Vehicle brake control method and electronic device

Abstract

The vehicle brake control method and the electronic device provided by the application relate to the intelligent driving technical field. The method comprises the following steps: acquiring a first risk parameter for indicating a rearward collision risk between a host vehicle and a rear target and a second risk parameter for indicating a forward collision risk between the host vehicle and a front target when the host vehicle is in a deceleration working condition; comparing the rearward collision risk and the forward collision risk based on the first risk parameter and the second risk parameter; and if the rearward collision risk is greater than the forward collision risk, controlling the host vehicle to perform an active intervention operation matched with the level of the rearward collision risk. Thus, the application compares the priority of the rearward collision risk and the forward collision risk to accurately identify the real approaching state of the vehicle behind the host vehicle lane, avoids the situation of not timely warning or false alarm, and improves the reliability and user experience of the warning system.

Description

Technical Field

[0001] This application relates to the field of intelligent driving technology, and more specifically, to a vehicle braking control method and electronic device. Background Technology

[0002] Rear-end collisions account for a high percentage of all traffic accidents, especially when a vehicle is decelerating or braking. When a vehicle approaches rapidly from behind, a rear-end collision is very likely to occur, causing not only vehicle damage and personal injury, but also exacerbating traffic congestion and the risk of secondary accidents. This is a major hidden danger to road safety.

[0003] To address the aforementioned risk of rear-end collisions, most vehicles are equipped with Rear Collision Warning (RCW) systems. These systems typically rely on a single sensor, such as millimeter-wave radar or a rearview camera, to detect the distance and speed of objects behind the vehicle. When the system determines that a vehicle is approaching from behind, it issues a basic warning to the driver to reduce the occurrence of rear-end collisions.

[0004] However, most existing rear collision warning systems use a single sensor solution of radar or camera, which makes it impossible for the vehicle to accurately identify the actual approach of vehicles behind in the same lane during deceleration. This results in untimely warnings or false alarms, affecting the reliability of the warning system and reducing the user experience. Summary of the Invention

[0005] This application provides a vehicle braking control method and electronic device. The method compares the priority of rear collision risk and forward collision risk to accurately identify the actual approaching state of vehicles behind in the same lane, avoids untimely warnings or false alarms, and improves the reliability of the warning system and user experience.

[0006] In a first aspect, a vehicle braking control method is provided, the method comprising: when the vehicle is in a deceleration condition, acquiring a first risk parameter representing the risk of a rearward collision between the vehicle and a rearward target, and a second risk parameter representing the risk of a forward collision between the vehicle and a forward target; comparing the rearward collision risk and the forward collision risk based on the first risk parameter and the second risk parameter; if the rearward collision risk is greater than the forward collision risk, controlling the vehicle to perform an active intervention operation that matches the level of the rearward collision risk.

[0007] In the above technical solution, when the vehicle is decelerating, it integrates millimeter-wave radar and rearview camera multi-sensor perception, and obtains a first risk parameter to characterize the rearward collision risk between the vehicle and the target behind by using three-level target attribution determination, dynamic threshold of collision time, and multi-component confidence screening. Simultaneously, it calculates a second risk parameter to characterize the forward collision risk between the vehicle and the target ahead by combining forward sensing information. This overcomes the shortcomings of insufficient detection accuracy and inaccurate target attribution of a single sensor, reducing false alarms and missed alarms. Then, based on the front and rear risk weights obtained by quantifying the first and second risk parameters, it further assesses the rearward and forward collision risks. Priority comparison avoids the drawbacks of traditional solutions that only assess rear risks and ignore frontal driving safety. If the rear collision risk is determined to be greater than the front collision risk, the risk level is divided according to the rear collision time interval, and layered intervention operations such as graded audible and visual warnings, driving guidance, and gradient active braking are matched and executed. This achieves accurate identification, reasonable decision-making, and graded active protection of rear-end collision risks under deceleration conditions. This solves the technical problems of traditional single-sensor solutions, such as inaccurate identification, delayed warning timing, easy generation of false alarms, inability to take into account both front and rear safety, and lack of matched active intervention. It improves the detection reliability, decision-making rationality, and driving experience of rear collision warning.

[0008] Secondly, a vehicle braking control device is provided, comprising: an acquisition module, configured to acquire, when the vehicle is in a deceleration condition, a first risk parameter representing the risk of a rearward collision between the vehicle and a rearward target, and a second risk parameter representing the risk of a forward collision between the vehicle and a forward target; a comparison module, configured to compare the rearward collision risk and the forward collision risk based on the first risk parameter and the second risk parameter; and a control module, configured to control the vehicle to perform an active intervention operation matching the level of the rearward collision risk if the rearward collision risk is greater than the forward collision risk.

[0009] Thirdly, an electronic device is provided, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, causing the electronic device to perform the methods of the first aspect or any possible implementation thereof.

[0010] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of this application more easily understood, specific embodiments of this application are given below. Attached Figure Description

[0011] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application; Figure 2 A flowchart illustrating a vehicle braking control method provided in this application embodiment. Figure 1 ; Figure 3 A flowchart illustrating a vehicle braking control method provided in this application embodiment. Figure 2 ; Figure 4 A flowchart illustrating a vehicle braking control method provided in this application embodiment. Figure 3 ; Figure 5 This is a schematic diagram of the structure of a vehicle braking control device provided in an embodiment of this application. Detailed Implementation

[0012] The technical solutions of this application will now be described clearly and in detail with reference to the accompanying drawings. In the description of the embodiments of this application, unless otherwise stated, " / " signifies "or," for example, A / B can mean A or B. "And / or" in the text merely describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Furthermore, in the description of the embodiments of this application, "multiple" refers to two or more.

[0013] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0014] Rear-end collisions are a frequent type of traffic accident. When a vehicle is decelerating and braking, a rapidly approaching vehicle from behind can easily cause a rear-end collision, resulting in personal injury and property damage, as well as secondary accidents and traffic congestion. Currently, most vehicles are equipped with rear collision warning systems. These systems mostly use single sensors such as millimeter-wave radar or rearview cameras, which can only detect the distance and relative speed of rear targets and provide only basic warning functions. Existing single-sensor detection solutions are limited in dimensionality and lack multi-sensor fusion detection mechanisms. They cannot determine the attribution of rear targets within the same lane and cannot accurately identify the actual approaching state of the following vehicle. At the same time, they lack logic for quantifying and comparing front and rear risks and tiered intervention strategies, making it impossible to reconcile the conflicting decisions of avoiding collisions with the vehicle in front and preventing rear-end collisions when the vehicle is decelerating. This can easily lead to problems such as delayed warnings, false alarms, and missed warnings, resulting in insufficient system reliability and adaptability to different operating conditions. Based on these issues, this application proposes a vehicle braking control scheme to solve the above problems.

[0015] The vehicle braking control method and electronic equipment of this application will be described in detail below with reference to the accompanying drawings and through multiple embodiments.

[0016] Figure 1 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 1 As shown, the electronic device 100 may include a processor 110 and a memory 120.

[0017] The memory 120 stores machine-executable instructions that can be executed by the processor 110. When the electronic device 100 is running, these machine-executable instructions are executed. The processor 110 and the memory 120 communicate via a bus. The processor 110 can execute the machine-executable instructions to implement a vehicle braking control method.

[0018] The memory 120, processor 110, and various bus components are electrically connected directly or indirectly to enable data transmission or interaction. For example, these components can be electrically connected to each other via one or more communication buses or signal lines. The memory 120 includes at least one software functional module, which is stored or embedded in the operating system (OS) of the electronic device in the form of software or firmware. This software functional module includes at least one executable module. The processor 110 is used to execute the executable modules stored in the memory 120, such as the software functional modules and computer programs included in a vehicle braking control method.

[0019] The memory 120 may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.

[0020] The electronic device 100 can be selected according to the actual situation; for example, it can be selected as an in-vehicle controller or a computer device. Furthermore, the electronic device 100 has software capable of executing vehicle braking control methods.

[0021] The vehicle braking control method provided in this application embodiment can be executed by a processor in the electronic device 100. The vehicle braking control method provided in this application embodiment will be explained further below. Figure 2 A flowchart illustrating a vehicle braking control method provided in this application embodiment. Figure 1 .like Figure 2 As shown, the method may include: S210. When the vehicle is decelerating, acquire a first risk parameter representing the risk of a rearward collision between the vehicle and a target behind it, and a second risk parameter representing the risk of a forward collision between the vehicle and a target ahead it.

