A brake control method and device of a vehicle, a vehicle and a medium

Abstract

The application discloses a brake control method and device of a vehicle, the vehicle and a medium. The method comprises the following steps: acquiring attribute information of a traffic participant; the attribute information is used for representing at least a motion behavior of the traffic participant; determining a stability weight used for representing a motion stability degree of the traffic participant according to the attribute information; determining a collision risk level of the vehicle and the traffic participant according to the stability weight and a plurality of relative motion parameters; the relative motion parameters are used for representing a relative motion trend between the vehicle and the traffic participant; and performing a graded brake control on the vehicle with different decelerations based on the collision risk level. Thus, the motion stability degree of the traffic participant can be evaluated based on the stability weight, and the jumping problem caused by sensor noise, target occlusion and the like can be avoided. In addition, the graded brake control is performed based on the collision risk level, so that unnecessary emergency braking can be reduced under the premise of ensuring the driving safety of the vehicle.

Description

Technical Field

[0001] This application relates to the field of vehicle braking control technology, and in particular to a vehicle braking control method, device, vehicle, and medium. Background Technology

[0002] With the rapid development of autonomous driving technology, driverless systems have been widely applied in scenarios such as urban roads and complex intersections. Current driverless systems mainly include a perception module, a prediction module, a decision-making and planning module, and a control module. Among them, the perception module is used to identify traffic participants around the vehicle (such as vehicles, pedestrians, non-motorized vehicles, etc.), the prediction module is used to predict the future movement trajectories of traffic participants, and the decision-making and planning module generates the vehicle's driving trajectory and behavioral commands based on the perception and prediction results.

[0003] However, in actual use, current autonomous driving systems are affected by factors such as sensor noise, target occlusion, and fluctuations in detection algorithms. The position, speed, category, and other attributes of the same traffic participant may change in consecutive frames, leading to misjudgments of their behavioral intentions by the system.

[0004] Furthermore, when the predicted trajectory intersects with the planned path of the autonomous vehicle, the autonomous driving system often immediately triggers emergency braking or even comes to a complete stop to avoid potential collisions, thus causing a large number of "unnecessary emergency braking" or "unnecessary stops". This not only reduces passenger comfort and traffic efficiency, but may also increase secondary risks such as rear-end collisions.

[0005] Therefore, how to reduce unnecessary sudden braking or stopping while ensuring vehicle safety and improving passenger comfort is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] In view of this, one aspect of this application provides a braking control method for a vehicle, the method comprising: Obtain attribute information of traffic participants during the vehicle's journey along the planned path; the attribute information is used to characterize the movement behavior of the traffic participants at least. Based on the attribute information, a stability weight is determined to characterize the degree of motion stability of the traffic participants; The collision risk level between the vehicle and the traffic participant is determined based on the stability weight and multiple relative motion parameters; the relative motion parameters are used to characterize the relative motion trend between the vehicle and the traffic participant. Based on the collision risk level, the vehicle is subjected to graded braking control with different deceleration rates.

[0007] Optionally, the attribute information is time-series based sequence information, including a position coordinate sequence, a motion speed sequence, and a category label sequence; The step of determining the stability weights used to characterize the degree of motion stability of the traffic participants based on the attribute information includes: Based on the location coordinate sequence, determine the position jump amplitude of the traffic participant within a specified number of consecutive frames; Based on the motion velocity sequence, determine the rate of velocity change between the consecutive frames; Based on the category label sequence, determine the number of times the category label changes between the consecutive frames; The stability weight is determined based on the position jump amplitude, the velocity change rate, and the number of changes.

[0008] Optionally, determining the stability weight based on the position jump amplitude, the velocity change rate, and the number of changes includes: If the position jump amplitude is greater than the amplitude threshold, the velocity change rate is greater than the change rate threshold, or the number of changes is greater than the number threshold, the corresponding attribute information will be marked as an unstable item; wherein, the amplitude threshold is determined based on the maximum physical velocity corresponding to the preset category label; Count the target number of the unstable terms; Based on the pre-constructed mapping relationship, the weight value corresponding to the target quantity is used as the stability weight; the target quantity is positively correlated with the stability weight; the mapping relationship is the correspondence between the quantity of unstable items and the weight value.

[0009] Optionally, determining the stability weight based on the position jump amplitude, the velocity change rate, and the number of changes includes: The number of frames in which the position jump amplitude is greater than an amplitude threshold, the number of frames in which the velocity change rate is greater than a change rate threshold, and the number of frames in which the number of changes is greater than a number threshold are determined. The largest of the first frame number, the second frame number, and the third frame number is taken as the target abnormal frame number; Obtain the maximum allowed number of abnormal frames from the specified quantity; and determine the ratio of the target number of abnormal frames to the maximum number of abnormal frames; The stability weight is determined based on the ratio.

[0010] Optionally, the relative motion parameters include the minimum collision time and relative speed between the vehicle and the traffic participant, and the angle between the direction of motion of the traffic participant and the path normal of the vehicle; The step of determining the collision risk level between the vehicle and the traffic participant based on the stability weight and multiple relative motion parameters includes: Obtain the preset minimum risk threshold; The basic collision risk value between the vehicle and the traffic participant is determined based on the minimum collision time, the relative speed, and the included angle. Based on the stability weight, the minimum risk base and the basic collision risk value are weighted and summed to obtain the comprehensive collision risk value; The collision risk level is determined based on the comprehensive collision risk value; wherein, the larger the comprehensive collision risk value, the higher the collision risk level, and the higher the degree of collision risk.

[0011] Optionally, determining the basic collision risk value between the vehicle and the traffic participant based on the minimum collision time, the relative speed, and the included angle includes: An initial collision risk value is determined based on the minimum collision time; wherein, the smaller the minimum collision time, the larger the initial collision risk value, and the higher the degree of collision risk. The initial collision risk value is corrected by using the relative speed and the included angle to obtain the basic collision risk value; wherein, if the relative speed is less than 0, the initial collision risk value is reduced; if the included angle is greater than the included angle threshold, the initial collision risk value is increased.

[0012] Optionally, the collision risk level includes a low risk level, a medium risk level, and an emergency risk level; the collision risk of the emergency risk level is higher than that of the medium risk level, and the collision risk of the medium risk level is higher than that of the low risk level. The step of performing graded braking control on the vehicle with different decelerations based on the collision risk level includes: When the collision risk level is the low risk level, control the vehicle to maintain the current driving state; When the collision risk level is medium risk, the vehicle is controlled to drive defensively; defensive driving is a driving mode that involves slight deceleration and tracking observation. When the collision risk level is the emergency risk level, the vehicle is controlled to perform emergency braking.

[0013] Optionally, controlling the vehicle to perform defensive driving includes: Within the tracking and observation window, the vehicle is controlled to decelerate at a specified slight deceleration, and it is determined whether the collision risk level changes abruptly. If the risk level changes to low and the duration reaches the preset duration, the vehicle is controlled to exit the defensive driving mode and accelerate at a specified slow acceleration rate to smoothly accelerate to the set cruise speed. If the emergency risk level changes, control the vehicle to perform emergency braking. If no change occurs, return to the steps of controlling the vehicle to decelerate at a comfortable deceleration and determining whether the collision risk level has changed.

[0014] Optionally, the medium-risk level includes multiple sub-risk levels; wherein, the higher the level of the sub-risk level, the smaller the tracking and observation window; and the greater the stability weight, the smaller the tracking and observation window. Within the tracking observation window, if the collision risk level jumps from a low sub-risk level to a high sub-risk level, the comfort deceleration is increased. Within the tracking observation window, if there is no intersection between the vehicle and the traffic participant's trajectory, it indicates that the collision risk has been eliminated, and the vehicle is controlled to exit the defensive driving mode.

[0015] Optionally, the braking control method for the vehicle further includes: If the vehicle performs emergency braking and it is determined that the traffic participant does not pose a collision threat, the current scenario is marked as an unnecessary braking case. Collect scene feature information of the unnecessary braking cases; the scene feature information includes at least attribute information, the stability weight, the relative motion parameters, and the collision risk level; The unnecessary braking cases and the scene feature information are associated and stored to construct a historical scene database.

