Decision method and device, electronic equipment and storage medium

Abstract

The invention discloses a decision-making method and device, electronic equipment and a storage medium, and the method comprises the steps: traversing an obstacle and a predicted track of the obstacle, and obtaining the obstacle appearing on a path plan of a vehicle; projecting an obstacle appearing on the path planning of the own vehicle to the S-T diagram, and calculating a transverse distance from a predicted trajectory of the obstacle close to the own vehicle to the planned path of the own vehicle; on the basis of a pre-configured transverse risk estimation model, according to the speed of the obstacle at the target moment t and the transverse distance between the obstacle and the vehicle, the maximum passing speed of safe passing of the vehicle at the target moment t is calculated; and when the penalty value of the vehicle at the target moment t is calculated, increasing the penalty value that the speed of the vehicle at the target moment t exceeds the maximum passing speed of safe passing of the vehicle, and outputting a decision track. According to the invention, on one hand, the maximum safe passing speed of the vehicle is calculated, and on the other hand, sufficient transverse safety is ensured.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a decision-making method, device, electronic device, and storage medium. Background Art

[0002] The decision-making and planning algorithms in autonomous driving usually adopt a horizontal and vertical separation architecture. Specifically, in the horizontal decision-making and planning, a smooth path curve without speed information is first planned, and in the vertical decision-making and planning, a safe, comfortable and efficient speed curve, i.e., trajectory, is planned based on the path.

[0003] The mainstream lateral decision-making planning algorithm can ensure that the output path is safe and collision-free with static obstacles in space. It also ensures that the final output trajectory is safe and collision-free with all obstacles in time and space. The premise for the mainstream longitudinal decision-making planning algorithm to ensure safety is that the obstacle must be within a certain lateral distance threshold of the path (generally set to half the width of the vehicle + buffer value), that is, if the lateral distance of the obstacle from the path is greater than the threshold, the longitudinal decision-making planning will not consider the obstacle at all. Although this can theoretically ensure safety and collision-free in time and space, due to a series of lateral uncertainties between the vehicle and obstacles, such as intention, perception, and control errors, it is impossible to guarantee absolute lateral safety by simply setting a fixed threshold. Summary of the invention

[0004] The embodiments of the present application provide a decision-making method, device, electronic device, and storage medium to ensure sufficient horizontal security.

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

[0006] In a first aspect, an embodiment of the present application provides a decision-making method, wherein the method comprises:

[0007] Traversing obstacles and predicted trajectories of the obstacles to obtain obstacles that appear on the vehicle's path planning;

[0008] Projecting obstacles that appear on the planned path of the ego vehicle onto the ST graph, and calculating the lateral distance from the predicted trajectory of the obstacle near the ego vehicle to the planned path of the ego vehicle, the ST graph is used to characterize the relationship between the planned path length and time, and the obstacles near the ego vehicle are obtained through screening;

[0009] Based on the pre-configured lateral risk estimation model, according to the speed of the obstacle at the target time t and the lateral distance between the obstacle and the vehicle, the maximum speed at which the vehicle can safely pass at the target time t is calculated;

[0010] When calculating the penalty value of the ego vehicle at the target time t, the penalty value of the ego vehicle speed exceeding the maximum speed at which the ego vehicle can safely pass at the target time t is added, and the decision trajectory is output.

[0011] In some embodiments, the method further comprises:

[0012] Determine whether the speed of the ego vehicle at the target time t exceeds the maximum speed at which the ego vehicle can safely pass, and if so, penalize the decision-making process of the ego vehicle that exceeds the maximum speed;

[0013] Otherwise, there is no penalty for the car's decision-making process.

[0014] In some embodiments, the method further comprises:

[0015] According to the predicted trajectory of the obstacle close to the ego vehicle at the lateral distance, a lateral distance between the ego vehicle and the predicted trajectory of the obstacle close to the ego vehicle at the lateral distance is obtained, and the lateral distance is added to the decision process. In some embodiments, the preconfigured lateral risk estimation model includes:

[0016] Based on the TTC model or RSS model, lateral collision risk modeling is performed;

[0017] According to the lateral collision risk modeling result, the collision time TTC=L / (V1+V1') is obtained;

[0018] Wherein, L is the lateral distance between the vehicle and the front vehicle, V1 is the lateral speed of the front vehicle, V1' is the lateral speed of the vehicle, and TTC is used to determine whether the current lateral distance and lateral speed are safe.

[0019] In some embodiments, the preconfigured lateral risk estimation model further comprises:

[0020] The vehicle has a fixed steering wheel vibration error during driving, and the error is transmitted to the vehicle chassis and manifested as the deflection angle of the vehicle's front wheels.

[0021] θ = f * δ;

[0022] Wherein, f is the transmission ratio from the vehicle steering wheel to the front wheel steering, and δ is the vehicle steering wheel angle;

[0023] Based on the deviation angle of the vehicle's front wheels, additional lateral speed is obtained

[0024] V1=v*θ

[0025] Wherein, V1 is proportional to the vehicle's forward speed v, and the proportionality coefficient is the deflection angle θ of the vehicle's front wheels.