[0022] Rear collision risk refers to the potential danger of a rear-end collision when a vehicle is decelerating and a target vehicle in the same lane rapidly approaches it from behind. It is characterized by the rear collision time (TTC_rear) and the rear risk weight (W_1), and is influenced by relative distance, relative speed, vehicle deceleration intensity, lane affiliation, ambient weather, and road type. Forward collision risk refers to the potential danger of a head-on rear-end collision or collision between the vehicle and a target vehicle in front when the vehicle is decelerating. It is characterized by the forward collision time (TTC_front) and the forward risk weight (W_2), and is constrained by the relative motion of the vehicles in front and behind and the vehicle's braking intensity. The first risk parameter is used to quantify the characteristic parameters representing the rear collision risk between the vehicle and a target vehicle behind it. It includes at least the rear collision time (TTC_rear) and the rear risk weight (W_1), and serves as the basis for determining the rear risk level, the front-rear risk game, and tiered intervention. The second risk parameter is a characteristic parameter used to quantify the risk of a forward collision between the vehicle and the target ahead. It includes at least the forward collision time TTC_front and the forward risk weight W_2. It is used to compare the weights with the first risk parameter to prioritize overall driving safety.

[0023] In one possible implementation, the vehicle collects real-time deceleration data and brake pedal opening rate of change data via the onboard CAN (Controller Area Network) bus or onboard accelerometer, and determines the operating condition according to a preset deep deceleration judgment threshold rule. This preset deep deceleration judgment threshold rule includes three criteria; meeting any one of them indicates the vehicle has entered a deceleration state: First, the primary judgment: the vehicle's real-time deceleration is greater than or equal to 2.5 m / s²; second, the first auxiliary judgment: the vehicle's real-time deceleration is greater than or equal to 1.5 m / s² and the time to collision (TTC) is less than or equal to 2.0 seconds; third, the second auxiliary judgment: the brake pedal change rate is greater than or equal to 30% / s, used to predict that the vehicle is about to enter a deep deceleration state. The above judgment process continuously samples and calculates in real time. Only after determining that the vehicle meets any of the above deceleration conditions can the subsequent collection and calculation of front and rear target risk parameters be initiated.

[0024] Since this vehicle is equipped with a multi-sensor perception system consisting of millimeter-wave radar and a rear-view camera, the millimeter-wave radar outputs the distance d, relative speed v_rel, and azimuth angle θ of the object behind in real time. After distortion correction and ROI (Region of Interest) cropping, the rear-view camera completes multi-layer convolutional feature extraction, feature pyramid fusion, detection head bounding box regression, and Softmax (flexible maximum activation function) category probability output through a lightweight preset target detection model to obtain the type and pixel coordinate box of the object behind. At the same time, the left and right lane line position information is output through a preset lane line segmentation model. Then, based on the distance d, relative speed v_rel, azimuth angle θ, type of the object behind, and left and right lane line position information, the target comprehensive confidence C is calculated using a five-component weighted method and the following formula (1).

[0025] C = 0.25×stability + 0.30×matching degree + 0.20×classification + 0.15×motion + 0.10×trajectory formula (1) In the above formula (1), stability refers to radar detection stability, which is used to characterize the degree of continuous and stable occurrence of the target. It is calculated by the ratio of the number of times the target appears in the past 10 frames to 10. Matching degree refers to spatial matching degree, which is used to characterize the alignment between the radar target projection point and the visual frame. If the projection point falls within the visual frame, it scores a full score of 1. If it does not fall within the frame, the score is attenuated according to the distance from the projection point to the nearest visual frame. Classification refers to target classification reliability, which is used to characterize the credibility of the visual model in determining whether the target is a vehicle. The vehicle category Softmax probability output by the visual model is directly used. If it is not a vehicle category, it is assigned a value of 0. Motion refers to motion consistency, which is used to characterize whether the target's motion acceleration conforms to the vehicle's physical motion constraint range. When the target's longitudinal acceleration is not greater than 8 m / s² and its lateral acceleration is not greater than 4 m / s², it scores a full score of 1. If it exceeds this constraint range, the score is linearly attenuated to 0 based on the difference. The trajectory refers to the continuity of historical trajectories, used to characterize the smoothness of the target's historical motion trajectory. It is evaluated based on the distance-time fitting residuals. The smaller the residuals, the higher the trajectory smoothness. The score is assigned according to the exponential decay rule.

[0026] Valid rear objects are selected based on the overall confidence score C. If the overall confidence score C ≥ 0.70, the target is considered a high-confidence target and can directly participate in subsequent risk decisions. If C ≤ 0.40 < 0.70, the target is considered a medium-confidence target, and the collision time to collision (TTC) trigger threshold is tightened, typically by 0.5 seconds. If C < 0.40, the target is considered a low-confidence target and is directly removed from the calculation. If either the millimeter-wave radar or the rear-view camera fails, the corresponding confidence component is set to zero, and the remaining weights are renormalized before continuing to calculate the overall confidence score C. When there is inconsistency between radar and visual detection results, a redundancy-tolerant strategy of radar priority followed by visual verification is used.

[0027] The target attribution for this lane is determined through a three-tiered system of lateral offset Δy = d × tan(θ), target lateral velocity motion consistency verification, and visual lane line boundary constraints. A maximum of three candidate targets for this lane are retained per frame, with the target having the smallest Time-to-Cross (TTC) selected as the rear target. The target lateral velocity motion consistency verification means that within a 1-second window, if the target's lateral velocity |v_lat| < 0.5 m / s and |Δy| shows no significant drift, it is retained; if the target's lateral velocity > 1.0 m / s, it is considered to be changing lanes and is not included in the risk assessment. Visual lane line boundary constraints involve projecting the radar target onto the image; if it falls within the left or right lane lines, it is ultimately marked as a valid target for this lane. If the camera cannot extract the lane lines, the first two criteria take effect directly.

[0028] The collision time TTC is calculated based on the following formula (2). Then, the collision time TTC is dynamically thresholded by combining the vehicle speed, road type, weather, and the type of object behind the vehicle. Finally, the rear collision time TTC_rear is calculated.

[0029] TTC = d ÷ v_rel formula (2) Then, calculate the rear risk weight W_1 using the following formula (3).

[0030] W_1 = min(1, (2.5 ÷TTC_rear)×0.8) formula (3) The rear collision time TTC_rear and the rear risk weight W_1 are used as the first risk parameters to characterize the rear collision risk, thus completing the acquisition.

[0031] The vehicle collects the relative distance and relative speed information between itself and the target ahead in real time through the forward-facing onboard sensors. The forward collision time TTC_front is calculated according to the calculation principle of the above formula (2). Combined with the real-time deceleration of the vehicle and the driving status of the vehicle ahead, the forward risk weight W_2 is calculated according to the preset mature model logic. Hard constraints are set, that is, when the forward collision time TTC_front ≤ 1.5 seconds or the vehicle deceleration ≥ 3.5m / s², the forward risk weight W_2 is directly assigned to 1. The forward collision time TTC_front and the forward risk weight W_2 are updated in real time with the system's operation frequency of 5 times / second throughout the process. This set of parameters is used as the second risk parameter to characterize the forward collision risk, so as to complete the synchronous acquisition.

[0032] It should be noted that, in order to objectively quantify the simultaneous forward collision risk and rear-end collision risk faced by the vehicle under deceleration conditions, and to avoid the shortcomings of existing rear-end warning systems that only focus on the risk of rear approach while ignoring the safety of the vehicle in front, by separately obtaining the first risk parameter representing the risk of rear collision and the second risk parameter representing the risk of forward collision, a unified assessment benchmark that can be quantified and compared for both front and rear risks can be established. This provides reliable data input and calculation basis for subsequent front and rear risk weight game, safety priority decision, graded warning and active braking intervention.

[0033] S220. Based on the first risk parameter and the second risk parameter, compare the rearward collision risk and the forward collision risk.

[0034] The first risk parameter includes at least: rear collision time TTC_rear and rear risk weight W_1; the second risk parameter includes at least: front collision time TTC_front and front risk weight W_2.

[0035] In one possible implementation, the rear risk weight W_1 calculated based on the above formula (3) is directly assigned to 1 when the rear collision time TTC_rear ≤ 1.2s; and directly assigned to 0 when the rear collision time TTC_rear > 4.0s.

[0036] At the same time, according to the preset rules, the forward risk weight W_2 is calculated using the following formula (4). Formula (4) in, The real-time deceleration of this vehicle is collected via CAN bus or accelerometer, in m / s².