[0016] Optionally, the vehicle braking control method further includes: Construct a cost function for evaluating the quality of the vehicle's braking decisions; the cost function includes an unnecessary penalty term, and at least one of a comfort penalty term, a benefit penalty term, and a safety penalty term; the unnecessary penalty term is used to characterize the degree of contribution of unnecessary braking to the cost function; During the vehicle's operation, if the current scene can be matched with a historical scene in the historical scene database that meets the similarity criteria, the contribution of the unnecessary penalty term to the cost function is increased.

[0017] Another aspect of this application provides a braking control device for a vehicle, the device comprising: The attribute information acquisition module is used to acquire attribute information of traffic participants during the process of a vehicle traveling along a planned path; the attribute information is used to characterize the movement behavior of the traffic participants at least. The stability determination module is used to determine stability weights that characterize the degree of motion stability of the traffic participants based on the attribute information. The collision risk level determination module is used to determine the collision risk level between the vehicle and the traffic participant based on the stability weight and multiple relative motion parameters; the relative motion parameters are used to characterize the relative motion trend between the vehicle and the traffic participant. The braking control module is used to perform graded braking control of the vehicle with different deceleration rates based on the collision risk level.

[0018] Another aspect of this application provides a vehicle including a memory and a processor, the memory storing a computer program executable on the processor, the processor executing the computer program to implement the steps of a braking control method for the vehicle.

[0019] Another aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the braking control method for the vehicle.

[0020] The braking control method, device, vehicle, and medium provided in this application offer the following beneficial effects: By using stability weight analysis based on attribute information, the motion stability of traffic participants can be assessed, avoiding attribute jumps caused by sensor noise, target occlusion, and detection algorithm fluctuations, thereby reducing the misjudgment rate of traffic participant intentions. Furthermore, by combining stability weights and relative motion parameters, a comprehensive collision risk level is determined, avoiding aggressive strategies such as emergency braking upon vehicle-traffic participant intersection. Simultaneously, graded braking control based on collision risk levels, i.e., differentiated deceleration for different risk levels, reduces unnecessary sudden braking or stopping while ensuring vehicle driving safety, improves ride comfort, and reduces the risk of rear-end collisions. Attached Figure Description

[0021] Figure 1 A schematic flowchart illustrating a vehicle braking control method provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a vehicle braking control system provided in an embodiment of this application; Figure 3 A schematic diagram illustrating the principle of a vehicle braking control method provided in an embodiment of this application; Figure 4A schematic diagram illustrating the principle of determining stability weights provided in an embodiment of this application; Figure 5 A schematic diagram illustrating the principle of a vehicle braking control method provided in another embodiment of this application; Figure 6 A schematic diagram illustrating the result of a vehicle braking control device provided in an embodiment of this application; Figure 7 This is a structural schematic diagram of a vehicle provided in an embodiment of this application.

[0022] The attached diagram is labeled as follows: 20 is the perception module, 21 is the prediction module, 22 is the stability assessment module, 23 is the risk level determination module, 24 is the decision control module, 25 is the control execution module, 60 is the attribute information acquisition module, 61 is the stability determination module, 62 is the collision risk level determination module, 63 is the braking control module, 70 is the memory, 71 is the processor, 72 is the display screen, 73 is the input / output interface, 74 is the communication interface, 75 is the power supply, 76 is the communication bus, 701 is the computer program, 702 is the operating system, and 703 is the data. Detailed Implementation

[0023] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0024] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0025] Figure 1 This is a schematic flowchart of a vehicle braking control method provided in an embodiment of this application, as shown below. Figure 1 As shown, the method includes: S10: Obtain attribute information of traffic participants during the vehicle's journey along the planned path; the attribute information shall at least be used to characterize the movement behavior of the traffic participants. First, it should be noted that the vehicle braking control method provided in the embodiments of this application can be applied to the vehicle controller, domain controller or advanced driver assistance system (ADAS) controller of the target vehicle, and this application does not limit it.

[0026] In addition, it should be noted that the target vehicle may include, but is not limited to, sedans, sport utility vehicles (SUVs), multi-purpose vehicles (MPVs), off-road vehicles, pickup trucks, or other powered, non-rail-borne vehicles.

[0027] Figure 2 This is a schematic diagram of a vehicle braking control system provided in an embodiment of this application. For ease of understanding, in an optional embodiment, the execution subject is described below. Figure 2 The system shown is used as an example for illustration. Figure 2 As shown, the system includes a perception module 20, a prediction module 21, a stability assessment module 22, a risk level determination module 23, a decision control module 24, and a control execution module 25.

[0028] In a specific embodiment, when a vehicle travels along a planned path, it needs to perceive the movement status of surrounding traffic participants in real time to assess the collision risk between the vehicle and these participants, and take timely braking measures to avoid a collision and ensure the vehicle's driving safety. These traffic participants include, but are not limited to, motor vehicles, pedestrians, and non-motorized vehicles other than the vehicle itself.

[0029] Based on this Figure 2 The sensing module 20 shown collects attribute information of traffic participants in real time and stores the attribute information based on time series. The sensing module 20 may be composed of multiple sensors, including but not limited to lidar, cameras, etc., and this application does not limit its composition.

[0030] In a specific embodiment, attribute information is used to characterize the motion behavior of traffic participants. For example, it may include, but is not limited to, information such as the location coordinates, speed, and acceleration of traffic participants. It can reflect the current motion state of traffic participants in order to assess whether there is a risk of collision between the vehicle (automobile) and the traffic participants.

[0031] It should be noted that, in an optional embodiment, in order to provide a data basis for subsequent analysis of the stability of traffic participants, when the vehicle is driving on the planned path, it maintains time-series cached attribute information for each perceived traffic participant. For example, it can save the historical attribute information data of the 10 most recent historical frames. This application does not limit the number of frames stored.

[0032] It is worth noting that, in specific embodiments, there may be one or more traffic participants, and this application does not limit this.

[0033] S11: Determine the stability weights used to characterize the degree of motion stability of traffic participants based on the attribute information; Figure 3 This is a schematic diagram illustrating the principle of a vehicle braking control method provided in an embodiment of this application, as shown below. Figure 2 The perception module 20 transmits the perceived attribute information to the prediction module 21 so that the prediction module 21 can assess whether there is a collision risk between the vehicle and traffic participants.

[0034] Specifically, such as Figure 3 As shown, in one optional embodiment, the vehicle can obtain its current planned path in real time. Simultaneously, based on the attribute information of traffic participants collected by the perception module 20, the prediction module 21 can predict the movement trajectory of the traffic participants. When the planned path of the vehicle intersects with the movement trajectory of a traffic participant, it indicates a risk of collision between the vehicle and the traffic participant.

[0035] It should be noted that when predicting the movement trajectories of traffic participants, the prediction module 21 employs a model-based method (such as an interactive multi-model approach) or a learning-based method (such as a Transformer trajectory prediction network) to output the movement trajectory of each traffic participant within a specified future time period (such as 3 to 8 seconds). This application does not limit the prediction method or the specified time period.

[0036] In one optional embodiment, spatiotemporal collision detection is performed between the movement trajectory of each traffic participant and the planned path of the vehicle. This can be achieved using a time-axis-based interval overlap detection method. Specifically, for each future time point, the Euclidean distance between the vehicle's position and the positions of the traffic participants is calculated. If the Euclidean distance is less than a safe distance threshold (e.g., 3-5 meters, related to vehicle speed), it is determined that the vehicle and the traffic participant intersect, i.e., there is a risk of collision. If there is no intersection, it indicates that there is no risk of collision between the vehicle and the traffic participants within a specified time period. Figure 3 As shown, the system continues to monitor in real time whether there is a risk of collision between the vehicle and other road users in the driving path.

[0037] Based on the determination of collision risk, if a collision risk is determined, in order to avoid misjudgment of collision risk due to problems such as sensor noise and obstruction, i.e., instability of the detected traffic participants, in an optional embodiment, it is necessary to perform a stability judgment on the attribute information of the traffic participants.