[0026] In some embodiments, the preconfigured lateral risk estimation model further comprises:

[0027] The lateral distance is determined by the distance between the obstacle outline and the vehicle body.

[0028] A time threshold is set according to the obstacle type, where the time threshold is the shortest time calculated according to the TTC or RSS model based on the current lateral speed and lateral distance.

[0029] In some embodiments, the preconfigured lateral risk estimation model further comprises:

[0030] According to the longitudinal speed v, lateral distance L and time threshold TTC of the obstacle, the maximum safe longitudinal speed of the vehicle is calculated.

[0031]

[0032] Among them, the TTC time threshold and the front wheel turning angle of the other vehicle are determined according to the type of obstacle, and the deflection angle of the front wheel of the vehicle is an adjustable parameter.

[0033] In a second aspect, an embodiment of the present application further provides a decision-making device, wherein the device includes:

[0034] A traversal processing module, used to traverse obstacles and predicted trajectories of the obstacles to obtain obstacles that appear on the vehicle path planning;

[0035] A projection calculation module is used to project the obstacles that appear in the planned path of the ego vehicle onto the ST graph, and calculate the lateral distance from the predicted trajectory of the obstacles near the ego vehicle to the planned path of the ego vehicle. The ST graph is used to characterize the relationship between the length of the planned path and time. The obstacles near the ego vehicle are obtained through screening.

[0036] A speed calculation module, for calculating a maximum speed at which the vehicle can safely pass at the target time t based on a pre-configured lateral risk estimation model and the speed of the obstacle at the target time t and the lateral distance between the obstacle and the vehicle;

[0037] The decision module is used to increase the penalty value of the vehicle speed exceeding the maximum speed at which the vehicle can safely pass at the target time t when calculating the penalty value of the vehicle at the target time t, and output the decision trajectory.

[0038] 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.

[0039] 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.

[0040] At least one of the above technical solutions adopted in the embodiment of the present application can achieve the following beneficial effects: Obstacles appearing on the self-vehicle path planning are obtained by traversing obstacles and the predicted trajectory of the obstacles. Then the obstacles appearing on the self-vehicle path planning are projected onto the ST graph, and the lateral distance from the predicted trajectory of the obstacle adjacent to the self-vehicle to the self-vehicle planned path is calculated. Based on the pre-configured lateral risk estimation model, the maximum speed at which the self-vehicle passes safely at the target time t is calculated according to the speed of the obstacle at the target time t and the lateral distance between the obstacle and the self-vehicle. Finally, when calculating the penalty value of the self-vehicle at the target time t, the penalty value of the self-vehicle speed exceeding the maximum speed at which the self-vehicle passes safely is increased, and the decision trajectory is output. Through the above method, a decision process with spatiotemporal joint characteristics is realized. At the same time, the adopted lateral risk estimation model includes uniformly modeling lateral uncertainty as lateral speed, and the TTC time threshold and front wheel steering angle can be flexibly set according to the obstacle type. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] 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:

[0042] Figure 1 This is a schematic diagram of a lateral risk estimation model in an embodiment of the present application;

[0043] Figure 2 A flowchart of a decision-making method in an embodiment of the present application;

[0044] Figure 3 This is a schematic diagram of the structure of the decision-making device in the embodiment of the present application;

[0045] Figure 4 This is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0046] 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.

[0047] The processing flow of mainstream vertical decision planning algorithms can be divided into:

[0048] (1) Generate ST projection diagram.

[0049] (2) Longitudinal decision: Generate a rough trajectory based on the ST projection map.

[0050] (3) Longitudinal planning: Generate the smoothed final trajectory based on the ST projection map and the rough trajectory.

[0051] First, the process of generating the ST projection map can be simplified as follows: the obstacle contour on the predicted obstacle trajectory (the trajectory of a static obstacle can be considered as a fixed value of the current position).

[0052] Secondly, decision-making based on the ST projection graph: use sampling or search algorithms to obtain a safe rough trajectory with the minimum penalty value COST on the ST graph.

[0053] Finally, the final trajectory is at least one safe, comfortable and efficient trajectory generated by optimizing the rough trajectory and the convex space in which it is located.

[0054] However, the longitudinal decision-making planning scheme can only guarantee no collision with obstacles that appear in the ST projection diagram, but ignores the potential lateral collision risk of obstacles outside the ST diagram. The reason is: in order to appear on the ST projection diagram, the lateral distance of the obstacle's contour from the path curve is required to be within the threshold (half the vehicle width + buffer value). The buffer value represents the minimum lateral distance allowed between the obstacle contour and the vehicle contour, and is generally small, such as 0.2m. Through analysis, it can be seen that if the lateral distance value is greater than the buffer value, the longitudinal decision planning will ignore the obstacle. Since the lateral distance is greater than the buffer value, it means that there will be no lateral collision in theory, so it does not appear in the ST diagram. However, in reality, the collision risk caused by the lateral uncertainty between the vehicle and the obstacle is ignored.