[0037] When the forward collision time TTC_front ≤ 1.5s or the vehicle's deceleration ≥ 3.5m / s2, the forward risk weight W_2 is forcibly set to 1.

[0038] The system synchronously updates and caches the solved rear risk weights W_1 and front risk weights W_2 in real time, operating at a fixed calculation cycle of 5 times per second. The rear risk weight W_1 quantifies the level of rear collision risk, while the front risk weight W_2 quantifies the level of front collision risk. Based on this, the system directly compares the numerical values ​​of the rear risk weights W_1 and W_2 to determine the priority of rear and front collision risks, thus establishing a judgment on the relative levels of the two risks. This provides the sole basis for deciding whether to activate the rear-level tiered intervention strategy. The entire comparison process achieves risk balancing solely through the comparison of objective parameter values, matching the actual scenario of front-rear safety conflicts under vehicle deceleration conditions while ensuring the reproducibility and implementability of the decision logic, meeting the real-time and reliability requirements of onboard active safety control.

[0039] It should be noted that the comparison process does not directly compare the collision time (TTC), but rather compares the normalized and quantified rear risk weight W_1 and forward risk weight W_2. If the rear risk weight W_1 > the forward risk weight W_2, then the rear collision risk is determined to be superior; if the forward risk weight W_2 ≥ the rear risk weight W_1, then the forward collision risk is determined to be higher or the two are equivalent.

[0040] S230. If the risk of a rear-end collision is greater than the risk of a front-end collision, then control the vehicle to perform active intervention operations that match the level of the rear-end collision risk.

[0041] In a possible implementation, when it is determined by comparing the front and rear risk weights that the rear risk weight W_1 is greater than the front risk weight W_2, that is, the rear collision risk is greater than the front collision risk, first, the rear collision risk level is defined according to the time to collision (TTC_rear) of the main target behind in this lane, and it is divided according to a preset interval. Among them, low risk corresponds to 2.5 s < TTC_rear ≤ 4.0 s; medium risk corresponds to 1.5 s < TTC_rear ≤ 2.5 s; high risk corresponds to TTC_rear ≤ 1.5 s. The corresponding risk level is matched according to the interval in which the actually detected TTC_rear falls, and then the preset hierarchical response and vehicle control logic are called according to different risk levels to perform active intervention operations that match their risk levels one by one. Among them, only background data monitoring is performed at the low risk level, and no vehicle control and audible and visual prompt actions that can be perceived by humans are output; at the medium risk level, four intervention actions are synchronously performed, namely, voice broadcast "Please gently step on the brake to prompt the vehicle behind", the instrument and HUD (Head Up Display) interface display a warning icon approaching from behind, the vehicle's hazard lights are controlled to switch to a 4 Hz high-frequency flashing mode, and at the same time, a control suggestion is output to the driver, guiding to adopt a pulse-like gentle braking method of 3 to ⑤ times per second, and the brake lights are frequently lit to convey a warning message to the vehicle behind. When it is determined to be at the high risk level, on the basis of reusing all the intervention actions at the medium risk level, the active assisted braking hierarchical control logic is further triggered, and the braking deceleration is applied in three stages according to the threshold corresponding to the TTC_rear. For example, when TTC_rear ≤ 1.5 s, it enters the first stage, and a slight braking of 0.5 to 1.0 m / s² is actively applied to light the brake lights to warn the vehicle behind; when TTC_rear ≤ 0.8 s and the front safety condition is established, it enters the second stage, and a medium braking of 2.0 to 3.0 m / s² is applied; when TTC_rear ≤ 0.5 s, it enters the third stage, and the rear automatic emergency braking (RAEB) is activated, with a maximum deceleration of 6 to 8 m / s² output; at the same time, a hard safety constraint is set. If the detected time to collision in front (TTC_front) ≤ 1.0 s, all active braking interventions are immediately prohibited, and the risk of collision with the vehicle in front is preferentially avoided. All active intervention actions follow the subsequent smooth exit mechanism, and smoothly exit the control according to the logic of light fading, voice ending, visual prompt fading out, and brake pressure ramping linearly releasing after the rear risk is解除, ensuring the driving comfort of the whole vehicle and the rigor and safety of the intervention logic.

[0042] Exemplarily, if the rear collision risk is less than or equal to the front collision risk, a braking inhibition instruction is generated to inhibit the execution of the active intervention operation.

[0043] It should be noted that there is an unclear "⑤" in the original text which is retained as is in the translation. Also, the "解除" in the original Chinese seems to be an incomplete or incorrect expression, and the translation is adjusted as best as possible based on the context.In one possible implementation, if the rear collision risk is less than or equal to the front collision risk (i.e., the rear risk weight W_1 is less than or equal to the front risk weight W_2), the vehicle controller immediately generates and issues a braking suppression command. This command blocks the execution authority of all subsequent rear active intervention operations, including at least: suppressing voice prompts, instrument HUD icon warnings, 4Hz high-frequency hazard light control, and pulse braking warnings under medium-risk conditions; and suppressing the activation process of phased active auxiliary braking, moderate-intensity braking intervention, and rear automatic emergency braking under high-risk conditions. The system only retains basic rear warning prompts and does not change the driver's original braking control strategy or the vehicle's deceleration status. Prioritizing the safety of the vehicle in front, the system blocks rear graded intervention and forced braking intervention at the command level, preventing excessive deceleration to avoid a rear-end collision and thus avoiding a collision with the vehicle in front. Furthermore, if TTC_front ≤ 1.0s is detected, all rear braking intervention channels are further locked, achieving objective judgment of front and rear risk priorities and reliable suppression control of intervention commands.

[0044] It should be noted that, in this application, vehicle braking control should be interpreted broadly. It includes not only the process of the braking system actively establishing hydraulic or electrical braking force to bring the vehicle to a stop, but also the entire dynamic process of the driver pressing the brake pedal and the vehicle responding to the braking system's deceleration request. The related deceleration condition specifically refers to the operating state where the vehicle's longitudinal acceleration is negative under the aforementioned braking control logic. Furthermore, it must be emphasized that the deceleration addressed in this application does not refer to passive deceleration caused by factors such as wind resistance or coasting after releasing the accelerator, but specifically to active deceleration behavior triggered by the intervention of the vehicle's braking system or the driver's braking intention. The core of this application lies in how to implement graded auxiliary intervention by coordinating the trade-off between front and rear collision risks when the vehicle is in this braking-related deceleration condition.

[0045] The vehicle braking control method provided by this application, when the vehicle is in a deceleration working condition, fuses millimeter-wave radar and rear-view camera multi-sensor perception, and through the three-level attribution determination of the target in the current lane, the dynamic threshold of the time to collision, and the screening of multi-component confidence, obtains a first risk parameter used to characterize the rear collision risk between the vehicle and the rear target. At the same time, it combines the forward sensing information to calculate a second risk parameter used to characterize the forward collision risk between the vehicle and the front target, so as to overcome the defects of insufficient detection accuracy and inaccurate target attribution of a single sensor, reduce false alarms and missed alarms; then, based on the front and rear risk weights quantified by the first risk parameter and the second risk parameter, compare the priorities of the rear collision risk and the forward collision risk, avoiding the disadvantages of traditional solutions that only evaluate the rear risk alone and ignore the safety of the vehicle ahead; if it is determined that the rear collision risk is greater than the forward collision risk, then divide the risk level according to the rear time to collision interval, and match and execute hierarchical intervention operations such as hierarchical sound and light warning, driving guidance, and gradient active assisted braking, realizing accurate identification, reasonable decision-making, and hierarchical active protection of the rear-end collision risk under deceleration conditions, fundamentally solving the technical problems of inaccurate identification, lagging warning timing, easy generation of false alarms, inability to take into account the safety of both the front and rear, and lack of matching active intervention in traditional single-sensing solutions, and improving the detection reliability, decision-making rationality, and driving experience of rear collision warning.

[0046] Optionally, in the above method, controlling the vehicle to perform an active intervention operation matching the level of the rear collision risk includes: If the level of the rear collision risk is level one, control the vehicle to maintain the current driving condition and do not trigger an active intervention operation.