[0038] It is understandable that, due to sensor noise, occlusion, or algorithm fluctuations, the attribute information of traffic participants may undergo non-physical jumps between consecutive frames (e.g., a stationary vehicle is misidentified as a moving vehicle, or the position of a pedestrian changes abruptly due to occlusion).

[0039] To address the aforementioned technical issues, it is necessary to assess the stability of traffic participants, such as... Figure 2 The sensing module 20 transmits the collected attribute information to the stability assessment module 22, which evaluates the stability of traffic participants in real time by calculating stability weights based on the attribute information. The stability weights reflect the degree of stability of the traffic participants' movements; a higher stability weight indicates more stable movement.

[0040] In one alternative embodiment, time-series analysis can be used to determine indicators such as position jump amplitude and velocity change rate based on attribute information, thereby determining whether the movement of traffic participants is stable, and then determining stability weights based on the stability determination results.

[0041] S12: Determine the collision risk level between the vehicle and traffic participants based on stability weights and multiple relative motion parameters; relative motion parameters are used to characterize the relative motion trend between the vehicle and traffic participants. like Figure 2 As shown, the prediction module 21 predicts the movement trajectory of traffic participants based on attribute information and determines whether there is a collision risk with a vehicle. If a collision risk exists, the prediction result is transmitted to the risk level determination module 23. At the same time, the stability assessment module 22 transmits the determined stability weights to the risk level determination module 23 so that the risk level determination module 23 can assess the current collision risk level.

[0042] It should be noted that, in an optional embodiment, steps S11 to S13 can be executed in real time to determine the collision risk level between the vehicle and traffic participants in real time, and to perform real-time graded braking control on the vehicle based on the collision risk level.

[0043] It is understandable that during vehicle operation, there is not a real-time collision risk with other road users. If the collision risk level were assessed in real time, it would inevitably lead to a significant waste of computational resources when there is no collision risk between the vehicle and other road users. Therefore, in another optional embodiment, in order to save computational resources, it can be determined whether there is a collision risk between the vehicle and other road users before executing step S11. If there is a collision risk, then steps S11 to S13 are executed to perform graded braking control on the vehicle, thereby improving the reliability of vehicle braking control.

[0044] Specifically, after obtaining the stability weights through step S11, such as Figure 3 As shown, the collision risk level is calculated by combining multiple relative motion parameters. In a specific embodiment, the relative motion parameters include, but are not limited to, the minimum time to collision (TTC) between the vehicle and the traffic participant, the relative speed, and the angle between the direction of motion of the traffic participant and the normal of the vehicle's path.

[0045] In one alternative embodiment, an initial collision risk value can be calculated based on multiple relative motion parameters. The initial collision risk value is then multiplied by a stability weight to obtain a final collision risk value, which is used to determine the collision risk level.

[0046] In one optional embodiment, the collision risk level can be divided into four levels: low risk, medium risk, high risk, or emergency risk. The risk level can be determined based on the final collision risk value. It should be noted that this application does not limit the number of levels.

[0047] S13: Based on the collision risk level, apply graded braking control to the vehicle with different deceleration rates.

[0048] In one optional embodiment, when performing graded braking control based on the collision risk level, if the risk level is low, the vehicle maintains its current driving state, issuing only a warning or taking no intervention. If the risk level is medium or high, the vehicle enters a defensive driving mode, performs slight deceleration (e.g., 0.15g) and enters a tracking observation window to continuously reassess the risk. If the risk level is emergency, the vehicle performs emergency braking (e.g., 0.8g~1.0g), bringing it to a complete stop if necessary.

[0049] It should be noted that graded braking control ensures vehicle safety while significantly reducing unnecessary emergency braking and stopping. Even in scenarios where the vehicle's planned path intersects with the movement trajectory of other road users, emergency braking will not be immediately triggered if the target is unstable or the relative movement trend indicates that the risk is manageable.

[0050] Therefore, the vehicle braking control method provided in this application, through stability weight analysis based on attribute information, can assess the motion stability of traffic participants, avoiding attribute jumps caused by sensor noise, target occlusion, and detection algorithm fluctuations, thereby reducing the misjudgment rate of traffic participant intentions. Furthermore, by combining stability weights and relative motion parameters, a comprehensive collision risk level is determined, avoiding aggressive strategies such as emergency braking upon vehicle-traffic participant intersection. Simultaneously, graded braking control is implemented based on the collision risk level, i.e., differentiated deceleration for different risk levels, thereby reducing unnecessary emergency braking or stopping while ensuring vehicle driving safety, improving ride comfort, and reducing the risk of rear-end collisions.

[0051] Figure 4 This is a schematic diagram illustrating the principle of determining stability weights according to an embodiment of this application. In an optional embodiment, the attribute information is time-series based sequence information, including a position coordinate sequence, a motion speed sequence, and a category label sequence. Based on attribute information, stability weights are determined to characterize the degree of motion stability of traffic participants, including: Based on the position coordinate sequence, determine the position jump amplitude of traffic participants in a specified number of consecutive frames; Determine the rate of change of velocity between consecutive frames based on the motion velocity sequence; Based on the category label sequence, determine the number of times the category label changes between consecutive frames; The stability weights are determined based on the position jump amplitude, velocity change rate, and number of changes.

[0052] In a specific embodiment, the attribute information collected by the perception module 20 integrates data from multiple sensors such as lidar, camera, and millimeter-wave radar, and outputs information such as the location coordinates, movement speed, movement acceleration, category label, size, timestamp, and confidence level of the category label of surrounding traffic participants. It also maintains a time-series cache for each detected traffic participant (such as saving the most recent 10 frames of historical data).

[0053] like Figure 2 As shown, the perception module 20 transmits the collected time-series-based attribute information to the stability assessment module 22, which performs real-time assessment of the motion stability of traffic participants. Specifically, it calculates the rate of change of attributes for each traffic participant between consecutive frames (including but not limited to position jump amplitude, speed change rate, category confidence fluctuation, etc.).

[0054] Furthermore, as an optional embodiment, the position jump amplitude, velocity change rate, and number of changes are compared with the corresponding thresholds to determine whether the collected attribute information is stable. Based on the stability determination result, the stability weight is determined, that is, the stability flag (stable / unstable) and the stability weight (with a value range of 0 to 1) are output.

[0055] In a specific embodiment for determining the magnitude of position jumps between consecutive frames, such as Figure 4 As shown, read the timing buffer and calculate according to formula (1): (1) in, The magnitude of the position change. For the first The location coordinates of traffic participants at each moment. For the first The location coordinates of traffic participants at a given time. In an optional embodiment, if there are multiple frames with location jumps exceeding a threshold between a specified number of consecutive frames, then... Figure 4 As shown, position coordinates can be marked as unstable terms.

[0056] In a specific embodiment for determining the rate of change of velocity, the stability of the motion velocity sequence in the attribute information can be determined by calculating the rate of change of velocity magnitude or direction between adjacent frames, see calculation formula (2): (2) in, For the first Rate of change of speed of traffic participants at each time point For the first The speed of traffic participants at any given moment For the first The speed of traffic participants at any given moment The frame interval duration.

[0057] like Figure 4 As shown, in one alternative embodiment, the rate of change of velocity When the rate of change exceeds a preset threshold, the speed of motion can be marked as an unstable term, or it can characterize the instability of the motion of traffic participants.

[0058] In a specific embodiment for determining the number of consecutive inter-frame category label changes, the number of times the category label changes (e.g., the category label changes from "vehicle" to "pedestrian") in the time-series data of N consecutive frames of category labels can be counted. If a specified number of consecutive inter-frame changes exceeds a preset threshold, for example, if the number of changes is greater than or equal to 3 in a 10-frame consecutive category label sequence, the category label can be marked as an unstable item.

[0059] Furthermore, a stability weight is determined comprehensively based on the position jump magnitude, velocity change rate, and number of changes. In one optional embodiment, if any one of the three items is determined to be unstable, the traffic participant is marked as an "unstable target," and the stability weight can be set to a lower value (e.g., 0.2). If all items are stable, the stability weight can be set to a higher value (e.g., 0.9). In another optional embodiment, a more granular weight assignment can be made based on the proportion of abnormal frames or the number of abnormal types. This application does not limit the specific method for determining the stability weight.

