision detection boundary according to the target obstacle.

[0058] According to the motion state information of the obstacle, assuming that the obstacle moves according to the uniform acceleration model, the virtual trajectory within a period of time in the future is calculated in the Frenet reference system, and the intersection information of the virtual trajectory and the collision detection boundary is the collision information. That is, for the selected target obstacle, the position and time when it first reaches the collision boundary are calculated.

[0059] like Figure 2 As shown, V S and V L They represent the speed in different directions based on the center of the lane line (Frenet), and also include a virtual trajectory and a collision point on the collision detection boundary based on the virtual trajectory.

[0060] Step S430, calculating the expected acceleration value of the ego vehicle according to the collision information and the current state of the ego vehicle, so as to allow the ego vehicle to perform an evasive action in the decision-making planning.

[0061] The collision information includes information such as the collision position and collision time. The current state of the vehicle includes: the speed, acceleration and position information of the vehicle.

[0062] Based on the above information, the expected acceleration value of the ego vehicle is calculated, that is, the longitudinal motion of the ego vehicle is modeled to convert the collision information into the expected acceleration of the ego vehicle to guide the future avoidance action of the ego vehicle in decision planning.

[0063] The above method determines the collision detection boundary and the target obstacle, and calculates the collision information of the target obstacle reaching the collision detection boundary based on the target obstacle, thereby ensuring the rationality of the obstacle decision. Different from the obstacle prediction scheme based on learning in the related art, it is often affected by the noise of the training data and the potential constraints of the road information, and cannot timely and effectively identify the intention of small but stable changes in state quantity. In particular, the recognition of the stable lateral low-speed obstacle cutin (cut-in) scene is often delayed or even unable to give effective results. Through the above method, it can be guaranteed that the intention of small but stable changes in state quantity is identified in a timely and effective manner, thereby ensuring the rationality of the obstacle decision and reducing the collision risk of longitudinal planning.

[0064] In one embodiment of the present application, the calculation of collision information of the target obstacle reaching the collision detection boundary based on the target obstacle includes: obtaining a virtual obstacle trajectory of the target obstacle in a future period of time based on the motion state information of the target obstacle in the Frenet system; calculating whether the virtual obstacle trajectory intersects with the collision detection boundary considering the geometric shape of the obstacle; if so, calculating the position information of the first intersection point, the first intersection point being the position point of the first arrival at the collision detection boundary; calculating the intersection time based on the position information of the first intersection point; if not, it is considered that the target obstacle poses no risk to the driving of the self-vehicle.

[0065] In the process of determining the collision information of the target obstacle reaching the collision detection boundary, it is necessary to calculate the virtual trajectory of the obstacle and calculate the intersection information. Specifically:

[0066] First, the virtual trajectory of the target obstacle in the future can be obtained by integrating the current position and speed information of the target obstacle;

[0067] Secondly, calculate whether the virtual trajectory of the target obstacle intersects with the collision detection boundary (determine the first intersection point and intersection time). If so, calculate the position information of the first intersection point. If not, it is considered that the obstacle poses no risk to the vehicle and does not need to be considered.

[0068] Finally, after obtaining the intersection position, the intersection time can be further calculated.

[0069] In one embodiment of the present application, the method further includes: establishing a Frenet reference system based on a structured road with the lane centerline as the X-axis and the vertical lane centerline as the Y-axis; according to the Frenet reference system, projecting the corner point positions of the obstacles perceived by the upstream module in the Frenet reference system, retaining the corner point closest to the lane centerline for calculating collision information, and the upstream module includes a prediction module and / or a fusion perception module; projecting the perceived obstacle position and obstacle speed in the Frenet reference system to obtain the position and speed state of the corresponding obstacle in the Frenet reference system.

[0070] In order to meet the requirements of calculating the projection state of obstacles, a new reference coordinate system needs to be created. Considering that on structured roads, vehicles usually drive according to the guidance of lane lines. Therefore, in the embodiment of the present application, the lane centerline is used as the X-axis and the vertical lane centerline is used as the Y-axis to establish a Frenet reference system. Based on the Frenet reference system, the positions of the corner points of the perceived obstacles are projected under this reference system, and the corner point closest to the lane centerline is retained as (P s , P l ); Project the perceived obstacle speed in this reference system, and obtain the position state and speed state as (V s , V l ). Among them, (P fs , P fl ) is the component of the corresponding state along the lane centerline direction, (P s , P l ) is the component in the direction perpendicular to the lane centerline in the corresponding state.