[0047] In a possible implementation manner, when the level of the rear collision risk is level one, it means that the rear collision risk is at a low risk, and the corresponding interval is 2.5s < TTC_rear ≤ 4.0s, and the vehicle is in a deceleration working condition, and the result of the front and rear risk weight comparison is that the rear risk weight W_1 is greater than the front risk weight W_2. At this time, do not change the original vehicle driving state, brake pedal opening, and power output state of the vehicle, maintain the current driving and deceleration conditions under the driver's control unchanged, and do not initiate voice prompts, instrument HUD visual warnings, high-frequency double-flash controls, and any active braking intervention-type active intervention operations. Only continuously and real-time monitor the distance, relative speed, time to collision value, and lane attribution status of the rear target in the background, continuously refresh the risk parameter and level determination, and do not output any perceivable control instructions externally, so as to ensure that the driver's autonomous control authority is not interfered by the system.

[0048] If the level of the rear collision risk is level two, control the vehicle to perform a warning intervention operation.

[0049] Among them, the warning intervention operation is to output at least one warning signal to prompt driving risks and turn on the brake light to convey the intention of deceleration to the following vehicle.

[0050] In a possible implementation manner, when the level of the rear collision risk is level two, it means that the rear collision risk is at a medium risk, and the corresponding interval is 1.5 s < TTC_rear ≤ 2.5 s, and the result of the comparison of the front and rear risk weight ratios is that the rear risk weight W_1 is greater than the front risk weight W_2. At this time, the vehicle multi-dimensional warning intervention operation is triggered. On the one hand, multiple warning signals are generated, and the warning signals at least include: in-vehicle voice broadcast to prompt the driver to gently step on the brake to warn the following vehicle, the combination instrument and the HUD interface synchronously display the exclusive icon warning information of the approaching following vehicle, control the vehicle's double flash signal lamp to switch to a 4 Hz high-frequency flashing mode, etc., to form at least one type of audible, visual and optical warning signal to visually prompt the driver of the rear-end collision risk behind; on the other hand, the brake light of this vehicle is continuously turned on through the logic instruction of the vehicle body controller, and by using the constant on of the brake light and the frequent on and off changes of the brake light brought by the driver's pulsed gentle stepping on the brake, the driving intention of this vehicle to continuously decelerate and the need to pay attention to maintaining a safe following distance are visually conveyed to the following vehicle. Only warning prompts and signal lamp state control are performed, and the adjustment of the vehicle's braking pressure is not intervened, and the driver's braking control dominance is retained.

[0051] If the level of the rear collision risk is level three, then control this vehicle to perform warning intervention operation and auxiliary braking.

[0052] Among them, the auxiliary braking is to gradually change the braking pressure of this vehicle until the rear emergency braking system is activated.

[0053] Among them, level one is less than level two is less than level three. Among them, level one corresponds to a low risk, with the lowest risk level and the lowest intervention level; level two corresponds to a medium risk, with a risk level and intervention intensity higher than level one; level three corresponds to a high risk, with the highest risk level and intervention intensity. Among them, the smaller the rear collision time TTC_rear, the higher the risk level, the more comprehensive the intervention measures that need to be intervened by the vehicle control system, and the higher the braking control level.

[0054] In one possible implementation, when the rear collision risk level is three, it means that the rear collision risk is high, corresponding to the interval TTC_rear ≤ 1.5s, and the comparison result of the front and rear risk weights shows that the rear risk weight W_1 is greater than the front risk weight W_2. In this case, all the warning intervention operations corresponding to the level two are fully reused first, and the phased auxiliary braking control is executed simultaneously. The phased auxiliary braking control refers to adjusting the vehicle's braking pressure and deceleration in stages according to the threshold corresponding to the rear collision time TTC_rear. In the first stage, when the rear collision time TTC_rear ≤ 1.5s, the braking pressure is gradually increased, applying a slight deceleration of 0.5~1.0m / s², and illuminating the brake lights to provide a primary warning. In the second stage, when the rear collision time TTC_rear ≤ 0.8s and the front collision time TTC_front > 1.0s, and the forward direction is in a safe state, the braking pressure is further increased to reach a moderate deceleration level of 2.0~3.0m / s², enhancing the deceleration warning effect. In the third stage, when the rear collision time TTC_rear ≤ 0.5s, the braking pressure is further increased in stages to the limit threshold, officially activating the rear automatic emergency braking (RAEB) system, outputting a maximum deceleration of 6~8m / s² for hazard avoidance. The auxiliary braking control uses the threshold corresponding to the rear collision time TTC_rear as the basis for graded switching throughout the entire process. It achieves smooth acceleration and deceleration by continuously and gradually changing the braking pressure to avoid abrupt braking. At the same time, a safety lock-in logic is set up, that is, if the forward collision time TTC_front is detected to be ≤1.0s, all auxiliary braking and rear emergency braking intervention will be terminated immediately to prioritize avoiding the risk of collision with the vehicle in front.

[0055] The vehicle braking control method provided in this application, when the rear collision risk level is divided into three progressively increasing levels: Level 1, Level 2, and Level 3, with Level 1 risk level and intervention intensity lower than Level 2, and Level 2 lower than Level 3, if the rear collision risk is determined to be Level 1, the vehicle is controlled to maintain its current driving condition without triggering any active intervention operations. Only the background continuously monitors the motion state and risk parameters of the rear target, avoiding redundant interventions that interfere with the driver's normal driving operation when there is no actual safety threat. If the rear collision risk is determined to be Level 2, a matching warning intervention operation is immediately executed, that is, at least one warning signal is output to promptly alert the driver to the risk of rear-end collision, while the brake lights are illuminated to visually convey the vehicle's intention to decelerate to the rear vehicles, achieving two-way risk notification between the driver and the vehicle, and avoiding potential rear-end collision accidents in advance. If the rear collision risk is determined to be Level 3, on the basis of executing all the warning intervention operations corresponding to Level 2, auxiliary braking control is simultaneously superimposed. The braking pressure of the vehicle is adjusted in stages according to the threshold corresponding to the rear collision time, and the braking intensity is gradually increased, progressively advancing until the rear emergency braking system is activated for extreme avoidance. Therefore, the entire set of hierarchical matching control logic in this application can adapt the intervention level according to the degree of danger of rear approach, which not only preserves the driver's driving control and improves the ride comfort, but also achieves full-gradient safety protection from early warning prompts to active braking under different risk conditions, so as to resolve the decision contradiction between preventing frontal collisions and preventing rear-end collisions during the deceleration of the vehicle.

[0056] Optionally, the first risk parameter includes: backward collision time TTC_rear.

[0057] The auxiliary braking in the above method includes: If the first risk parameter is less than or equal to the first preset risk parameter threshold, the vehicle is controlled to perform pulse braking at the first preset frequency and the brake lights are illuminated.

[0058] The first preset risk parameter threshold can be selected according to the actual situation; for example, the first preset risk parameter threshold can be selected as 1.5s. The first preset frequency can also be selected according to the actual situation; for example, the first preset frequency can be selected as 0.5~1.0m / s².

[0059] In one possible implementation, the collision time TTC is calculated according to the above formula (2), and then the collision time TTC is dynamically thresholded by combining the vehicle speed, road type, weather, and the type of object behind the vehicle. Finally, the calculated rear collision time TTC_rear is used. When the rear collision time TTC_rear is determined to be ≤1.5s, and the preconditions of the vehicle being in deceleration condition and the rear risk weight W_1 being greater than the front risk weight W_2 are met, the vehicle controller outputs the first control command according to the first preset frequency (such as 0.5~1.0m / s²), controls the braking system to perform intermittent and periodic pulse braking operation; at the same time, the second control command is sent to the vehicle body light controller to continuously illuminate the vehicle's brake lights. The alternating effect of the brake lights brought by the pulse braking is used to intuitively convey the vehicle's deceleration warning diagram to the following vehicle, realizing the rear vehicle reminder without changing the vehicle's driving speed, and retaining the driver's control authority throughout the process.

[0060] If the first risk parameter is less than or equal to the second preset risk parameter threshold, the second risk parameter is greater than the third preset risk parameter threshold, and the deceleration of the vehicle is less than the first preset deceleration threshold, then the vehicle is controlled to perform pulse braking at the second preset frequency.

[0061] Both the second and third preset risk parameter thresholds can be selected according to actual conditions. For example, the second preset risk parameter threshold can be selected as 0.8s, and the third preset risk parameter threshold can be selected as 1.5s. The third preset risk parameter threshold is the safety threshold for forward collision risk. It should be noted that the third preset risk parameter threshold can be equal to or different from the first preset risk parameter threshold, depending on the actual situation.