[0060] Therefore, the vehicle braking control method provided in this application evaluates the temporal stability of multiple dimensions such as position, speed, and category to obtain non-physical jumps in the perception information, providing reliable data support for subsequent determination of collision risk level, suppressing unnecessary braking caused by perception noise from the data source, and improving vehicle safety and ride comfort.

[0061] Based on the above embodiments, as an optional embodiment, stability weights are determined according to the position jump amplitude, velocity change rate, and number of changes, including: If the position change amplitude exceeds the amplitude threshold, the velocity change rate exceeds the change rate threshold, or the number of changes exceeds the number threshold, the corresponding attribute information will be marked as an unstable item; whereby the amplitude threshold is determined based on the maximum physical velocity corresponding to the preset category label; The target number of unstable items is counted; Based on the pre-built mapping relationship, the weight value corresponding to the target quantity is used as the stability weight; the target quantity is positively correlated with the stability weight; the mapping relationship is the correspondence between the quantity of unstable items and the weight value.

[0062] In a specific embodiment, such as Figure 4 As shown, based on the above embodiments, if the position jump amplitude When, the position coordinates are marked as unstable terms, where, The maximum physical speed corresponding to the category label of the traffic participant. The frame interval duration This is for location noise tolerance.

[0063] Based on the above calculation method, for example, if the category label of the traffic participant is pedestrian, and the maximum physical speed of the pedestrian is 3 m / s, and the frame interval is 0.1s, then the amplitude threshold can be set to 0.36 m. If the traffic participant is a vehicle, and the maximum physical speed of the vehicle is 30 m / s, then the amplitude threshold can be set to 3.6 m.

[0064] Furthermore, if the rate of change of velocity is greater than the rate of change threshold, i.e. Then the velocity is marked as an unstable term, where, This represents the maximum physical acceleration corresponding to the category label of a traffic participant. If the number of changes exceeds a threshold, the category label is marked as unstable.

[0065] Then, see Figure 4 The number of unstable targets is counted. It is understood that, in the embodiments of this application regarding location coordinates, movement speed, and category labels, the target number ranges from 0 to 3. After obtaining the target number, a search is performed based on the target number within a pre-constructed mapping relationship to determine the weight matching the target number, and this weight is used as the final stability weight. Table 1 is a schematic table of a mapping relationship provided by an embodiment of this application. As shown in Table 1, in an optional embodiment, a correspondence can be established between unstable items and stability weights.

[0066] Table 1 is a schematic diagram of a mapping relationship. As shown in Table 1 and Figure 4 As shown in the specific embodiment, after determining the target number of unstable items, a search is performed in the above mapping relationship, thereby quickly determining the stability weight.

[0067] It should be noted that, in specific embodiments, this application does not limit the correspondence between the target quantity and the stability weight, nor the specific value of the stability weight; these can be set and modified according to actual business needs. Furthermore, as shown in the table above, the target quantity and the stability weight are negatively correlated; that is, the more unstable items there are, the lower the weight.

[0068] Therefore, by simplifying the stability judgment of traffic participants into independent stability counts, the stability weights can be quickly obtained by using a mapping relationship retrieval method, thereby improving the efficiency of vehicle braking control.

[0069] Based on the above embodiments, as another optional embodiment, stability weights are determined according to the position jump amplitude, velocity change rate, and number of changes, including: Determine the number of frames in which the position jump amplitude is greater than the amplitude threshold, the number of frames in which the velocity change rate is greater than the change rate threshold, and the number of frames in which the number of changes is greater than the number threshold; The largest of the first, second, and third frame counts is taken as the target abnormal frame count; Obtain the maximum allowed number of abnormal frames from a specified number; and determine the ratio of the target number of abnormal frames to the maximum number of abnormal frames; The stability weights are determined based on the ratio.

[0070] In one optional embodiment, stability weights are determined based on the number of anomalous frames. Specifically, the number of first frames where the position jump amplitude exceeds an amplitude threshold is determined, i.e., the specific number of frames in the time buffer where position jump anomalies occur. Simultaneously, the number of second frames where the velocity change rate exceeds a change rate threshold and the number of third frames where the number of category changes exceeds a number threshold are determined.

[0071] To make it easier to understand, an example will be given below. For instance, if the specified number of consecutive frames is 10 frames, there may be 3 frames with abnormal position, 2 frames with abnormal speed, and 1 frame with abnormal category. That is, the first frame has 3 frames, the second frame has 2 frames, and the third frame has 1 frame.

[0072] Furthermore, the largest of the first, second, and third frame counts is used as the target abnormal frame count, so that the dimension with the most severe anomaly serves as a benchmark reflecting the stability of traffic participants. Simultaneously, the maximum allowed abnormal frame count within a specified number (e.g., 10 frames) is obtained. The maximum abnormal frame count can be set empirically or based on different traffic participants (e.g., a smaller maximum abnormal frame count can be set for vehicles, and a larger one for pedestrians), and this application does not impose any limitations on this.

[0073] Finally, the stability weight is determined based on the ratio of the target number of abnormal frames to the maximum number of abnormal frames, as shown in formula (3): (3) in, For stability weights, For the target number of abnormal frames, This represents the maximum number of abnormal frames. It should be noted that, in a specific embodiment, the target number of abnormal frames... Less than the maximum number of abnormal frames .

[0074] Based on the above embodiments, as an optional embodiment, different severity coefficients can be assigned to different anomaly dimensions (including location coordinates, motion speed, and category labels), and then the weighted anomaly frame count can be calculated. The weighted anomaly frame count can be used as the target anomaly frame count in the above embodiments, and finally the stability weight can be determined by formula (3). .

[0075] Therefore, by statistically analyzing the proportion of abnormal frames in consecutive frames, changes in the attribute information of traffic participants can be perceived more accurately, thereby avoiding drastic fluctuations in weights caused by extreme jumps in single-frame data and ensuring the accuracy of stability weight calculation.

[0076] In one alternative embodiment, the relative motion parameters include the minimum collision time and relative velocity between the vehicle and the traffic participant, and the angle between the direction of motion of the traffic participant and the path normal of the vehicle. Based on stability weights and multiple relative motion parameters, the collision risk level between the vehicle and traffic participants is determined, including: Obtain the preset minimum risk threshold; The basic collision risk value between the vehicle and traffic participants is determined based on the minimum collision time, relative speed, and angle. Based on stability weights, the minimum risk base and the basic collision risk value are weighted and summed to obtain the comprehensive collision risk value; The collision risk level is determined based on the comprehensive collision risk value; the higher the comprehensive collision risk value, the higher the collision risk level, and the greater the degree of collision risk.

[0077] In a specific embodiment, to improve the accuracy of collision risk level assessment, this application establishes a multi-dimensional, multi-level dynamic risk assessment model. During the assessment process, not only the time to collision (TTC) is considered, but also the relative speed between the vehicle and traffic participants, and the angle between the traffic participants' direction of motion and the vehicle's path normal.

[0078] In this context, TTC refers to the time it takes for a vehicle and a traffic participant to reach the point of collision while traveling at their current speeds. A positive relative speed indicates that the vehicle and traffic participant are approaching each other, while a negative value indicates that they are moving away from each other. Furthermore, the angle between the two can assess the relative motion trend between the vehicle and traffic participant, i.e., whether they are moving in opposite directions, in the same direction, parallel, or crossing each other.

[0079] In one alternative embodiment, for each traffic participant, their relative motion to the vehicle is calculated. Specifically, the vehicle's position coordinates are denoted as... The driving speed is recorded as The location coordinates of traffic participants are set as The speed of motion is denoted as See formula (4) to solve for TTC: (4) Find the minimum positive solution of formula (4) to obtain TTC. If the above formula (4) has no solution or the solution is infinite, TTC is set to a preset value, such as 10 seconds.