[0055] Taking driving common sense as an example, if there are stationary cars on both sides of the road ahead, if the car wants to pass, the minimum distance from the stationary cars on both sides is only slightly greater than 0.2m (for example, 0.21m). According to the existing longitudinal decision-making planning algorithm, the lateral distance of the stationary cars on both sides from the path is greater than the buffer value and will not enter the ST diagram. Therefore, the longitudinal decision will completely ignore the stationary cars on both sides, and may eventually plan a speed curve for high-speed passing. However, human drivers do the opposite. When passing the stationary cars on both sides, they will reduce the speed to a reasonable range and pass at a low speed. If the buffer value is artificially increased in the existing algorithm and the stationary cars on both sides are forced to enter the ST diagram, the longitudinal speed curve will be planned to completely stop in front of the stationary car, making it impossible to pass. The reason for this difference is that the existing algorithm ignores the risk of lateral collision with obstacles.

[0056] In this way, due to the error between the control of the ego vehicle and the perception of the stationary vehicle, there is a certain lateral risk, and as the speed of the ego vehicle increases, the lateral wind direction also increases, so the ego vehicle cannot pass at a faster speed. Only by controlling the speed of the ego vehicle to a certain range (the value is related to the lateral distance) can the lateral risk be reduced to below the psychological expectations of human drivers, that is, safe passage. In addition, the existing longitudinal decision planning algorithm cannot reasonably handle the risk of lateral collision with dynamic obstacles.

[0057] In view of the above shortcomings, an improved longitudinal decision algorithm is provided in the embodiment of the present application. The algorithm fully considers all adjacent obstacles that have not entered the ST diagram, calculates the maximum passing speed according to the lateral risk estimation model, and punishes the self-vehicle decision that exceeds the maximum passing speed in the longitudinal decision, thereby achieving safe passage. At the same time, a mathematical model for lateral risk estimation is also provided in the longitudinal decision algorithm, which can calculate the maximum speed of the self-vehicle to pass safely through the known speed of the obstacle and the lateral distance.

[0058] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0059] The present application embodiment provides a decision-making method, such as Figure 2 As shown, a flowchart of a decision-making method in an embodiment of the present application is provided, and the method at least includes the following steps S210 to S240:

[0060] Step S210, traversing obstacles and predicted trajectories of the obstacles to obtain obstacles appearing on the vehicle path planning.

[0061] Obstacle information can be obtained by the vehicle positioning and perception module and passed to the downstream module. The predicted trajectory of the obstacle is determined based on the distance from the vehicle and the speed of the obstacle. It can be understood that obstacles include but are not limited to the vehicle in front, stationary obstacles, and other traffic participants traveling at low speeds.

[0062] By traversing the above obstacles and their trajectory information, the obstacles that appear on the vehicle path planning can be obtained. It can be understood that the obstacles here include all possible obstacles.

[0063] Step S220, projecting obstacles appearing on the planned path of the ego vehicle onto the ST diagram, and calculating the lateral distance from the predicted trajectory of the obstacle near the ego vehicle to the planned path of the ego vehicle. The ST diagram is used to characterize the relationship between the length of the planned path and time, and the obstacles near the ego vehicle are obtained through screening.

[0064] According to the method in the related art, the obstacles appearing on the planned path of the ego vehicle are projected onto the ST map, and the lateral distance from the predicted trajectory of other adjacent obstacles to the planned path of the ego vehicle is calculated based on the projection result of the ST map. It should be noted that other adjacent obstacles are obtained through pre-screening rules, which can reduce the amount of calculation compared to the case where obstacles are not screened.

[0065] In the related art, the longitudinal decision-making process only considers obstacles that appear in the ST diagram, which is equivalent to treating obstacles as two-dimensional objects in the ST space of the vehicle path, where S represents the path length and T represents the time, thus ignoring the lateral distance dimension of the obstacle.

[0066] The above process not only needs to consider the obstacles in the ST diagram, but also the obstacles adjacent to the vehicle in the lateral distance. The lateral distance is directly incorporated into the decision-making process, thereby restoring the lateral distance between the obstacle and the path, and having the three-dimensional spatiotemporal joint decision-making characteristics under the vehicle path.

[0067] Step S230, based on a pre-configured lateral risk estimation model, according to the speed of the obstacle at the target time t and the lateral distance between the obstacle and the vehicle, calculate the maximum speed at which the vehicle can safely pass at the target time t.

[0068] like Figure 1 As shown, the lateral distance between the vehicle and the front vehicle is L, and the lateral speed of the front vehicle is V L , the lateral velocity of the vehicle is V L Based on the pre-configured lateral risk estimation model, the maximum speed at which the ego vehicle can safely pass is calculated according to the speed of the obstacle at the target time t and the lateral distance between the obstacle and the ego vehicle.

[0069] It can be understood that the target time t is any time and is determined according to actual needs.

[0070] Step S240, when calculating the penalty value of the ego vehicle at the target time t, the penalty value of the ego vehicle speed exceeding the maximum speed at which the ego vehicle can safely pass at the target time t is added, and the decision trajectory is output.