[0071] It is understandable that the perceived obstacle may be based on a regular quadrilateral or a convex polygon, and therefore the state of the nearest corner point of the obstacle may be used to calculate the state of the first arrival at the collision point taking into account the geometric shape of the obstacle.

[0072] Furthermore, when calculating the virtual trajectory of the obstacle, the virtual trajectory of the obstacle in the future period of time is obtained by integrating the current position and speed information of the obstacle. The corresponding calculation formula is:

[0073] P fs =P s +V s *t

[0074] P fl =P l +V l *t

[0075] Among them, (P fs , P fl ) is the future time interval t (P s, P l ) corresponds to the position in the Frenet reference frame.

[0076] Calculate the intersection information. Calculate whether the obstacle virtual trajectory intersects with the collision detection boundary. If so, calculate the position information P of the first intersection point. c (P cs , P cl ), if they do not intersect, it is considered that the obstacle poses no risk to the vehicle and does not need to be considered. After obtaining the intersection position, the intersection time T can be calculated according to the following formula: c :

[0077]

[0078] In one embodiment of the present application, the expected acceleration value of the ego vehicle is calculated based on the collision information and the current state of the ego vehicle so as to allow the ego vehicle to perform an evasive action in the decision-making planning, including: projecting the collision point in the collision information onto the lateral route to obtain the lateral route length S corresponding to the collision point c , the transverse route length S c is the length between the vehicle and the collision point; based on the geometry of the vehicle and / or the obstacle and the preset safety distance, the overtaking point S is obtained by adding the expected overtaking distance to the lateral route length corresponding to the collision point. o Or subtract the expected courtesy distance to get the courtesy point S y .

[0079] like Figure 3 As shown, the collision information is used to obtain the collision information and the current state of the vehicle, and the desired acceleration of the vehicle is calculated, and it is applied to the existing autonomous driving planning framework in the form of a cost.

[0080] like Figure 3 As shown in FIG. 1 , a new reference system is established based on the lateral path (route) planned by the ego vehicle, and the collision point, overtaking point and courtesy point are represented in this reference system. Specifically, the collision point in the collision information is projected onto the lateral path (route), and the lateral path (route) length S corresponding to the collision point is obtained. c Based on the geometry of the vehicle and / or obstacle, the preset safety distance, the lateral path length S at the corresponding collision point c Based on the expected overtaking distance, the overtaking point S is obtained. o Or subtract the expected courtesy distance to get the courtesy point S y .

[0081] like Figure 3As shown, based on the horizontal path (route), different positions of the path can be divided into different intervals according to the yield points and overtake points. The yield interval is the yield interval, and the overtake interval is the overtake interval. y , Overtaking point S o , collision projection S c Wherein, the collision projection S c It is obtained by projection mapping the collision point Pc in the Frenet reference system.

[0082] In one embodiment of the present application, the acceleration value expected by the vehicle includes: the expected acceleration of the overtaking decision, and the method further includes: the current position of the vehicle defaults to an initial value of 0, and the current speed state of the vehicle V ego As the initial state, the future driving trajectory of the vehicle is modeled as a uniform acceleration model; the upper limit of the acceleration when limiting the overtaking decision is a b , and the overtaking decision of the ego vehicle is equivalent to the decision at time T c The vehicle's current position S ego Passing overtaking point S o , the corresponding problem can be modeled as a first vehicle motion problem; if the first vehicle motion problem has a solution, there is no need to make a courtesy decision; if the first vehicle motion problem has a solution, the expected acceleration of the overtaking decision is determined.

[0083] Please continue to refer to Figure 3 , is the process of calculating the expected acceleration of overtaking decision:

[0084] 1) The current position of the vehicle is assumed to be 0, and the current speed state of the vehicle is V ego As the initial state, the future driving trajectory of the ego vehicle is modeled as a uniform acceleration model;

[0085] 2) At the same time, the upper limit of acceleration during overtaking decision is a b The overtaking decision of the ego vehicle is equivalent to c At time , the vehicle is at position S ego Passing overtaking point S o .