[0080] For the relative velocity, in an alternative embodiment, it can be determined according to formula (5): (5) in, This is the angle between the direction of movement of the traffic participant and the direction of movement of the vehicle. Motion trend factor. Defined as the angle between the direction of movement of a traffic participant and the path normal of the vehicle, it is used to determine whether a traffic participant is moving across, in the same direction, or in the opposite direction.

[0081] In a specific embodiment for determining the collision risk level, firstly, a preset minimum risk baseline is obtained. (e.g., 0.1). In a specific embodiment, the minimum risk base. Ensure that the system retains a minimum level of risk awareness even if a traffic participant is deemed highly unstable.

[0082] Furthermore, based on the minimum collision time, relative speed, and angle, the basic collision risk value between the vehicle and traffic participants is determined. Basic collision risk value Through nonlinear mapping functions The calculation method for this function can be implemented using a calibration table or a neural network, and the output is a value between 0 and 1. This application does not limit the calculation method for this function.

[0083] For example, in an alternative embodiment, in an embodiment determined by a calibration table, when TTC < 2 seconds, =0.9. When 2≤TTC<4 seconds, Gradually decrease. When TTC ≥ 6 seconds =0.1.

[0084] Then, based on the stability weights determined in the above embodiments. For the lowest risk base and basic collision risk value The weighted summation yields the comprehensive collision risk value, as shown in formula (6): (6) in, This is the overall collision risk value.

[0085] In one alternative embodiment, the overall collision risk value is... A value < 0.3 indicates low risk, 0.3 ≤ R < 0.6 indicates medium risk, 0.6 ≤ R < 0.9 indicates high risk, and R ≥ 0.9 indicates emergency risk. Therefore, the comprehensive collision risk value... The larger the value, the higher the collision risk level, indicating a greater degree of collision risk.

[0086] Building upon the above embodiments, to further improve the accuracy of collision risk level assessment, in addition to the aforementioned relative motion parameters, a risk correction factor can be added. This risk correction factor may include, but is not limited to, the size of traffic participants and the ratio of the vehicle's current speed to the road speed limit. Therefore, in a specific embodiment, a basic collision risk value is determined based on the relative motion parameters and the risk correction factor, resulting in a more accurate comprehensive collision risk value.

[0087] Therefore, the vehicle braking control method provided in this application dynamically integrates stability weights with multi-dimensional relative motion parameters to calculate an accurate comprehensive collision risk value, thereby enabling a more precise classification of collision risk levels.

[0088] Based on the above embodiments, as an optional embodiment, the basic collision risk value between the vehicle and traffic participants is determined according to the minimum collision time, relative speed, and included angle, including: The initial collision risk value is determined based on the minimum collision time; the smaller the minimum collision time, the larger the initial collision risk value, which indicates a higher degree of collision risk. The initial collision risk value is corrected by relative speed and included angle to obtain the basic collision risk value; if the relative speed is less than 0, the initial collision risk value is reduced; if the included angle is greater than the included angle threshold, the initial collision risk value is increased.

[0089] In specific embodiments, the initial collision risk value can be determined using a piecewise linear function or a lookup table. For example, in a specific embodiment where the initial collision risk value is determined using a piecewise linear function, see formula (7): (7) in, This is the initial collision risk value.

[0090] As can be seen from formula (7), the minimum collision time The smaller the initial collision risk value The larger the value, the higher the collision risk, i.e., the minimum collision time. Compared with the initial collision risk value It shows a negative correlation.

[0091] Specifically, when the minimum collision time When the first interval [0,2) is included, the initial collision risk value is... The maximum value of 0.9 was reached. This is the minimum collision time. When the value belongs to the second interval [0,6), the initial collision risk value is... With minimum collision time It increases linearly as the minimum collision time decreases. Belongs to the third interval (i.e.) When the initial collision risk value is ≥6, The minimum value of 0.1 was obtained.

[0092] Initial collision risk value Based on the established parameters, further analysis is conducted using relative velocity and angle to determine the initial collision risk value. Make corrections to obtain the basic collision risk value. Specifically, if road users are moving away from vehicles, i.e., relative speed... This can reduce the initial collision risk value. If a traffic participant crosses the road at an angle greater than a threshold (e.g., angle...), the traffic participant will be crossing the road at an angle greater than a threshold threshold (e.g., angle...). When this happens, the initial collision risk value can be increased. .

[0093] Therefore, by establishing an initial risk benchmark through TTC and then correcting it using relative speed and direction angle, the basic collision risk value can reflect high-risk scenarios such as "rapid approach of the target" and "crossing", while also appropriately reducing the risk when the target moves away, thereby improving the fit between the basic collision risk value and the actual scenario, that is, improving the calculation accuracy of the basic collision risk value.

[0094] Figure 5 This is a schematic diagram of a vehicle braking control method provided in another embodiment of this application. As an optional embodiment, the collision risk level includes a low risk level, a medium risk level, and an emergency risk level; the collision risk of the emergency risk level is higher than that of the medium risk level, and the collision risk of the medium risk level is higher than that of the low risk level. Based on the collision risk level, the vehicle is subjected to graded braking control with different deceleration rates, including: When the collision risk level is low, control the vehicle to maintain the current driving state; When the collision risk level is medium risk, control the vehicle to drive defensively; defensive driving is a driving mode that involves slowing down slightly and keeping a close watch. When the collision risk level is classified as an emergency risk level, control the vehicle to apply emergency braking.

[0095] In one alternative embodiment, such as Figure 3 As shown, collision risk levels can be divided into three levels: low risk, medium risk, and emergency risk. The low risk level has the lowest collision risk, followed by the medium risk level, and the emergency risk level has the highest collision risk.

[0096] Different braking controls are set for different levels of collision risk. Specifically, in a specific embodiment, such as... Figure 2 As shown, the risk level determination module 23 determines the collision risk level between the current vehicle and traffic participants based on the stability weight transmitted by the stability assessment module 22, and further outputs the collision risk level to the decision control module 24, so that the decision control module 24 outputs acceleration or deceleration commands to the control execution module 25. Finally, the control execution module 25 sends the underlying control signal to the vehicle actuator, thereby realizing the braking control of the vehicle.

[0097] like Figure 3 and 5 As shown, when the collision risk level is low, it indicates that the traffic participants are far away, the TTC is large, and the actual risk is low. At this time, vehicle control is not intervened and the current driving state is maintained.

[0098] If the collision risk level is medium, indicating a relatively high risk, the vehicle needs to slow down, but unnecessary braking should be avoided. At the medium risk level, the system enters defensive driving mode. Defensive driving refers to a driving mode that involves slight deceleration and close observation.

[0099] In defensive driving, the vehicle is first controlled to slow down gently, specifically at a preset comfortable deceleration rate (e.g., 0.15g, approximately 1.47 m / s²), in order to limit the vehicle speed within the safe speed limit. While slowing down gently, the collision risk level between the vehicle and other road users is tracked and observed in real time to continuously monitor the status and risk changes of road users.

[0100] In one alternative embodiment, the safe speed limit It can be set to ,in, The vehicle's current speed. For road speed limits, for safety factors In one alternative embodiment, the safety factor It can be set to 0.8, which means limiting the vehicle speed to within 0.8 times the road speed limit.

[0101] If the collision risk level is medium risk, indicating an extremely high collision risk, in order to ensure vehicle safety, the maximum braking force (such as 0.8g~1.0g) can be used for braking control. If necessary, the vehicle should be brought to a complete stop, and a braking warning should be issued to the following vehicles (such as illuminating the hazard lights).

[0102] It should be noted that, in one optional embodiment, for the medium-risk level of defensive driving, the medium-risk level can be further divided into multiple levels. Among the multiple levels, the higher the collision risk level, the greater the degree of collision risk, and the greater the corresponding deceleration value of the light deceleration.

[0103] For example, the medium risk level can be divided into medium-low risk level and medium-high risk level. The collision risk of medium-high risk level is higher than that of medium-low risk level, and the deceleration value corresponding to medium-high risk level is greater than that of medium-low risk level (e.g., the deceleration value of medium-low risk level is 0.1g, and the deceleration value of medium-high risk level is 0.25g).