[0071] When calculating the COST penalty value of the ego vehicle at the target time t, add the value of the ego vehicle’s target time t speed exceeding V t The COST penalty value is used as the lateral risk COST penalty value and added to the decision-making process of the vehicle.

[0072] Since the predicted obstacle trajectory is taken into account, the actual moment between the obstacle and the vehicle can be calculated based on the lateral distance and the travel speed. In this way, a decision can be made to slow down at a reasonable time and space position to avoid lateral risks, which has certain time and space joint planning characteristics.

[0073] Through the above method, in the entire longitudinal decision-making process, decisions with lateral collision risks will receive certain penalties, thereby outputting a safe and risk-free decision trajectory.

[0074] Through the above method, all adjacent obstacles that have not entered the ST diagram are fully considered. The maximum speed for the vehicle to pass safely can be calculated based on the known speed and lateral distance of the obstacle. The maximum passing speed is then calculated based on the lateral risk estimation model. In the longitudinal decision, the vehicle decision that exceeds the maximum passing speed is punished, thereby achieving safe passage.

[0075] Through the above method, the lateral uncertainty is uniformly modeled as the lateral velocity. At the same time, the TTC time threshold and the front wheel steering angle can be flexibly set according to the obstacle type.

[0076] By treating obstacles as two-dimensional objects in the ST space of the ego vehicle path (S: path length, T: time), the lateral distance dimension of the obstacles is ignored. The optimized longitudinal decision not only needs to consider the obstacles in the ST diagram, but also the obstacles adjacent to the lateral distance, and directly incorporates the lateral distance into the decision-making process, thus restoring the three-dimensional properties of the obstacles (the distance between the obstacle and the ego vehicle, the lateral speed of the obstacle, and the lateral distance from the planned path of the ego vehicle), and has the three-dimensional spatiotemporal joint decision-making characteristics under the ego vehicle path.

[0077] Traverse all obstacles and their predicted trajectories, project obstacles that appear on the vehicle's path to the ST graph, and calculate the lateral distance from the predicted trajectories of other adjacent obstacles to the vehicle's path. Calculate the maximum speed V of the vehicle based on the speed of the obstacle at a certain time t and the lateral distance from the vehicle. tWhen calculating the cost of the vehicle at time t, increase the speed of the vehicle at time t to exceed V t The COST is added to the decision-making process of the vehicle as a horizontal risk COST.

[0078] It can be understood that, except for the above-mentioned penalty process, other decision processes are by default consistent with the mainstream decision processes based on ST diagrams.

[0079] Since the obstacle prediction trajectory is taken into account, the lateral distance and speed calculation can be accurate to the moment when the obstacle and the vehicle actually interact. In this way, it can be decided to slow down at a reasonable time and space position to avoid lateral risks, and it has certain time and space joint planning characteristics. In the entire longitudinal decision-making process, decisions with lateral collision risks will receive certain penalties, thereby outputting safe and risk-free decision trajectories.

[0080] In one embodiment of the present application, the method also includes: determining whether there is a situation where the speed of the self-vehicle at the target time t exceeds the maximum speed for the self-vehicle to pass safely, and if so, penalizing the decision-making process of the self-vehicle that exceeds the maximum speed; otherwise, not penalizing the decision-making process of the self-vehicle.

[0081] It can be understood that the autonomous driving decision penalty cost refers to a mechanism introduced in the autonomous driving system to optimize the decision-making process, which punishes those behaviors or decisions that do not meet expectations by adding additional costs. For example, penalties include obstacle avoidance penalties, penalties for violating traffic rules, penalties for path smoothness, penalties for distance from the reference line, etc. The design and adjustment of these penalty costs need to be optimized according to the specific driving environment and needs to ensure the safe and efficient operation of the autonomous driving system.

[0082] When determining the penalty value, it is necessary to determine whether the speed of the ego vehicle at the target time t exceeds the maximum speed at which the ego vehicle can safely pass. If so, the decision-making process of the ego vehicle that exceeds the maximum speed will be penalized. In the entire longitudinal decision-making process, decisions with a risk of lateral collision will receive a certain penalty to optimize the decision, thereby outputting a safe and risk-free decision trajectory. Similarly, if otherwise, no penalty will be imposed on the decision-making process of the ego vehicle.

[0083] In one embodiment of the present application, the method also includes obtaining the lateral distance between the self-vehicle and the predicted trajectory of the obstacle close to the self-vehicle at the lateral distance based on the predicted trajectory of the obstacle close to the self-vehicle, and adding the lateral distance to the decision-making process. By fully considering all adjacent obstacles that have not entered the ST diagram, the maximum passing speed is calculated according to the lateral risk estimation model, and the decision of the self-vehicle that exceeds the maximum passing speed is punished in the longitudinal decision, thereby achieving safe passage.

[0084] In one embodiment of the present application, the pre-configured lateral risk estimation model includes: lateral collision risk modeling based on the TTC model or the RSS model; according to the lateral collision risk modeling result, obtaining the collision time TTC=L / (V1+V1'); wherein, L is the lateral distance between the vehicle and the front vehicle, V1 is the lateral speed of the front vehicle, V1' is the lateral speed of the vehicle, and the TTC is used to determine whether it is safe at the current lateral distance and lateral speed.