[0086] 3) Strategy based on expected acceleration of overtaking decision: that is, the main strategy is for the ego car to overtake the other car first. The ego car does not need a large acceleration, and the minimum acceleration that meets a preset condition is sufficient. The corresponding problem can be modeled as:

[0087] min a

[0088]

[0089] S ego ≥S o

[0090] 0≤a≤a b

[0091] a b The size of can reflect the aggressiveness of driving, which can be adjusted by a b If the above problem has a solution, the overtaking decision is executed without calculating other subsequent decisions.

[0092] In one embodiment of the present application, the acceleration value expected by the vehicle includes: the expected acceleration of the courtesy decision, and the method further includes: the current position of the vehicle defaults to an initial value of 0, and the current speed state of the vehicle V ego As the initial state, the future driving trajectory of the vehicle is modeled as a uniform acceleration model; at time T c The position of the vehicle is S y After that, the courtesy decision of the ego vehicle needs to satisfy the condition that the courtesy speed is always greater than 0, and the corresponding problem can be modeled as the second vehicle motion problem; if the second vehicle motion problem has a solution, the expected acceleration of the courtesy decision is determined and the speed of the ego vehicle does not need to be uniformly decelerated to 0; if the second vehicle motion problem has no solution, the ego vehicle needs to first be uniformly decelerated to 0, and then stop to yield, and the resulting problem can be modeled as the third vehicle motion problem.

[0093] Please continue to refer to Figure 3 , is the process of calculating the expected acceleration of the courtesy decision:

[0094] 1) The current position of the vehicle is assumed to be 0, and the current speed state of the vehicle is V ego As the initial state, the future driving trajectory of the ego vehicle is modeled as a uniform acceleration model;

[0095] 2) Courtesy decision needs to satisfy the following conditions at time T c The position of the vehicle is S y To avoid reversing, the courtesy speed must be greater than or equal to 0.

[0096] 3) Based on the expected acceleration decision of the courtesy decision, the speed of the ego vehicle has a characteristic, that is, the speed of the ego vehicle is greater than 0, which means that the ego vehicle does not need to stop completely. The corresponding problem can be modeled as:

[0097] max a

[0098]

[0099] V e =V ego +a*T c

[0100] S ego≤S y

[0101] V e >0

[0102] a≤0

[0103] If the above problem has a solution, the obtained acceleration is used as the expected acceleration of the courtesy decision.

[0104] If the above problem has no solution, it means that the vehicle needs to first decelerate uniformly to 0 (stop), and then stop to yield. The corresponding problem can be modeled as:

[0105] max a

[0106]

[0107] S ego ≤S y

[0108] V e =0

[0109] a≤0

[0110] In one embodiment of the present application, the method further includes: using the desired acceleration value of the ego-vehicle as a penalty term in a dynamic programming search process based on the ST graph.

[0111] Through the above decision-making process, the desired driving acceleration value can be finally obtained, and then it is used as a penalty item in the dynamic planning search process of the ST diagram. The search process can refer to the speed planning part in the prior art, and this solution will not elaborate on it in detail.

[0112] The embodiment of the present application also provides a device 500 for identifying and avoiding collision risk intentions, such as Figure 5 As shown, a schematic diagram of the structure of a device for identifying and avoiding collision risk intention in an embodiment of the present application is provided, wherein the device 500 for identifying and avoiding collision risk intention at least includes: a determination module 510, a collision detection module 520, and an expected acceleration calculation module 530, wherein:

[0113] In one embodiment of the present application, the determination module 510 is specifically used to: determine a collision detection boundary and a target obstacle, wherein the collision detection boundary is determined based on route information and map lane data.

[0114] "Obstacles" are obtained through the upstream perception module, and are screened out based on the loaded map data, perception data, and the vehicle's driving path. "Collision detection boundaries" are generated by coupling the vehicle's lateral path (route) with the lane boundary. It can be understood that the lane boundary is the adjacent lane line as the boundary.

[0115] For the "target obstacle", considering the high uncertainty of future changes in the surrounding environment, the target obstacle cannot be too far away from the collision detection boundary. At the same time, considering that the traffic regulations do not allow continuous lane changes, the position range is limited to within one lane width from the collision detection boundary in the embodiment of the present application, and static obstacles are filtered out, such as Figure 1 In addition, some other filtering conditions can be added to avoid interference from irrelevant obstacles. For example, filtering conditions can be added according to the characteristics of the vehicle or engineering experience, so as to filter out unnecessary obstacles for a specific scene.