[0104] It should be noted that this application does not limit the number of different levels within the medium-risk category; the number can be set according to actual business needs.

[0105] Therefore, different braking strategies are set for different collision risk levels. Smooth driving is maintained when the risk is low, slight deceleration is used to gain observation time when the risk is medium, and full braking is only used when there is an emergency risk. This ensures safety and minimizes unnecessary emergency braking, thus improving vehicle ride comfort while ensuring vehicle safety.

[0106] Based on the above embodiments, as an optional embodiment, controlling the vehicle to perform defensive driving includes: Within the tracking and observation window, control the vehicle to decelerate at a specified slight deceleration and determine whether the collision risk level changes abruptly; If the risk level drops to low and remains so for the preset duration, the vehicle will exit defensive driving mode and accelerate at a specified rate to gradually reach the set cruising speed. If the risk level changes to an emergency, control the vehicle to apply emergency braking. If no change occurs, return to the steps of controlling the vehicle to decelerate at a comfortable deceleration and determining whether the collision risk level has changed.

[0107] In a specific embodiment of defensive driving, a tracking and observation window is set, for example, the tracking and observation window can be set to last for 3 seconds or 30 control cycles. This application does not limit the setting of the tracking window. After the vehicle enters defensive driving mode, the status of traffic participants and changes in collision risk are monitored in real time within the tracking and observation window.

[0108] Specifically, such as Figure 5 As shown, in defensive driving mode, the vehicle is controlled to decelerate at a specified light deceleration speed, while continuously acquiring the latest attribute information and expected cross-prediction results of traffic participants at a higher frequency (e.g., 20Hz, compared to the normal 10Hz), and dynamically reassessing the collision risk level, i.e., determining whether the collision risk level changes during this process.

[0109] See Figure 5 If the vehicle transitions from a medium-risk level to a low-risk level, and the transition lasts for a preset duration (e.g., 1 second), it indicates that the collision risk has been eliminated, and the vehicle can proceed with the planned path. Therefore, the vehicle is then deactivated from defensive driving and accelerated at a specified gradual rate (e.g., 0.1g) to smoothly reach the set cruising speed.

[0110] In an optional embodiment, the exit process employs a first-order low-pass filter for a smooth transition, avoiding discomfort caused by rapid acceleration and further improving vehicle ride comfort. Specifically, see formula (8) for a smooth transition: (8) in, This refers to the deceleration after smoothing (i.e., the deceleration of the current instruction after filtering). The target deceleration output based on the current state of the system (i.e., the raw, unfiltered expected deceleration).

[0111] like Figure 5 As shown, if the risk level jumps from medium to high, it indicates that the collision risk is increasing as the vehicle moves. In order to ensure the safety of the vehicle, the vehicle should be controlled to brake suddenly to avoid collision with other road users.

[0112] In another possible embodiment, see Figure 5 If the collision risk level does not change abruptly, then it is sufficient to continuously monitor the status of traffic participants and the collision risk level in real time, that is, to continue to perform comfortable deceleration and repeatedly determine the collision risk level.

[0113] In summary, in this embodiment, during the tracking and observation period, if the threat from the traffic participant is eliminated (e.g., turning away or slowing down to yield), the system smoothly exits the defensive mode and resumes normal driving. If the threat from the traffic participant persists but does not escalate, the system maintains slight deceleration and continuous observation. If the threat from the traffic participant escalates to an emergency threshold, the system immediately triggers emergency braking.

[0114] Thus, by "slightly decelerating," the system gains time for observation and decision-making, and by "tracking and observing the status," it achieves continuous tracking of dynamic changes in risk, avoiding the two extremes of "stopping as soon as there is risk" or "sudden braking only when risk is imminent." This makes the system behavior closer to the driver's defensive driving strategy, ensuring safety while avoiding unnecessary sudden braking.

[0115] Based on the above embodiments, as an optional embodiment, the medium-risk level includes multiple sub-risk levels; wherein, the higher the level of the sub-risk level, the smaller the tracking and observation window; the greater the stability weight, the smaller the tracking and observation window. Within the tracking and observation window, if the collision risk level jumps from a low sub-risk level to a high sub-risk level, increase the comfort deceleration. Within the tracking and observation window, if there is no intersection between the vehicle and traffic participants' trajectories, it indicates that the collision risk has been eliminated, and the vehicle is controlled to exit defensive driving.

[0116] In a specific embodiment, Figure 2 The decision control module 24 shown A three-state finite state machine can be implemented. In a specific embodiment, normal driving can be set to state 0. When the vehicle is in state 0, it indicates that there is no potential risk or the risk has been eliminated, and normal planning and control can be performed. Furthermore, defensive driving mode can be set to state 1. When the vehicle is in state 1, it indicates that there is a non-emergency risk, and it can perform slight deceleration and enter tracking and observation mode. Simultaneously, emergency braking can be set to state 2. When the vehicle is in state 2, it indicates that there is an emergency risk, and full braking is required.

[0117] In a specific embodiment, when the vehicle transitions from state 0 to state 1, it indicates that a non-emergency risk (medium or high risk) has been detected. When the vehicle transitions from state 1 to state 2, it indicates that the risk has escalated to an emergency risk during the monitoring period. When the vehicle transitions from state 1 to state 0, it indicates that the risk has been eliminated during the monitoring period. When the vehicle transitions from state 2 to state 0, it indicates that emergency braking has been completed and the risk has been eliminated.

[0118] Based on the above, as an optional embodiment, different tracking observation window lengths can be dynamically set according to different scenarios. Specifically, if the medium-risk level includes multiple sub-risk levels, such as low-medium-risk and medium-high-risk levels, defensive driving is performed in all of these sub-risk levels. However, to further improve vehicle safety and passenger comfort, as an optional embodiment, different tracking windows can be set for different sub-risk levels.

[0119] Understandably, the higher the risk level, the faster the decision-making needs to be made, and the less time can be allowed to wait; therefore, the smaller the tracking and observation window should be. At the same time, the larger the stability weight, the more reliable the perception of traffic participants, meaning that the confidence level of traffic participants' state perception is higher, and there is no need for long-term verification. In this case, the tracking and observation window can be shortened.

[0120] For example, for low to medium risk levels, the tracking and observation window can be set to 3 seconds, while for medium to high risk levels, it can be set to 2 seconds. If the stability weight is lower, the tracking and observation window can be appropriately extended. The specific value can be set according to the actual scenario, and this application does not limit it.

[0121] Based on the above, during defensive driving, if the collision risk level changes from a low sub-risk level to a high sub-risk level within the tracking and observation window, for example, from a low-to-medium risk level to a medium-to-high risk level, then the comfort deceleration needs to be increased, for example, from 0.15g to 0.25g, to control the collision risk more quickly. The deceleration adjustment can be based on a preset ratio (e.g., an increase of 0.05g for each level) or through continuous function mapping; this application does not limit this.

[0122] Within the tracking and observation window, if there is no intersection between the vehicle and the traffic participant's trajectory, that is, the predicted trajectory of the traffic participant no longer intersects with the vehicle's planned path or the traffic participant has left, it indicates that the collision risk has been eliminated, and the system can control the vehicle to exit defensive driving and resume normal driving.

[0123] It should be noted that this application does not limit the number of sub-risk levels included in the medium-risk level. For example, it can be set to 3 different levels, each corresponding to a different tracking observation window length, deceleration value and exit threshold, thereby achieving more precise defensive driving control.

[0124] Therefore, the vehicle braking control method provided in this application further subdivides the medium-risk level into multiple sub-risk levels and configures different observation windows and decelerations for different sub-risk levels, ensuring that the vehicle can perform more precise braking control according to the severity of the risk, thereby further improving the vehicle's ride comfort.

[0125] In an optional embodiment, the vehicle braking control method provided in this application further includes: If a vehicle brakes suddenly and it is determined that no other road users pose a collision threat, the current scenario will be marked as an unnecessary braking case. Collect scene feature information for non-essential braking cases; scene feature information should include at least attribute information, stability weights, relative motion parameters, and collision risk level; Unnecessary braking cases and scene feature information are linked and stored to build a historical scene database.