[0085] TTC (Time To Collision) is a key indicator used to measure the time required for a vehicle to travel normally until a collision occurs. It plays a vital role in improving automobile safety.

[0086] RSS (Responsibility Sensitive Safety) aims to transform human concepts of safe driving and the division of accident responsibilities into reference parameters for mathematical models and decision-making control.

[0087] Considering that the lateral risk of static obstacles comes from perception errors and potential starting intentions, the lateral risk of dynamic obstacles is more complicated, including not only perception errors, but also their complex movement intentions (whether to change lanes or cut in or turn), speed, size, type, control errors, etc.

[0088] Generally, the greater the speed of the obstacle, the greater the perceived error. When the steering wheel control error is constant, the greater the speed of the obstacle, the smaller the steering wheel fluctuation will cause the vehicle to have a larger lateral speed fluctuation. Similarly, if the obstacle type is an electric vehicle, bicycle or pedestrian, its complex movement intention can be reflected in its flexible steering ability. In addition, if the obstacle type is a large vehicle, due to its huge size and mass advantages, the vehicle needs to maintain a larger lateral safety distance. Therefore, a mathematical model is needed to model it reasonably.

[0089] Preferably, in an embodiment of the present application, the lateral risk of an obstacle is uniformly modeled as a TTC or RSS mathematical relationship between the lateral speed and the lateral distance.

[0090] TTC and RSS models, TTC and RSS are currently mature collision safety models, and the same models are also used in the embodiments of this application to model lateral collision risks. Figure 1 As shown, the lateral distance between the vehicle and the front vehicle is L, and the lateral speed of the other vehicle is V L , the lateral velocity of the vehicle is V L '.

[0091] TTC = L / (V1+V1') (Formula 1)

[0092] If TTC is less than the set time threshold, it is judged that there is a risk of lateral collision. RSS is a responsibility safety sensitive model, similar to TTC, which can be used to determine whether the current lateral distance and lateral speed are safe.

[0093] In one embodiment of the present application, the pre-configured lateral risk estimation model further includes: a fixed steering wheel vibration error exists during the driving process of the vehicle, and the error is transmitted to the vehicle chassis and manifested as the deflection angle of the front wheel of the vehicle

[0094] θ = f * δ;

[0095] Wherein, f is the transmission ratio from the vehicle steering wheel to the front wheel steering, and δ is the vehicle steering wheel angle;

[0096] Based on the deviation angle of the vehicle's front wheels, additional lateral speed is obtained

[0097] V L =v*θ

[0098] Wherein, V1 is proportional to the vehicle's forward speed v, and the proportionality coefficient is the deflection angle θ of the vehicle's front wheels.

[0099] From the TTC and RSS models, it is known that the lateral velocity modeling of the vehicle and obstacles is crucial. However, a normal straight-moving vehicle does not generate lateral velocity, so in the embodiments of the present application, it is proposed to model the potential lateral velocity through the forward velocity of the vehicle. Considering that the vehicle may have a fixed steering wheel jitter error during driving, the error is transmitted to the vehicle chassis and manifested as the deflection angle of the vehicle's front wheels, as shown in the following formula:

[0100] θ=f*δ (Formula 2)

[0101] Among them, f is the transmission ratio from the vehicle steering wheel to the front wheel steering, which is determined by mechanical characteristics and generally floats around a fixed value and can be defaulted to this fixed value. δ is the vehicle steering wheel angle.

[0102] Due to the front wheel deflection, additional lateral speed is generated, as shown in the V in the figure above. L As shown,

[0103] V L =v*θ(Formula 3)

[0104] That is, the lateral speed is proportional to the vehicle's forward speed, and the proportionality coefficient is the front wheel deflection angle θ.

[0105] It should be noted that the front wheel deflection angle needs to be set according to the flexibility and intention of the obstacle:

[0106] For vehicles with flexible steering, such as bicycles and electric vehicles, the steering wheel vibration error is larger and the steering intention is more difficult to predict. They can be uniformly modeled as a larger front wheel turning angle, resulting in a larger lateral speed. Therefore, when passing such obstacles, it is necessary to maintain a larger lateral safety distance from such obstacles or reduce the vehicle speed. The above model can calculate a reasonable lateral safety distance or passing speed for different types of obstacles, showing human-like behavior.

[0107] In one embodiment of the present application, the pre-configured lateral risk estimation model also includes: using the distance of the obstacle contour relative to the vehicle body as the lateral distance; setting a time threshold according to the obstacle type, and the time threshold is the shortest time calculated according to the TTC or RSS model based on the current lateral speed and lateral distance.

[0108] In specific implementation, the lateral distance is the lateral distance between the obstacle outline and the vehicle body. Generally, the perception error can be considered and a certain contraction can be made based on the absolute distance.

[0109] Time threshold: The collision time is the minimum time calculated based on the current lateral speed and lateral distance according to the TTC or RSS model. If the time is greater than the set time threshold, it is considered that there is no collision risk; otherwise, it is considered that there is a risk.