[0116] like Figure 1 As shown, the left and right collision detection boundaries can be generated according to the path information of the upstream module and the map lane data. It should be noted that the map lane data can preferably use an HD map.

[0117] Please continue to refer to Figure 1 Considering the driving habits of drivers, they often use the lane line of their own vehicle as an important judgment sign for whether the vehicles in the adjacent lanes will invade their own lanes. However, it is unsafe to rely solely on the lane line as the collision detection boundary for the active line-crossing behavior of the own vehicle due to other purposes such as the need to bypass obstacles. Therefore, it is necessary to couple the lateral path (route) of the own vehicle with the lane boundary to generate the final collision detection boundary. Specifically, taking the right collision detection boundary as an example to illustrate this process, first calculate the intersection of the right boundary of the lateral path (route) and the right boundary of the current lane. These intersections divide the right boundary of the lateral path (route) and the right boundary of the current lane into multiple intervals. The intervals that form the rightmost boundary are spliced ​​together to form the final right collision detection boundary. The left boundary is processed in the same way to obtain Figure 1 The collision detection boundaries are shown.

[0118] In one embodiment of the present application, the collision detection module 520 is specifically configured to calculate, according to the target obstacle, collision information of the target obstacle reaching the collision detection boundary.

[0119] According to the motion state information of the obstacle, assuming that the obstacle moves according to the uniform acceleration model, the virtual trajectory within a period of time in the future is calculated in the Frenet reference system, and the intersection information of the virtual trajectory and the collision detection boundary is the collision information. That is, for the selected target obstacle, the position and time when it first reaches the collision boundary are calculated.

[0120] like Figure 2 As shown, V S and V LThey represent the speed in different directions based on the center of the lane line (Frenet), and also include a virtual trajectory and a collision point on the collision detection boundary based on the virtual trajectory.

[0121] In one embodiment of the present application, the expected acceleration calculation module 530 is specifically used to calculate the expected acceleration value of the ego vehicle according to the collision information and the current state of the ego vehicle, so as to allow the ego vehicle to perform evasive action in decision planning.

[0122] The collision information includes information such as the collision position and collision time. The current state of the vehicle includes: the speed, acceleration and position information of the vehicle.

[0123] Based on the above information, the expected acceleration value of the ego vehicle is calculated, that is, the longitudinal motion of the ego vehicle is modeled to convert the collision information into the expected acceleration of the ego vehicle to guide the future avoidance action of the ego vehicle in decision planning.

[0124] In one embodiment of the present application, the collision detection module 520 is further used to:

[0125] According to the motion state information of the target obstacle in the Frenet system, a virtual obstacle trajectory of the target obstacle in a future period of time is obtained;

[0126] Considering the geometric shape of the obstacle, whether the obstacle virtual trajectory intersects with the collision detection boundary;

[0127] If they intersect, the position information of the first intersection point is calculated, where the first intersection point is the position point that first reaches the collision detection boundary;

[0128] Calculating the intersection time according to the position information of the first intersection point;

[0129] If they do not intersect, it is considered that the target obstacle poses no risk to the vehicle.

[0130] In one embodiment of the present application, it further includes: a Frenet coordinate system module, which is used to:

[0131] Based on the structured road, the Frenet reference system is established with the lane centerline as the X-axis and the vertical lane centerline as the Y-axis;

[0132] According to the Frenet reference system, the positions of the corner points of the obstacles perceived by the upstream module are projected in the Frenet reference system, and the corner points closest to the center line of the lane are retained to calculate the collision information, and the upstream module includes a prediction module and a fusion perception module;

[0133] The perceived obstacle position and speed are projected in the Frenet reference system to obtain the position and speed state of the corresponding obstacle in the Frenet reference system.

[0134] In one embodiment of the present application, the expected acceleration calculation module 530 is further used to:

[0135] Projecting the collision point in the collision information onto the lateral route to obtain the lateral route length corresponding to the collision point;

[0136] Based on the geometry of the vehicle and / or the obstacle and the preset safety distance, the overtaking point S is obtained by adding the expected overtaking distance to the lateral route length corresponding to the collision point. o Subtract the expected courtesy distance to get the courtesy point S y .