[0126] Based on the above embodiments, in order to reduce unnecessary emergency braking of vehicles, in a specific embodiment, unnecessary braking cases from historical braking cases can be stored so that unnecessary emergency braking can be further avoided when the same case occurs in the future.

[0127] Specifically, if a vehicle applies emergency braking during its historical driving process, and it is determined after braking that the corresponding road user did not actually pose a collision threat—for example, the road user swerved away, slowed down to yield, or came to a complete stop during braking—then the vehicle's emergency braking is considered unnecessary braking, and this scenario can be marked as an unnecessary braking case.

[0128] Furthermore, scene feature information for non-essential braking cases is collected. This scene feature information includes, but is not limited to, traffic participant attribute information (time series such as location coordinates, speed, and category labels) for a period of time prior to braking, calculated stability weights, relative motion parameters used (TTC, relative speed, and angle), and the final collision risk level. In addition, scene feature information may also include case record timestamps, geographical location, and weather conditions.

[0129] Finally, unnecessary braking cases and their corresponding scene feature information are associated and stored to construct a historical scene database. It should be noted that the database can be a vector database or a relational database to support subsequent scene similarity retrieval; this application does not limit the database type.

[0130] Based on the above embodiments, as an optional embodiment, the vehicle braking control method further includes: Construct a cost function to evaluate the quality of vehicle braking decisions; the cost function includes an unnecessary penalty term, and at least one of a comfort penalty term, a benefit penalty term, and a safety penalty term; the unnecessary penalty term is used to characterize the degree of contribution of unnecessary braking to the cost function; During vehicle operation, if the current scene can be matched with a historical scene in the historical scene database that meets the similarity criteria, the contribution of unnecessary penalty terms in the cost function is increased.

[0131] In a specific embodiment, a cost function is constructed to evaluate the quality of vehicle braking decisions. This cost function includes at least an unnecessary penalty term, and at least one of a comfort penalty term, a benefit penalty term, and a safety penalty term. The unnecessary penalty term characterizes the degree to which unnecessary braking contributes to the cost function.

[0132] In an optional embodiment, the cost function includes a non-essential penalty term, a comfort penalty term, a benefit penalty term, and a safety penalty term, and the cost function is given by formula (9): (9) in, The cost function value, For comfort penalty items, As a benefit-based penalty item, As a security penalty item, This is a non-essential penalty.

[0133] In an alternative embodiment, non-essential penalty terms Defined as: (10) In formula (10), The penalty coefficient is... This is an indicator function. When emergency braking in historical scenario cases in historical data is marked as unnecessary, the indicator function takes a value of 1; otherwise, it takes a value of 0. This refers to the acceleration during emergency braking in the corresponding scenario.

[0134] During vehicle operation, for the currently encountered scene, a similarity search is performed in the historical scene database. If the similarity between the current scene and one or more scenes in the historical scene database meets preset conditions, such as the Euclidean distance of the feature vectors being less than a threshold, or the cosine similarity being greater than a similarity threshold, the system will proceed accordingly.

[0135] At this point, the scenario is deemed likely to lead to unnecessary braking, thus increasing unnecessary penalties. The degree to which a factor contributes to the cost function makes the planner inclined to choose no braking or a lighter braking strategy, thereby reducing the value of the cost function. This gradually reduces the system's conservative tendencies.

[0136] It should be noted that this applies to increasing unnecessary penalties. One approach could be to increase the penalty coefficient. Alternatively, the indicator function I_brake_unnecessary can be activated and multiplied by a weight coefficient greater than 1. This application does not limit the method of increasing the weight.

[0137] In a specific embodiment, such as Figure 2 As shown, an "unnecessary penalty term" is added to the cost function of the decision control module 24. "Non-essential penalty items" This is used to quantify unnecessary braking behaviors (including emergency braking and stopping) performed due to excessive conservatism in the system, enabling the system to adaptively adjust the decision boundaries through online learning or offline calibration.

[0138] Specifically, the system records historical data showing that, after braking in specific scenarios, traffic participants did not actually pose a real threat. Furthermore, based on this historical data, the system dynamically adjusts the thresholds or weighting coefficients in the risk assessment model, causing the system to tend to choose a smoother, defensive driving strategy rather than aggressive braking in similar scenarios.

[0139] Thus, the system can "remember" which braking actions are unnecessary and automatically reduce its braking tendency in similar future scenarios, reducing conservative bias at the source of decision-making. Furthermore, it works in conjunction with the stability assessment, risk classification, and tracking observation mechanisms described in the above embodiments to effectively suppress unintended braking.

[0140] Table 2 is a schematic diagram of the working timing of a vehicle braking control method provided in the embodiments of this application. In order to make the technical solution provided by this application clearer to those skilled in the art, examples will be given below in conjunction with Table 2.

[0141] Table 2. Timing diagram of a vehicle braking control method In one alternative embodiment, the traffic participant is a motor vehicle. When the vehicle is traveling at 36 km / h (10 m / s) in the straight lane, a lateral motor vehicle approaches the intersection from the right 50 meters ahead. The predicted trajectory intersects with the vehicle's path, meaning there is a risk of collision between the two.

[0142] At this point, referring to Table 2, the vehicle transitions from normal driving at time T1 to defensive driving at time T2, and performs slight deceleration and continuous risk monitoring within the tracking and observation window until time T5, when the collision risk level jumps to a low risk level, meaning the collision risk is eliminated. At this point, the vehicle decelerates to zero, begins to accelerate, and the vehicle returns to normal driving.

[0143] Throughout the entire process, the vehicle did not trigger emergency braking, which ensured safety and avoided the impact of sudden braking on passenger comfort.

[0144] In the above embodiments, the braking control method for a vehicle has been described in detail. This application also provides an embodiment of a braking control device for a vehicle.

[0145] Figure 6 This is a schematic diagram of a vehicle braking control device provided in an embodiment of this application, as shown below. Figure 6 As shown, the device includes: The attribute information acquisition module 60 is used to acquire the attribute information of traffic participants during the process of the vehicle traveling along the planned path; the attribute information is used to characterize the movement behavior of the traffic participants at least. The stability determination module 61 is used to determine the stability weights that characterize the degree of motion stability of traffic participants based on the attribute information. The collision risk level determination module 62 is used to determine the collision risk level between the vehicle and traffic participants based on stability weights and multiple relative motion parameters; the relative motion parameters are used to characterize the relative motion trend between the vehicle and traffic participants. Braking control module 63 is used to perform graded braking control of the vehicle with different deceleration rates based on the collision risk level.

[0146] Figure 7 This is a schematic diagram of the structure of a vehicle provided in an embodiment of this application, such as... Figure 7 As shown, the vehicle includes: a memory 70 for storing computer programs; The processor 71 is used to execute a computer program to implement the steps of the vehicle braking control method as described in the above embodiments.

[0147] The vehicles provided in this application embodiment may include, but are not limited to, sedans, sport utility vehicles (SUVs), multi-purpose vehicles (MPVs), off-road vehicles, pickup trucks, or other power-driven, non-rail-borne vehicles.

[0148] The processor 71 may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor 71 may be implemented using at least one of the following hardware forms: Digital Signal Processor (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). The processor 71 may also include a main processor and a coprocessor. The main processor, also known as the Central Processing Unit (CPU), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 71 may integrate a Graphics Processing Unit (GPU), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, the processor 71 may also include an Artificial Intelligence (AI) processor, which is used to handle computational operations related to machine learning.

[0149] The memory 70 may include one or more computer-readable storage media, which may be non-transitory. The memory 70 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In this embodiment, the memory 70 is used to store at least the following computer program 701, which, after being loaded and executed by the processor 71, is capable of implementing the relevant steps of the vehicle braking control method disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory 70 may also include an operating system 702 and data 703, and the storage method may be temporary or permanent storage. The operating system 702 may include Windows, Unix, Linux, etc. The data 703 may include, but is not limited to, relevant data involved in the vehicle braking control method.

[0150] In some embodiments, the vehicle may also include a display screen 72, an input / output interface 73, a communication interface 74, a power supply 75, and a communication bus 76.