[0062] The second preset frequency can also be selected according to actual conditions. For example, the second preset frequency can be selected as 2.0~3.0 m / s². The second preset frequency is higher than the first preset frequency to match the progressive control relationship where a smaller rear collision time TTC_rear corresponds to a higher approach risk and a more frequent pulse braking warning. The first preset deceleration threshold is used to indicate that the driver has not performed extreme braking and there is still room for optimizing the safety strategy through pulse warnings. It can be selected according to actual conditions; for example, the first preset deceleration threshold can be selected as 3.5 m / s².

[0063] In one possible implementation, when the rear collision time TTC_rear is determined to be ≤0.8s, and the vehicle is in a deceleration state and the preconditions of rear risk weight W_1 being greater than front risk weight W_2 are met, and simultaneously, the forward collision time TTC_front >1.5s and the vehicle's deceleration is <3.5m / s², the vehicle is controlled to switch to a second preset frequency (e.g., 2.0~3.0m / s²) to perform pulse braking. This is to increase the flashing frequency of the brake lights by increasing the frequency of pulse braking start and stop, thereby strengthening the warning effect on vehicles approaching from behind.

[0064] If the first risk parameter is less than or equal to the fourth preset risk parameter threshold, then the vehicle is controlled to activate the rear automatic emergency braking system.

[0065] The first preset risk parameter threshold is greater than the second preset risk parameter threshold, which in turn is greater than the fourth preset risk parameter threshold. The fourth preset risk parameter threshold can be selected based on actual conditions; for example, it can be set to 0.5s.

[0066] In one possible implementation, when the rear collision time TTC_rear is detected to be ≤0.5s, it is determined that the rear vehicle has entered an extreme rear-end collision scenario of extremely close proximity and high speed, which meets the critical condition for activating the rear automatic emergency braking. At this time, the vehicle controller, while retaining the aforementioned warning and pulse braking intervention, sends a control command to the vehicle's brake actuator to formally activate the rear automatic emergency braking system and performs rear-end collision avoidance according to the preset maximum deceleration (e.g., 6~8m / s²).

[0067] It should be noted that if the forward collision time TTC_front ≤ 1.0s is detected, the activation command of the rear automatic emergency braking system is immediately blocked to avoid the rear-end collision caused by the vehicle decelerating too quickly due to forced rear braking. This achieves a safety control logic of front and rear risk interlocking and priority control.

[0068] The vehicle braking control method provided in this application, when the first risk parameter (such as rear-end collision time) is less than or equal to the first preset risk parameter threshold, controls the vehicle to perform pulse braking at the first preset frequency and simultaneously illuminates the brake lights. The regular illumination and extinguishing of the brake lights visually conveys the vehicle's deceleration status to the following vehicle, thereby achieving early and gentle warning and avoiding the risk of rear-end collisions caused by forced deep braking. When the first risk parameter is less than or equal to the second preset risk parameter threshold, the second risk parameter representing the risk of forward collision is greater than the third preset risk parameter threshold, and the actual deceleration of the vehicle does not reach the first preset deceleration threshold, the control switches to perform pulse braking at a higher second preset frequency. Under the premise of ensuring that the safety of the preceding vehicle is not affected, the warning intensity to the approaching vehicle is strengthened, making up for the lack of warning caused by insufficient braking intensity of the vehicle. When the first risk parameter is less than or equal to the fourth preset risk parameter threshold, the control activates the rear automatic emergency braking system to intervene in braking and avoid severe rear-end collision accidents. The system sets a first preset risk parameter threshold that is greater than a second preset risk parameter threshold, a second preset risk parameter threshold that is greater than a fourth preset risk parameter threshold, and a first preset frequency that is less than a second preset frequency. This hierarchical matching method, with risk thresholds ranging from large to small and pulse braking frequencies ranging from low to high, achieves a gradient-based intervention for rear-end collision risks, from mild to severe. This approach not only progressively increases the intensity of risk response and adapts to different approach conditions, but also balances the safety of vehicles ahead and behind, preserves the driver's basic control authority, avoids the decision-making contradiction between braking to prevent frontal collisions and preventing rear-end collisions when the vehicle is decelerating, and improves the smoothness and adaptability of active safety intervention.

[0069] Optionally, the first risk parameter includes: backward collision time TTC_rear. Figure 3 A flowchart illustrating a vehicle braking control method provided in this application embodiment. Figure 2 .like Figure 3 As shown, the above method also includes: S310. If the first risk parameter is greater than the fifth preset risk parameter threshold, or the deceleration of the vehicle is less than the second preset deceleration threshold and continues for a first preset duration, then the exit process will begin.

[0070] Among them, the fifth preset risk parameter threshold is greater than the fourth preset risk parameter threshold, so that the risk exit judgment threshold is higher than the rear automatic emergency braking activation threshold, forming a hysteresis control range to prevent the system from frequently switching intervention and exiting the state at the critical rear collision time; the first preset deceleration threshold is greater than the second preset deceleration threshold to distinguish between the deep deceleration intervention trigger threshold and the smooth deceleration exit threshold, and together with the preset smooth transition logic, effectively avoids control jitter at the critical point of the working condition, ensuring the system judgment is stable and reliable.

[0071] The fifth preset risk parameter threshold, the second preset deceleration threshold, and the first preset duration can all be selected according to actual conditions. For example, the fifth preset risk parameter threshold can be selected as 3.0s, the second preset deceleration threshold can be selected as 1.0m / s², and the first preset duration can be selected as 0.5s.

[0072] In one possible implementation, if the real-time calculated rear collision time TTC_rear is greater than the fifth preset risk parameter threshold (e.g., 3.0s), it means that the target vehicle behind has moved away from the vehicle, and the risk of rear-end collision is eliminated; or, if the real-time deceleration of the vehicle is detected to be less than the second preset deceleration threshold (e.g., 1.0m / s²), and this low deceleration state is continuously and stably maintained for the first preset duration (e.g., 0.5s), it can be determined that the vehicle has left the deep deceleration condition, and there is no longer a decision contradiction between front and rear collision avoidance when the vehicle decelerates.

[0073] S320. During the exit process, control the vehicle to reduce braking pressure according to the preset exit pattern and close the warning output according to the preset exit sequence.

[0074] In one possible implementation, after the system triggers the exit process, the braking pressure of the vehicle is gradually reduced according to a preset ramp exit pattern, using a linear ramp with a time constant of 0.2 seconds to avoid the sudden release of braking pressure and the resulting vehicle jerking sensation, ensuring smooth driving. For various warning outputs, they are shut down step-by-step according to a preset fixed exit sequence. Specifically, the high-frequency 4Hz hazard lights are first reduced to 2Hz and maintained for 0.5 seconds before resuming the standard 1Hz flashing frequency. Next, after the current warning voice phrase is fully played in the vehicle, the voice prompt output is terminated. Finally, the rear approach warning icons on the instrument panel and HUD fade out gradually within 0.5 seconds, disabling the visual warning. The entire exit process does not instantaneously cut off any intervention signal or braking output. A smooth transition of lights, voice, visual prompts, and braking status is achieved through graded delays and gradual decay control. Furthermore, the entire exit process continues to monitor the front and rear collision times and risk weight parameters in the background. If the rear risk escalates again, intervention can be immediately reinstated.

[0075] The vehicle braking control method provided in this application determines that the rear-end collision risk has been eliminated when the first risk parameter representing the rear-end collision risk is greater than the fifth preset risk parameter threshold, or when the vehicle deceleration is less than the second preset deceleration threshold and remains stable for a first preset duration. It then actively initiates a smooth exit process for the intervention strategy. The fifth preset risk parameter threshold is greater than the fourth preset risk parameter threshold, and the first preset deceleration threshold is greater than the second preset deceleration threshold. By setting differentiated thresholds to form a reasonable hysteresis range, the method effectively avoids the problems of frequent switching of system operating condition critical points and repeated false triggering of intervention logic. During the exit process, the system follows a pre-defined... The exit mechanism is designed to gradually reduce braking pressure in a linear, ramp-like manner to avoid vehicle jerking and driving shock caused by a sudden drop in braking pressure. At the same time, various warning outputs are turned off in a pre-set exit sequence. First, the high-frequency hazard lights are gradually reduced in frequency and then restored to the normal flashing frequency. After the current voice warning broadcast statement ends completely, the voice prompt stops, and the visual icons behind the instrument panel and head-up display system fade out smoothly within a set time. This achieves a stepless and smooth exit of braking status, light warnings, voice and visual prompts, maintaining vehicle smoothness and ride comfort, while preventing sudden changes in warning signals from misleading the driver of this vehicle and following vehicles.