[0137] In one embodiment of the present application, the acceleration value expected by the vehicle includes: the acceleration expected for overtaking decision, and the expected acceleration calculation module 530 is further used for:

[0138] The current position of the vehicle is assumed to be an initial value of 0, and the current speed state of the vehicle is V ego As the initial state, the future driving trajectory of the ego vehicle is modeled as a uniform acceleration model;

[0139] The upper limit of acceleration when limiting overtaking decision is a b , and the overtaking decision of the ego vehicle is equivalent to the decision at time T c The vehicle's current position S ego Passing overtaking point S o , the corresponding problem can be modeled as the first vehicle motion problem;

[0140] If the first vehicle motion problem has a solution, there is no need to make a courtesy decision;

[0141] If the first vehicle motion problem has a solution, the expected acceleration for the overtaking decision is determined.

[0142] In one embodiment of the present application, the desired acceleration value of the vehicle includes: the expected acceleration of the courtesy decision, and the expected acceleration calculation module 530 is also used to

[0143] The current position of the vehicle is assumed to be an initial value of 0, and the current speed state of the vehicle is V ego As the initial state, the future driving trajectory of the ego vehicle is modeled as a uniform acceleration model;

[0144] At time T c The position of the vehicle is S yAfter that, the courtesy decision of the ego vehicle must satisfy the courtesy speed that is always greater than 0, and the corresponding problem can be modeled as the second vehicle motion problem;

[0145] If the second vehicle motion problem has a solution, the expected acceleration of the courtesy decision is determined and the speed of the vehicle does not need to be uniformly decelerated to 0;

[0146] If the second vehicle motion problem has no solution, the vehicle needs to first decelerate uniformly to 0, then stop and yield, and the resulting problem can be modeled as the third vehicle motion problem.

[0147] In one embodiment of the present application, the expected acceleration calculation module 530 is further used to use the expected acceleration value of the vehicle as a penalty item in the dynamic programming search process based on the ST graph.

[0148] It can be understood that the above-mentioned device for collision risk intention identification and avoidance can implement the various steps of the method for collision risk intention identification and avoidance provided in the aforementioned embodiment. The relevant explanations about the method for collision risk intention identification and avoidance are applicable to the device for collision risk intention identification and avoidance, and will not be repeated here.

[0149] Figure 6 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 6 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0150] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0151] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0152] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming an avoidance device based on collision risk intention recognition at the logical level. The processor executes the program stored in the memory and is specifically used to perform the following operations:

[0153] Determining a collision detection boundary and a target obstacle, wherein the collision detection boundary is determined based on route information and map lane data;

[0154] Calculating collision information of the target obstacle reaching the collision detection boundary according to the target obstacle;

[0155] According to the collision information and the current state of the ego vehicle, the expected acceleration value of the ego vehicle is calculated so as to allow the ego vehicle to perform evasive action in decision planning.

[0156] The above application Figure 4The method for executing collision risk intention identification and avoidance disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module can be located in a mature storage medium in the field such as random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0157] The electronic device may also perform Figure 4 A method for executing a device for identifying and avoiding collision risk intentions in a vehicle, and implementing a device for identifying and avoiding collision risk intentions in a vehicle Figure 4 The functions of the illustrated embodiment will not be described in detail in the embodiments of the present application.

[0158] The present application also provides a computer-readable storage medium, which stores one or more programs, wherein the one or more programs include instructions, which, when executed by an electronic device including multiple application programs, enable the electronic device to execute Figure 4 The method executed by the avoidance device based on collision risk intention recognition in the illustrated embodiment is specifically used to execute:

[0159] Determining a collision detection boundary and a target obstacle, wherein the collision detection boundary is determined based on route information and map lane data;

[0160] Calculating collision information of the target obstacle reaching the collision detection boundary according to the target obstacle;

[0161] According to the collision information and the current state of the ego vehicle, the expected acceleration value of the ego vehicle is calculated so as to allow the ego vehicle to perform evasive action in decision planning.

[0162] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0163] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0164] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0165] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0166] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0167] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0168] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0169] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0170] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0171] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for identifying and avoiding collision risk intentions, wherein: The method comprises: Determining a collision detection boundary and a target obstacle, wherein the collision detection boundary is determined based on route information and map lane data; Calculating collision information of the target obstacle reaching the collision detection boundary according to the target obstacle; According to the collision information and the current state of the ego vehicle, the expected acceleration value of the ego vehicle is calculated so as to allow the ego vehicle to perform evasive action in decision planning.