[0151] Those skilled in the art will understand that Figure 7 The structure shown does not constitute a limitation on the braking control method of the vehicle and may include more or fewer components than shown.

[0152] The vehicle provided in this application includes a memory and a processor. When the processor executes the program stored in the memory, it can implement the braking control method of the vehicle in the above embodiments.

[0153] It should be noted that although the operations are depicted in a specific order in the accompanying drawings, this should not be construed as requiring these operations to be performed in the specific order shown or sequentially, or requiring all illustrated operations to be performed to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. Furthermore, the separation of various system modules and components in the above embodiments should not be construed as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

Claims

1. A braking control method for a vehicle, characterized in that, The method includes: Obtain attribute information of traffic participants during the vehicle's journey along the planned path; the attribute information is used to characterize the movement behavior of the traffic participants at least. Based on the attribute information, a stability weight is determined to characterize the degree of motion stability of the traffic participants; The collision risk level between the vehicle and the traffic participant is determined based on the stability weight and multiple relative motion parameters; the relative motion parameters are used to characterize the relative motion trend between the vehicle and the traffic participant. Based on the collision risk level, the vehicle is subjected to graded braking control with different deceleration rates.

2. The vehicle braking control method as described in claim 1, characterized in that, The attribute information is time-series based sequence information, and includes a position coordinate sequence, a motion speed sequence, and a category label sequence; The step of determining the stability weights used to characterize the degree of motion stability of the traffic participants based on the attribute information includes: Based on the location coordinate sequence, determine the position jump amplitude of the traffic participant within a specified number of consecutive frames; Based on the motion velocity sequence, determine the rate of velocity change between the consecutive frames; Based on the category label sequence, determine the number of times the category label changes between the consecutive frames; The stability weight is determined based on the position jump amplitude, the velocity change rate, and the number of changes.

3. The vehicle braking control method as described in claim 2, characterized in that, Determining the stability weight based on the position jump amplitude, the velocity change rate, and the number of changes includes: If the position jump amplitude is greater than the amplitude threshold, the velocity change rate is greater than the change rate threshold, or the number of changes is greater than the number threshold, the corresponding attribute information will be marked as an unstable item; wherein, the amplitude threshold is determined based on the maximum physical velocity corresponding to the preset category label; Count the target number of the unstable terms; Based on the pre-constructed mapping relationship, the weight value corresponding to the target quantity is used as the stability weight; the target quantity is positively correlated with the stability weight; the mapping relationship is the correspondence between the quantity of unstable items and the weight value.

4. The vehicle braking control method as described in claim 2, characterized in that, Determining the stability weight based on the position jump amplitude, the velocity change rate, and the number of changes includes: The number of frames in which the position jump amplitude is greater than an amplitude threshold, the number of frames in which the velocity change rate is greater than a change rate threshold, and the number of frames in which the number of changes is greater than a number threshold are determined. The largest of the first frame number, the second frame number, and the third frame number is taken as the target abnormal frame number; Obtain the maximum allowed number of abnormal frames from the specified quantity; and determine the ratio of the target number of abnormal frames to the maximum number of abnormal frames; The stability weight is determined based on the ratio.

5. The vehicle braking control method as described in claim 1, characterized in that, The relative motion parameters include the minimum collision time and relative speed between the vehicle and the traffic participant, as well as the angle between the direction of motion of the traffic participant and the path normal of the vehicle. The step of determining the collision risk level between the vehicle and the traffic participant based on the stability weight and multiple relative motion parameters includes: Obtain the preset minimum risk threshold; The basic collision risk value between the vehicle and the traffic participant is determined based on the minimum collision time, the relative speed, and the included angle. Based on the stability weight, the minimum risk base and the basic collision risk value are weighted and summed to obtain the comprehensive collision risk value; The collision risk level is determined based on the comprehensive collision risk value; wherein, the larger the comprehensive collision risk value, the higher the collision risk level, and the higher the degree of collision risk.

6. The vehicle braking control method as described in claim 5, characterized in that, The determination of the basic collision risk value between the vehicle and the traffic participant based on the minimum collision time, the relative speed, and the included angle includes: An initial collision risk value is determined based on the minimum collision time; wherein, the smaller the minimum collision time, the larger the initial collision risk value, and the higher the degree of collision risk. The initial collision risk value is corrected by using the relative speed and the included angle to obtain the basic collision risk value; wherein, if the relative speed is less than 0, the initial collision risk value is reduced; if the included angle is greater than the included angle threshold, the initial collision risk value is increased.

7. The vehicle braking control method as described in claim 1, characterized in that, The collision risk levels include low risk, medium risk, and emergency risk. The collision risk of the emergency risk level is higher than that of the medium risk level, and the collision risk of the medium risk level is higher than that of the low risk level. The step of performing graded braking control on the vehicle with different decelerations based on the collision risk level includes: When the collision risk level is the low risk level, control the vehicle to maintain the current driving state; When the collision risk level is medium risk, the vehicle is controlled to drive defensively; defensive driving is a driving mode that involves slight deceleration and tracking observation. When the collision risk level is the emergency risk level, the vehicle is controlled to perform emergency braking.

8. The vehicle braking control method as described in claim 7, characterized in that, Controlling the vehicle for defensive driving includes: Within the tracking and observation window, the vehicle is controlled to decelerate at a specified slight deceleration, and it is determined whether the collision risk level changes abruptly. If the risk level changes to low and the duration reaches the preset duration, the vehicle is controlled to exit the defensive driving mode and accelerate at a specified slow acceleration rate to smoothly accelerate to the set cruise speed. If the emergency risk level changes, control the vehicle to perform emergency braking. If no change occurs, return to the steps of controlling the vehicle to decelerate at a comfortable deceleration and determining whether the collision risk level has changed.

9. The vehicle braking control method as described in claim 8, characterized in that, The medium-risk level includes multiple sub-risk levels; wherein, the higher the level of the sub-risk level, the smaller the tracking and observation window; the greater the stability weight, the smaller the tracking and observation window. Within the tracking observation window, if the collision risk level jumps from a low sub-risk level to a high sub-risk level, the comfort deceleration is increased. Within the tracking observation window, if there is no intersection between the vehicle and the traffic participant's trajectory, it indicates that the collision risk has been eliminated, and the vehicle is controlled to exit the defensive driving mode.

10. The vehicle braking control method as described in claim 1, characterized in that, The method further includes: If the vehicle performs emergency braking and it is determined that the traffic participant does not pose a collision threat, the current scenario is marked as an unnecessary braking case. Collect scene feature information of the unnecessary braking cases; the scene feature information includes at least attribute information, the stability weight, the relative motion parameters, and the collision risk level; The unnecessary braking cases and the scene feature information are associated and stored to construct a historical scene database.

11. The vehicle braking control method as described in claim 10, characterized in that, The method further includes: Construct a cost function for evaluating the quality of the vehicle's braking decisions; the cost function includes an unnecessary penalty term, and at least one of a comfort penalty term, a benefit penalty term, and a safety penalty term; the unnecessary penalty term is used to characterize the degree of contribution of unnecessary braking to the cost function; During the vehicle's operation, if the current scene can be matched with a historical scene in the historical scene database that meets the similarity criteria, the contribution of the unnecessary penalty term to the cost function is increased.

12. A braking control device for a vehicle, characterized in that, The device includes: The attribute information acquisition module is used to acquire attribute information of traffic participants during the process of a vehicle traveling along a planned path; the attribute information is used to characterize the movement behavior of the traffic participants at least. The stability determination module is used to determine stability weights that characterize the degree of motion stability of the traffic participants based on the attribute information. The collision risk level determination module is used to determine the collision risk level between the vehicle and the traffic participant based on the stability weight and multiple relative motion parameters; the relative motion parameters are used to characterize the relative motion trend between the vehicle and the traffic participant. The braking control module is used to perform graded braking control of the vehicle with different deceleration rates based on the collision risk level.

13. A vehicle comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the braking control method for the vehicle according to any one of claims 1 to 11.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the braking control method for the vehicle according to any one of claims 1 to 11.

Images