[0110] It should be noted that the time threshold is set according to the obstacle type. For obstacles such as large vehicles, pedestrians and two-wheeled vehicles, human driving habits are to keep a longer safe distance or reduce the passing speed appropriately. This feature can be achieved by setting a larger time threshold for such obstacles.

[0111] In one embodiment of the present application, the pre-configured lateral risk estimation model further includes: calculating the maximum safe longitudinal speed of the vehicle according to the longitudinal speed v, lateral distance L, and time threshold TTC of the obstacle.

[0112]

[0113] Among them, the TTC time threshold and the front wheel turning angle of the other vehicle are determined according to the type of obstacle, and the deflection angle of the front wheel of the vehicle is an adjustable parameter.

[0114] Among them, v 障碍物 θ 障碍物 , lateral velocity and V 障碍物 The vehicle's forward speed is proportional to θ 障碍物 The proportionality factor is the front wheel deflection angle θ.

[0115] Among them, f 自 is the transmission ratio from the vehicle steering wheel to the front wheel steering, which can be set to this fixed value by default, δ 自is the vehicle steering wheel angle.

[0116] When the longitudinal speed, lateral distance, and time threshold of the obstacle vehicle are known, the maximum safe speed of the vehicle can be calculated by combining formulas (1, 2, and 3), that is, the longitudinal speed. The TTC time threshold and the front wheel angle of the other vehicle vary with the type of obstacle, and the steering wheel angle of the vehicle is also an adjustable parameter.

[0117] The present application embodiment also provides a decision-making device 300, such as Figure 3 As shown, a schematic diagram of the structure of the decision-making device in an embodiment of the present application is provided, wherein the decision-making device 300 at least includes: a traversal processing module 310, a projection calculation module 320, a speed calculation module 330 and a decision module 340, wherein:

[0118] In one embodiment of the present application, the traversal processing module 310 is specifically used to traverse obstacles and predicted trajectories of the obstacles to obtain obstacles that appear on the vehicle path planning.

[0119] Obstacle information can be obtained by the vehicle positioning and perception module and passed to the downstream module. The predicted trajectory of the obstacle is determined based on the distance from the vehicle and the speed of the obstacle. It can be understood that obstacles include but are not limited to the vehicle in front, stationary obstacles, and other traffic participants traveling at low speeds.

[0120] By traversing the above obstacles and their trajectory information, the obstacles that appear on the vehicle path planning can be obtained. It can be understood that the obstacles here include all possible obstacles.

[0121] In one embodiment of the present application, the projection calculation module 320 is specifically used to: project obstacles appearing on the planned path of the vehicle onto the ST diagram, and calculate the lateral distance from the predicted trajectory of the obstacle adjacent to the vehicle to the planned path of the vehicle. The ST diagram is used to characterize the relationship between the length of the planned path and time, and the obstacles adjacent to the vehicle are obtained through screening.

[0122] According to the method in the related art, the obstacles appearing on the planned path of the ego vehicle are projected onto the ST map, and the lateral distance from the predicted trajectory of other adjacent obstacles to the planned path of the ego vehicle is calculated based on the projection result of the ST map. It should be noted that other adjacent obstacles are obtained through pre-screening rules, which can reduce the amount of calculation compared to the case where obstacles are not screened.

[0123] In the related art, the longitudinal decision-making process only considers obstacles that appear in the ST diagram, which is equivalent to treating obstacles as two-dimensional objects in the ST space of the vehicle path, where S represents the path length and T represents the time, thus ignoring the lateral distance dimension of the obstacle.

[0124] The above process not only needs to consider the obstacles in the ST diagram, but also the obstacles adjacent to the vehicle in the lateral distance. The lateral distance is directly incorporated into the decision-making process, thereby restoring the lateral distance between the obstacle and the path, and having the three-dimensional spatiotemporal joint decision-making characteristics under the vehicle path.

[0125] In one embodiment of the present application, the speed calculation module 330 is specifically used to: based on a preconfigured lateral risk estimation model, calculate the maximum passing speed for the vehicle to safely pass at the target time t according to the speed of the obstacle at the target time t and the lateral distance between the obstacle and the vehicle.

[0126] like Figure 1 As shown, the lateral distance between the vehicle and the front vehicle is L, and the lateral speed of the front vehicle is V L , the lateral velocity of the vehicle is V L Based on the pre-configured lateral risk estimation model, according to the speed of the obstacle at the target time t and the lateral distance between the obstacle and the ego vehicle, the decision penalty calculation module 340 is specifically used to: when calculating the penalty value of the ego vehicle at the target time t, increase the penalty value of the ego vehicle at the target time t exceeding the maximum speed of the ego vehicle for safe passage, and output the decision trajectory.

[0127] When calculating the COST penalty value of the ego vehicle at the target time t, add the value of the ego vehicle’s target time t speed exceeding V t The COST penalty value is used as the lateral risk COST penalty value and added to the decision-making process of the vehicle.