2. The method of claim 1, wherein: The calculating, according to the target obstacle, collision information of the target obstacle reaching the collision detection boundary includes: According to the motion state information of the target obstacle in the Frenet system, a virtual obstacle trajectory of the target obstacle in a future period of time is obtained; Considering the geometric shape of the obstacle, determining whether the obstacle virtual trajectory intersects with the collision detection boundary; If they intersect, the position information of the first intersection point is calculated, where the first intersection point is the position point that first reaches the collision detection boundary; According to the position information of the first intersection point, the intersection time T is calculated. c ; If they do not intersect, it is considered that the target obstacle poses no risk to the vehicle.

3. The method of claim 2, wherein: The method further comprises: Based on the structured road, the Frenet reference system is established with the lane centerline as the X-axis and the vertical lane centerline as the Y-axis; According to the Frenet reference system, the positions of the corner points of the obstacles sensed by the upstream module are projected in the Frenet reference system, and the corner points closest to the center line of the lane are retained to calculate the collision information, wherein the upstream module includes a prediction module and / or a fusion perception module; The perceived obstacle position and obstacle speed are projected in the Frenet reference system to obtain the position and speed state of the corresponding obstacle in the Frenet reference system.

4. The method of claim 1, wherein: The step of calculating the expected acceleration value of the vehicle according to the collision information and the current state of the vehicle so as to allow the vehicle to perform an evasive action in decision planning includes: Project the collision point in the collision information onto the lateral route to obtain the lateral route length S corresponding to the collision point c , the transverse route length S c is the length between the ego vehicle and the collision point; Based on the geometry of the vehicle and / or the obstacle and the preset safety distance, the lateral path length S at the corresponding collision point is c Based on the expected overtaking distance, the overtaking point S is obtained. o Subtract the expected courtesy distance to get the courtesy point S y .

5. The method of claim 4, wherein: The acceleration value expected by the vehicle includes: the acceleration expected by the overtaking decision, and the method further includes: The current position of the vehicle is assumed to be an initial value of 0, and the current speed state of the vehicle is V ego As the initial state, the future driving trajectory of the ego vehicle is modeled as a uniform acceleration model; The upper limit of acceleration when limiting overtaking decision is a b , and the overtaking decision of the ego vehicle is equivalent to the decision at time T c The vehicle's current position S ego Passing overtaking point S o , the corresponding problem can be modeled as the first vehicle motion problem; If the first vehicle motion problem has a solution, there is no need to make a courtesy decision; If the first vehicle motion problem has a solution, the expected acceleration for the overtaking decision is determined.

6. The method of claim 4, wherein: The acceleration value expected by the vehicle includes: the acceleration expected by the courtesy decision, and the method further includes: The current position of the vehicle is assumed to be an initial value of 0, and the current speed state of the vehicle is V ego As the initial state, the future driving trajectory of the ego vehicle is modeled as a uniform acceleration model; At time T c The position of the vehicle is S y After that, the courtesy decision of the ego vehicle must satisfy the courtesy speed that is always greater than 0, and the corresponding problem can be modeled as the second vehicle motion problem; If the second vehicle motion problem has a solution, the expected acceleration of the courtesy decision is determined and the speed of the vehicle does not need to be uniformly decelerated to 0; If the second vehicle motion problem has no solution, the vehicle needs to first decelerate uniformly to 0, then stop and yield, and the resulting problem can be modeled as the third vehicle motion problem.

7. The method according to any one of claims 1 to 6, wherein: The method further comprises: The desired acceleration value of the ego vehicle is used as a penalty term in a dynamic programming search process based on the ST graph.

8. A device for identifying and avoiding collision risk intentions, wherein: The device comprises: A determination module, used to determine a collision detection boundary and a target obstacle, wherein the collision detection boundary is determined based on route information and map lane data; A collision detection module, used for calculating collision information of the target obstacle reaching the collision detection boundary according to the target obstacle; The expected acceleration calculation module is used to calculate the expected acceleration value of the ego vehicle according to the collision information and the current state of the ego vehicle, so as to allow the ego vehicle to perform evasive actions in decision planning.

9. An electronic device, comprising: processor; and a memory arranged to store computer executable instructions, which, when executed, cause the processor to perform the method of any one of claims 1 to 6.

10. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, causes the electronic device to execute any one of the methods of claims 1 to 6.

Images