[0076] Optionally, the above method includes controlling the vehicle to reduce braking pressure according to a preset withdrawal pattern and shutting off the warning output according to a preset withdrawal sequence, including: The braking pressure of the vehicle is controlled to decrease by a preset time constant, and the rate of decrease of the braking pressure is less than a preset maximum rate threshold.

[0077] The preset time constant and the preset maximum slope threshold can both be selected according to the actual situation. For example, the preset time constant can be selected as 0.2 seconds.

[0078] In one possible implementation, when the exit criteria are met—namely, the rearward collision time TTC_rear of the rear target increases by more than 3.0 seconds, or the vehicle's deceleration is less than 1.0 m / s² and remains in this state for 0.5 seconds—active braking pressure smoothing control is immediately initiated. Based on a 0.2-second ramp-based linear release, the currently applied auxiliary braking and rear emergency braking pressures are linearly reduced. Simultaneously, built-in braking pressure descent slope constraint logic limits the real-time braking pressure change slope to within a preset maximum slope threshold, preventing a sudden drop in braking pressure. Furthermore, a preset hysteresis and forgetting factor algorithm is used to smooth the pressure decay process, avoiding vehicle pitching and jerking caused by sudden braking pressure changes, achieving a smooth, controllable, and uniform braking pressure decline.

[0079] In the first stage, the flashing mode of the vehicle's brake lights is switched to the standard mode.

[0080] The blinking frequency for both the blinking mode and the standard mode can be selected according to the actual situation.

[0081] In one possible implementation, the first stage of the lighting control logic is initiated simultaneously with the risk exit process. For the 4Hz high-frequency hazard flashing mode activated during risk intervention, the flashing frequency is first reduced from 4Hz to 2Hz and maintained for a 0.5-second transition period before smoothly switching to the vehicle's standard 1Hz flashing mode. This completes the gradual transition of the braking and warning lights from the high-risk warning mode to the standard mode. This first stage is the first action executed in the entire exit process, employing a gradual frequency reduction followed by standard adjustment to avoid abrupt changes in light frequency and prevent visual disturbance to the driver of this vehicle and following vehicles.

[0082] In the second stage, the voice prompts will terminate after the current playback cycle is completed.

[0083] In one possible implementation, after the first stage of brake light mode switching is initiated, the second stage of voice control proceeds according to a preset sequence. Instead of forcibly truncating the voice broadcast, the system monitors the playback progress and cycle of the in-vehicle warning voice in real time. Based on a preset rule that the broadcast stops after the current phrase has completely finished, the system waits for the currently playing risk warning voice to complete its playback cycle before cutting off the voice output channel and terminating subsequent voice prompts. This ensures that each warning message is fully delivered to the driver while avoiding interruptions that could disrupt the human-machine interaction experience. The timing of this execution is later than the initiation of the first stage of light control.

[0084] In the third stage, the icons that appear behind the human-computer interaction interface are switched to a hidden state.

[0085] The execution start time of the first phase is earlier than that of the second phase, and the execution start time of the second phase is earlier than that of the third phase.

[0086] In one possible implementation, after the second-stage voice prompt enters the final termination process, the sequence proceeds sequentially to the third stage of human-machine interface visual control. For the rear approach warning icons that are permanently displayed on the instrument cluster, HUD, and other human-machine interaction interfaces, a 0.5-second gradual fade-out control logic is adopted to gradually reduce the icon display transparency and weaken the visual display intensity until the rear approach icon is switched to a hidden state, rather than disappearing instantly. The smooth removal of visual warning information is achieved through the gradual fading method, ensuring that the transition of the human-machine interface display state is natural and without abruptness.

[0087] The vehicle braking control method provided in this application, when the risk of a rear-end collision is eliminated and the exit conditions are met, controls the vehicle's braking pressure to decrease smoothly according to a preset time constant, while limiting the rate of decrease in braking pressure to less than a preset maximum rate threshold. A ramp-like linear release method is used to avoid sudden changes in braking pressure that could cause vehicle jerking, thus improving ride smoothness. Furthermore, the method executes exit control of multi-dimensional intervention signals sequentially and in stages. In the first stage, it prioritizes switching the high-frequency warning flashing mode of the vehicle's brake lights to a standard flashing mode, achieving a smooth transition of the light warning state. In the second stage, it waits for the voice prompt to complete the current playback. The broadcast is terminated after the cycle is completed to prevent the interactive experience from being interrupted midway. In the third stage, the rear approach warning icon in the human-machine interface such as the control instrument and head-up display is switched to a hidden state to complete the orderly withdrawal of visual warnings. The execution start time of the first stage is set to be earlier than the second stage, and the execution start time of the second stage is set to be earlier than the third stage. This is to achieve the gradual withdrawal of braking pressure, lights, voice, and visual warnings through staggered timing and smooth control mechanism with hysteresis and forgetting factors. This avoids the abruptness caused by the instantaneous start and stop of various intervention signals and ensures the continuity of the vehicle control logic and the natural and smooth human-machine interaction experience.

[0088] Figure 4 A flowchart illustrating a vehicle braking control method provided in this application embodiment. Figure 3 .like Figure 4 As shown, before obtaining the first risk parameter representing the rearward collision risk between the vehicle and a rearward target, the method further includes: S410. Based on the first distance and first relative velocity of multiple rear objects relative to the vehicle, calculate the lateral offset of each rear object relative to the vehicle.

[0089] The lateral offset is the vertical distance by which each object behind the vehicle deviates from the center line of the vehicle's lane.

[0090] In one possible implementation, the vehicle's millimeter-wave radar detects the road environment in real time and simultaneously outputs raw sensing data such as the first distance d, the first relative velocity v_rel, and the target azimuth angle θ for multiple objects behind it. The first distance d is the straight-line relative distance between the object behind it and the vehicle, and the first relative velocity v_rel is the longitudinal velocity of the object behind it relative to the vehicle. The lateral offset Δy of each object behind it is calculated one by one based on the following formula (5).

[0091] Δy= d × tan(θ) formula (5) The physical definition of the lateral offset Δy is the vertical distance by which a rear object deviates from the center line of the vehicle's lane in the lateral dimension of the road.

[0092] Millimeter-wave radar can simultaneously detect the motion and position parameters of multiple candidate objects behind the vehicle in multiple frames. The system independently applies the above formula (5) to solve the lateral offset for each detected object behind the vehicle, providing a quantitative basis for the initial screening of candidate targets in the lane. The entire calculation process is updated in real time with the radar detection frame rate to ensure the continuity and real-time nature of the calculation of the position of objects behind the vehicle.

[0093] S420. The acquired image data is processed to identify and obtain the type of each object behind and lane line information.

[0094] Among them, the type is used to distinguish between vehicles and non-vehicle obstacles, and the lane line information is used to define the boundaries of the vehicle's driving area.

[0095] In one possible implementation, the vehicle's rear-view camera acquires rear road image data in real time at a preset fixed frame rate. Preprocessing operations are then performed on the raw image data, including distortion correction, ROI cropping, adaptive brightness equalization, and Gaussian denoising, to remove invalid interference areas such as the sky and roadside debris. The preprocessed image data is then input into two types of lightweight deep learning models to complete the recognition process: one is a pre- set up The object detection model, this pre set up The target detection model extracts features through convolutional processing of the backbone network, multi-scale fusion of feature pyramids, bounding box regression of the detection head, and calculation of Softmax activation probability. It outputs the target category label and classification confidence score for each object behind it, distinguishing between vehicles such as sedans, SUVs (Sport Utility Vehicles), and trucks, as well as non-vehicle obstacles such as guardrails, road signs, and non-motorized vehicles. The Softmax activation probability of the vehicle category is directly used as the target classification reliability component in the subsequent fusion confidence score calculation. The second model is a preset lane detection model. Through encoder feature downsampling, pixel segmentation and vector clustering of the dual-branch decoder, multinomial fitting of the lane line discrete point set, and pixel-to-vehicle coordinate system transformation, it outputs stable left and right lane line positions and lane width parameters. The standard lane width can be set to 3.5~3.75m. Based on the spatial coordinates of the left and right lane lines, the physical boundaries of the left and right areas of the vehicle's current driving lane can be defined, providing a visual benchmark for subsequent target lane attribution determination.

[0096] S430. Based on the lateral offset, the type of each rear object, and lane line information, determine the rear targets belonging to this lane.