[0128] Since the predicted obstacle trajectory is taken into account, the actual moment between the obstacle and the vehicle can be calculated based on the lateral distance and the travel speed. In this way, a decision can be made to slow down at a reasonable time and space position to avoid lateral risks, which has certain time and space joint planning characteristics.

[0129] In one embodiment of the present application, a judgment module is also included for

[0130] Determine whether the speed of the ego vehicle at the target time t exceeds the maximum speed at which the ego vehicle can safely pass, and if so, penalize the decision-making process of the ego vehicle that exceeds the maximum speed;

[0131] Otherwise, there is no penalty for the car's decision-making process.

[0132] In one embodiment of the present application, a lateral distance module is also included for

[0133] According to the predicted trajectory of the obstacle close to the ego vehicle at the lateral distance, the lateral distance between the ego vehicle and the predicted trajectory of the obstacle close to the ego vehicle at the lateral distance is obtained, and the lateral distance is added to the decision process.

[0134] In one embodiment of the present application, the speed calculation module 330 is also used to

[0135] Based on the TTC model or RSS model, lateral collision risk modeling is performed;

[0136] According to the lateral collision risk modeling result, the collision time TTC=L / (V1+V1') is obtained;

[0137] Wherein, L is the lateral distance between the vehicle and the front vehicle, V1 is the lateral speed of the front vehicle, V1' is the lateral speed of the vehicle, and TTC is used to determine whether the current lateral distance and lateral speed are safe.

[0138] In one embodiment of the present application, the speed calculation module 330 is also used to

[0139] The vehicle has a fixed steering wheel vibration error during driving, and the error is transmitted to the vehicle chassis and manifested as the deflection angle of the vehicle's front wheels.

[0140] θ = f * δ;

[0141] Wherein, f is the transmission ratio from the vehicle steering wheel to the front wheel steering, and δ is the vehicle steering wheel angle;

[0142] Based on the deviation angle of the vehicle's front wheels, additional lateral speed is obtained

[0143] V1=v*θ

[0144] Wherein, V1 is proportional to the vehicle's forward speed v, and the proportionality coefficient is the deflection angle θ of the vehicle's front wheels.

[0145] In one embodiment of the present application, the preconfigured lateral risk estimation model further includes:

[0146] The lateral distance is determined by the distance between the obstacle outline and the vehicle body.

[0147] A time threshold is set according to the obstacle type, and the time threshold is used to calculate the shortest time according to the TTC or RSS model based on the current lateral speed and lateral distance.

[0148] In one embodiment of the present application, the speed calculation module 330 is further used to calculate the maximum safe longitudinal speed of the vehicle according to the longitudinal speed v, the lateral distance L, and the time threshold TTC of the obstacle.

[0149]

[0150] Among them, the TTC time threshold and the front wheel turning angle of the other vehicle are determined according to the type of obstacle, and the deflection angle of the front wheel of the vehicle is an adjustable parameter.

[0151] It can be understood that the above-mentioned decision-making device can implement each step of the decision-making method provided in the above-mentioned embodiment, and the relevant explanations about the decision-making method are applicable to the decision-making device and will not be repeated here.

[0152] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 4 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. 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.

[0153] 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 4 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.

[0154] 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.

[0155] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a decision-making device at the logical level. The processor executes the program stored in the memory and is specifically used to perform the following operations:

[0156] Traversing obstacles and predicted trajectories of the obstacles to obtain obstacles that appear on the vehicle's path planning;

[0157] Traversing obstacles and predicted trajectories of the obstacles to obtain obstacles that appear on the vehicle's path planning;

[0158] The obstacles appearing on the planned path of the ego vehicle are projected onto the ST graph, and the lateral distance from the predicted trajectory of the obstacle close to the ego vehicle to the planned path of the ego vehicle is calculated. The ST graph is used to characterize the relationship between the length of the planned path and time. The obstacles close to the ego vehicle are obtained through screening.

[0159] Based on the pre-configured lateral risk estimation model, according to the speed of the obstacle at the target time t and the lateral distance between the obstacle and the vehicle, the maximum speed at which the vehicle can safely pass at the target time t is calculated;

[0160] When calculating the penalty value of the ego vehicle at the target time t, the penalty value of the ego vehicle speed exceeding the maximum speed at which the ego vehicle can safely pass at the target time t is added, and the decision trajectory is output.

[0161] The above application Figure 2 The method performed by the decision-making device 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 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete 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 may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in a decoding processor. The software module can be located in a storage medium mature in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. 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.

[0162] The electronic device may also perform Figure 2 The method executed by the decision-making device in Figure 1The functions of the illustrated embodiment will not be described in detail in the embodiments of the present application.

[0163] 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 2 The method executed by the decision-making device in the illustrated embodiment is specifically used to execute:

[0164] Traversing obstacles and predicted trajectories of the obstacles to obtain obstacles that appear on the vehicle's path planning;

[0165] The obstacles appearing on the planned path of the ego vehicle are projected onto the ST graph, and the lateral distance from the predicted trajectory of the obstacle close to the ego vehicle to the planned path of the ego vehicle is calculated. The ST graph is used to characterize the relationship between the length of the planned path and time. The obstacles close to the ego vehicle are obtained through screening.