[0097] In one possible implementation, a pre-defined three-level joint judgment mechanism is used to screen and determine rear targets in the current lane based on the lateral offset of the rear object, the type of the rear object output by visual recognition, and lane line boundary information. The first level is the initial lateral offset screening criterion, setting the judgment condition that the lateral offset Δy ≤ standard lane width / 2 + 0.5m. If the condition is met, it is listed as a candidate target in the current lane; otherwise, it is directly judged as a target in the adjacent lane and eliminated. The second level is motion consistency verification, which checks the target's lateral speed within a 1-second time window. If |v_lat| < 0.5m / s and the lateral offset |Δy| shows no significant drift, it retains its candidate status. If the target's lateral speed is greater than 1.0m / s, it is judged to be in a lane-changing state and is temporarily excluded from the list of risk targets in the current lane. The third level is visual lane line fusion, which projects the radar target coordinates onto the camera image coordinate system and checks whether the projected point falls within the driving area boundary defined by the left and right lane lines. If the camera cannot extract the lane lines normally, only the results of the first two levels of judgment are used as the final judgment basis. At the same time, the above formula (1) can be used to verify the validity of the candidate targets. Only high-confidence targets with a comprehensive confidence level C≥0.70 are retained. Finally, a maximum of 3 valid candidate rear targets in this lane are selected for each frame, and the candidate target with the smallest rear collision time is selected as the rear target to complete the determination of the rear targets in this lane and avoid the problem of false detection and false warning caused by irrelevant targets in the adjacent lane.

[0098] The vehicle braking control method provided in this application is based on the acquisition of the first distance, first relative velocity, and target azimuth angle of multiple rear objects relative to the vehicle by millimeter-wave radar. It uses geometric calculations to solve for the lateral offset of each rear object relative to the vehicle. This lateral offset represents the vertical distance of the rear object from the centerline of the vehicle's driving lane, providing a quantitative positional basis for the initial screening of target lane affiliation. Simultaneously, it acquires rear road image data through a rear-view camera, and performs preprocessing and deep learning model recognition processing on the images to calculate the category attributes of each rear object and road lane line information. The category attributes can effectively distinguish between vehicles and non-vehicle obstacles, eliminating invalid interfering targets. The lane line information can accurately define the boundary area of ​​the vehicle's driving lane. Based on this, it integrates the lateral offset of the rear object, target type, and lane line information, along with a preset three-level joint judgment rule and a multi-component comprehensive confidence verification mechanism, to identify and determine valid rear targets within the current lane. This filters out targets in adjacent lanes and non-vehicle interference from the source, effectively reducing the probability of false detections and false alarms in rear risk warnings, and improving the accuracy and reliability of rear target identification within the current lane.

[0099] Optionally, the method described above obtains a first risk parameter representing the rearward collision risk between the vehicle and a rearward target, including: Based on the first distance and the first relative velocity of the target behind, a first risk parameter is calculated to represent the risk of a rearward collision between the vehicle and the target behind.

[0100] In one possible implementation, based on the first distance d and the first relative velocity v_rel of the rear target, the rearward basic collision time TTC is calculated using the above formula (2). At the same time, the TTC is adaptively corrected by combining preset dynamic threshold rules. For example, the baseline threshold of the rearward basic collision time TTC is automatically adjusted in the range of 2.0s to 3.5s by combining the vehicle speed, road type, weather conditions, and the type of the following vehicle. The thresholds of relative velocity and relative distance are matched synchronously. The corresponding parameter relaxation or tightening strategies are implemented for urban roads, highways, and rainy weather environments, respectively, to obtain the rearward collision time TTC_rear. Then, the rearward risk weight W_1 is calculated according to the above formula (3). According to the preset fixed constraint logic: when the collision time TTC_rear of the rearward target ≤ 1.2s, the rear risk weight W_1 is forcibly assigned to 1; when the collision time TTC_rear of the rearward target > 4.0s, the rear risk weight W_1 is forcibly assigned to 0. The dynamically corrected collision time TTC_rear of the rearward target and the solved rear risk weight W_1 are combined to form the first risk parameter, so as to comprehensively quantify the degree of rear-end collision risk between the vehicle and the rearward target from the time dimension and the risk weight dimension.

[0101] The vehicle braking control method provided in this application calculates a first risk parameter representing the rearward collision risk between the vehicle and the rearward target based on a first distance and a first relative velocity of the rearward target. Therefore, this application can objectively calculate the rapid approach state of the rearward target using the actually detected first distance and first relative velocity, avoiding the one-sidedness of judging with a single parameter. Simultaneously, it can provide a reliable and quantifiable calculation basis for subsequent comparison of front and rear risk weights, classification of rearward risk levels, and triggering of graded active intervention strategies, improving the real-time performance and accuracy of rearward collision risk identification.

[0102] Optionally, the above method determines rear targets belonging to the current lane based on lateral offset, the type of each rear object, and lane line information, including: Based on the type of each rear object, determine the vehicle type among multiple rear objects to obtain multiple first rear targets.

[0103] In one possible implementation, the type of each rear object can be used to distinguish between motor vehicles such as cars, SUVs, and trucks, and non-motorized vehicles, pedestrians, road signs, guardrails, and other non-vehicle obstacles. Therefore, the probability corresponding to the type of each rear object can be used as a classification reliability component according to the above formula (1). Type discrimination calculation is performed on all rear objects to eliminate all non-vehicle obstacles and retain only rear objects with the identification type of motor vehicles. All vehicle-type rear objects obtained after screening are defined as multiple first rear targets to complete the first round of target coarse screening.

[0104] Based on the lateral offset, multiple second rear targets are identified from multiple first rear targets.

[0105] In one possible implementation, the lateral offset Δy of each first rear target relative to the centerline of the vehicle's lane is calculated one by one. Based on the set initial screening criteria for the current lane, such as a standard lane width ranging from 3.5m to 3.75m, with |Δy| ≤ lane width / 2 + 0.5m as the judgment benchmark, the lateral offset Δy of each first rear target is verified one by one. If the lateral offset Δy of the first rear target meets the above judgment benchmark, it is determined to be a candidate target adjacent to the current lane and retained; if it exceeds the constraint range, it is directly determined to be a target in the adjacent lane and eliminated. Thus, by filtering through the quantitative threshold of lateral offset, targets that are significantly deviated from the current lane are filtered out from multiple first rear targets, and the remaining candidate targets that meet the lateral position conditions are determined as multiple second rear targets, completing the second level of initial position screening.

[0106] Based on lane line information, the rear target belonging to this lane is determined from multiple second rear targets.

[0107] Lane line information refers to the spatial position information of the left and right lane lines, which is used to define the left and right boundary range of the lane in which the vehicle is traveling.

[0108] In one possible implementation, based on lane line information, motion consistency verification is first performed. This involves setting a 1-second sliding time window. If the target's lateral velocity satisfies |v_lat| < 0.5 m / s and the lateral offset |Δy| shows no significant drift, it is retained as a candidate. If the target's lateral velocity v_lat is greater than 1.0 m / s, it is determined that the target is in a lane-changing state and is directly eliminated. Next, visual lane line fusion is performed, which involves projecting the coordinates of the second rear target detected by radar onto the image coordinate system and verifying whether the projected point falls within the boundary range of the current lane defined by the left and right lane lines. If the camera malfunctions and cannot extract the lane lines, only the lateral offset criterion and motion consistency verification results are used as the judgment basis. At the same time, combined with the comprehensive confidence threshold rule, only high-confidence targets with a comprehensive confidence C ≥ 0.70 are retained. Finally, rear targets located within the current lane are selected from multiple second rear targets, and a maximum of 3 candidate targets in the current lane are retained per frame, providing an effective target source for subsequent rear collision risk calculation and graded intervention.

[0109] The vehicle braking control method provided in this application distinguishes vehicles from non-vehicle obstacles such as guardrails, road signs, and non-motorized vehicles based on the type of each rear object, thereby determining the vehicle type among multiple rear objects and obtaining multiple first rear targets to eliminate invalid non-vehicle interference sources at the source. Then, based on the lateral offset and combined with the lane width constraint threshold, a preliminary screening is performed to initially screen out multiple second rear targets whose positions meet the lateral range requirements of the current lane from the multiple first rear targets, so as to quickly eliminate vehicle targets in the adjacent lanes that are significantly deviated from the current lane. Finally, based on the lane line information, the boundary of the vehicle's driving area is defined, and the rear targets belonging to the current lane are determined from the multiple second rear targets. Thus, this application effectively avoids the false detection problem caused by adjacent lane targets and road debris by using a hierarchical identification method of vehicle type classification filtering, lateral offset coarse screening, and accurate lane line boundary positioning, thereby improving the accuracy and reliability of the determination of the rear target attribution in the current lane.