[0166] Based on the pre-configured lateral risk estimation model, according to the speed of the obstacle at the target time t and the lateral distance between the obstacle and the vehicle, the maximum speed at which the vehicle can safely pass at the target time t is calculated;

[0167] When calculating the penalty value of the self-vehicle at the target time t, the penalty value of the self-vehicle speed exceeding the maximum speed of the self-vehicle to safely pass at the target time t is increased, and the decision trajectory is output. It should be understood by those skilled in the art that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention can 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.

[0168] 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.

[0169] 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.

[0170] 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.

[0171] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0172] 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.

[0173] 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.

[0174] 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.

[0175] 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.

[0176] 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 decision-making method, wherein: The method comprises: Traversing obstacles and predicted trajectories of the obstacles to obtain obstacles that appear on the vehicle's path planning; The obstacles appearing on the planned path of the ego vehicle are projected onto the ST graph, and the lateral distance from the predicted trajectory of the obstacle close to the ego vehicle to the planned path of the ego vehicle is calculated. The ST graph is used to characterize the relationship between the length of the planned path and time. The obstacles close to the ego vehicle are obtained through screening. Based on the pre-configured lateral risk estimation model, according to the speed of the obstacle at the target time t and the lateral distance between the obstacle and the vehicle, the maximum speed at which the vehicle can safely pass at the target time t is calculated; When calculating the penalty value of the ego vehicle at the target time t, the penalty value of the ego vehicle speed exceeding the maximum speed at which the ego vehicle can safely pass at the target time t is added, and the decision trajectory is output.

2. The method of claim 1, wherein: The method further comprises: Determine whether the speed of the ego vehicle at the target time t exceeds the maximum speed at which the ego vehicle can safely pass, and if so, penalize the decision-making process of the ego vehicle that exceeds the maximum speed; Otherwise, there is no penalty for the car's decision-making process.

3. The method of claim 2, wherein: The method further comprises: According to the predicted trajectory of the obstacle close to the ego vehicle at the lateral distance, the lateral distance between the ego vehicle and the predicted trajectory of the obstacle close to the ego vehicle at the lateral distance is obtained, and the lateral distance is added to the decision process.

4. The method of claim 1, wherein: The pre-configured horizontal risk estimation model includes: Based on the TTC model or RSS model, lateral collision risk modeling is performed; According to the lateral collision risk modeling result, the collision time TTC=L / (V1+V1') is obtained; Wherein, L is the lateral distance between the vehicle and the front vehicle, V1 is the lateral speed of the front vehicle, V1' is the lateral speed of the vehicle, and TTC is used to determine whether the current lateral distance and lateral speed are safe.

5. The method of claim 4, wherein: The pre-configured lateral risk estimation model further includes: The vehicle has a fixed steering wheel vibration error during driving, and the error is transmitted to the vehicle chassis and manifested as the deflection angle of the vehicle's front wheels. θ = f * δ; Wherein, f is the transmission ratio from the vehicle steering wheel to the front wheel steering, and δ is the vehicle steering wheel angle; Based on the deviation angle of the front wheels of the vehicle, additional lateral speed is obtained V1=v*θ Wherein, V1 is proportional to the vehicle's forward speed v, and the proportionality coefficient is the deflection angle θ of the vehicle's front wheels.

6. The method of claim 5, wherein: The pre-configured lateral risk estimation model further includes: The lateral distance is determined by the distance between the obstacle outline and the vehicle body. A time threshold is set according to the obstacle type, where the time threshold is the shortest time calculated according to the TTC or RSS model based on the current lateral speed and lateral distance.

7. The method of claim 6, wherein: The pre-configured lateral risk estimation model further includes: According to the longitudinal speed v, lateral distance L and time threshold TTC of the obstacle, the maximum safe longitudinal speed of the vehicle is calculated. Among them, the TTC time threshold and the front wheel turning angle of the other vehicle are determined according to the type of obstacle, and the deflection angle of the front wheel of the vehicle is an adjustable parameter.

8. A decision-making device, wherein: The device comprises: A traversal processing module, used to traverse obstacles and predicted trajectories of the obstacles to obtain obstacles that appear on the vehicle path planning; A projection calculation module is used to project obstacles appearing on the planned path of the ego vehicle onto the ST graph, and calculate the lateral distance from the predicted trajectory of the obstacle near the ego vehicle to the planned path of the ego vehicle. The ST graph is used to characterize the relationship between the length of the planned path and time. The obstacles near the ego vehicle are obtained through screening. A speed calculation module, for calculating a maximum speed at which the vehicle can safely pass at the target time t based on a pre-configured lateral risk estimation model and the speed of the obstacle at the target time t and the lateral distance between the obstacle and the vehicle; The decision module is used to increase the penalty value of the vehicle speed exceeding the maximum speed at which the vehicle can safely pass at the target time t when calculating the penalty value of the vehicle at the target time t, and output a decision trajectory.

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

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

Images