[0110] Optionally, the above method determines multiple second rear targets from multiple first rear targets based on lateral offset, including: The lateral offset of each first rear target is determined to be less than or equal to the sum of half the preset standard lane width and the preset margin.

[0111] The preset standard lane width and preset margin can be selected according to the actual situation. For example, the preset standard lane width can be 3.5~3.75m; the preset margin can be 0.5m.

[0112] In one possible implementation, the lateral offset Δy of each first rear target relative to the center line of the vehicle's lane is calculated one by one using the above formula (5); then the preset lateral offset is obtained by adding half of the preset standard lane width (e.g., 3.5~3.75m) to the preset margin using the following formula (6).

[0113] Preset lateral offset = lane width / 2 + preset margin formula (6) Then, iterate through all the first rear targets, and compare the calculated absolute value of the lateral offset |Δy| with the preset lateral offset for each one. Then, determine whether each first rear target meets the constraint condition of the preset lateral offset (i.e., the absolute value of the lateral offset |Δy| is less than or equal to the preset lateral offset) to complete the quantitative criterion verification of the initial screening of targets in this lane and provide an objective calculation basis for the division of candidate areas. If so, then each first rear target is determined to be in the candidate area of ​​this lane, and each first rear target in the candidate area of ​​this lane is determined to be a second rear target.

[0114] In one possible implementation, when any first rear target, after being verified by a preset lateral offset, meets the condition that the absolute value of the lateral offset |Δy| is less than or equal to the preset lateral offset, the first rear target is designated as a candidate area within the current lane. For first rear targets that do not meet the condition, they are directly identified as targets in adjacent lanes and are removed, and they are no longer included in the subsequent lane assignment verification process. All first rear targets that fall into the candidate area of ​​the current lane are uniformly relabeled and marked as second rear targets, so as to realize the hierarchical screening transition from the original radar target to the candidate target of the current lane, laying the foundation for subsequent identification of effective rear targets in the current lane and avoiding false detection of targets in adjacent lanes.

[0115] The vehicle braking control method provided in this application verifies, one by one, whether the lateral offset of each first rear target is less than or equal to the sum of half the preset standard lane width and a preset margin, based on the lateral offset of each first rear target. When the lateral offset of the first rear target meets this constraint, it can be determined that the first rear target falls within the candidate area of ​​this lane, and all first rear targets within the candidate area of ​​this lane are uniformly labeled as second rear targets. Thus, this application utilizes the geometric ranging and angle measurement characteristics of radar to complete the initial screening of lateral position quantification, which can quickly eliminate invalid targets obviously located in adjacent lanes, filter out roadside interference and targets unrelated to adjacent lanes in advance, and lay a preliminary screening foundation for determining the attribution of targets behind this lane.

[0116] Optionally, before determining the rear target belonging to the current lane from multiple second rear targets based on lane line information in the above method, the method further includes: Obtain the lateral velocity of each second rear target within a preset time window.

[0117] The preset time window can be selected according to the actual situation. For example, the preset time window can be selected as 1 second.

[0118] The second rear target is the rear vehicle target that has entered the candidate queue after the initial screening by lateral offset.

[0119] In one possible implementation, since the lateral position data detected by millimeter-wave radar in a single frame is easily affected by road bumps, radar detection noise, and instantaneous small vehicle swaying, relying solely on the lateral offset of a single frame cannot distinguish the true lateral driving state of the target behind. It is very easy to misjudge the short-term irregular small lateral displacement of the target vehicle as active lane changing behavior. Therefore, it is necessary to collect continuous multi-frame temporal position data by relying on a fixed preset time window, and obtain the target's continuous lateral movement speed through temporal difference calculation. This objectively reflects the true lateral driving trend of the target behind from a long-term perspective, thereby distinguishing whether the target is in a stable lane-keeping state or an active lane-changing state. Invalid rear interference targets that are changing lanes or about to leave the current lane are eliminated in advance. Invalid targets are avoided from participating in subsequent image coordinate system projection, lane line boundary verification, front and rear risk parameter calculation, and active intervention decision-making processes. This reduces the invalid computing power consumption of the on-board controller, and avoids false warnings and false braking caused by targets changing lanes in adjacent lanes. This improves the accuracy of the subsequent selection results of valid rear targets in the current lane and the reliability of the vehicle's active safety control. Based on this, since the system caches the lateral offset time-series data of all frames within a 1-second preset time window (i.e., the data corresponding to each second rear target), the system performs differential calculations based on the rate of change of the lateral offset Δy over time to calculate the lateral velocity v_lat corresponding to each second rear target. The system independently collects trajectory data within a 1-second time window for each candidate second rear target and completes the lateral velocity calculation, providing continuous and time-series motion parameter basis for subsequent lane keeping status determination. If the lateral speed of each second rear target is less than the first preset speed threshold, then each second rear target is determined to be in lane-keeping mode.

[0120] The first preset speed threshold can be selected according to the actual situation. For example, the first preset speed threshold can be selected as 0.5 m / s.

[0121] In one possible implementation, if the absolute value of the lateral velocity of each second rear target is less than a first preset speed threshold, i.e., |v_lat| < 0.5 m / s, and the absolute value of the lateral offset |Δy| of each second rear target within the same time window shows no significant drift, then the second rear target is determined to have a stable driving trajectory and no lateral lane-changing tendency, thus confirming that it is in a lane-keeping state. If the detected target lateral velocity |v_lat| is greater than 1.0 m / s, then the second rear target is directly determined to be in the process of changing lanes, removed from the candidate target queue of this lane, and not involved in subsequent image projection and lane line boundary determination, so as to achieve objective differentiation of the driving behavior of rear targets through quantified speed thresholds.

[0122] Project each second rear target in lane-keeping mode onto the coordinate system of the image data to obtain the projection point of each second rear target.

[0123] In one possible implementation, for a second rear target already determined to be in lane-keeping mode, a coordinate transformation matrix is ​​constructed based on the target vehicle coordinate system position coordinates output by the millimeter-wave radar and the pre-calibrated internal and external parameters of the rear-view camera. This matrix performs a spatial mapping transformation from the radar's three-dimensional world coordinate system to the rear-view camera's two-dimensional image pixel coordinate system. The physical position coordinates of each second rear target are then converted one by one into the pixel coordinate system of the image data, generating a unique corresponding pixel coordinate projection point. This projection point is used to subsequently determine whether the radar target projection point falls within the range of the left and right lane lines identified by the camera. It is also used to calculate the alignment degree between the radar projection point and the visually detected target box, providing raw coordinate data support for the weighted calculation of the spatial matching degree component in the target's overall confidence score.

[0124] The vehicle braking control method provided in this application acquires the lateral velocity of each second rear target within a preset time window. Using a first preset speed threshold as a judgment criterion, when the lateral velocity of each second rear target is less than the first preset speed threshold, it can be determined that each second rear target is in a stable lane-keeping driving state. This method can effectively filter and eliminate interfering targets with large lateral displacement and in lane-changing driving state, avoiding false warnings caused by lane-changing vehicles being included in the risk assessment of this lane. On this basis, only each second rear target that is already in a lane-keeping state undergoes coordinate transformation, projecting it from the radar vehicle coordinate system to the image data coordinate system collected by the rear-view camera, generating the corresponding projection point of each second rear target, providing coordinate basis for subsequent processing, thereby improving the accuracy of multi-sensor fusion matching and the reliability of lane ownership recognition.

[0125] Optionally, the method described above, based on lane line information, determines rear targets belonging to the current lane from multiple second rear targets, including: If the projection point of each second rear target is located within the area defined by the lane line information, then each second rear target is determined as a candidate target belonging to this lane.

[0126] In one possible implementation, the position coordinates of the second rear target in the vehicle coordinate system are mapped one by one to the image pixel coordinate system of the rearview camera to generate the image projection point corresponding to each second rear target. At the same time, the rearview camera outputs the pixel coordinates and spatial boundaries of the left and right lane lines through image preprocessing, ROI cropping, preset lane line model feature extraction, pixel segmentation, discrete point polynomial fitting and coordinate transformation. The left and right lane lines together form the limited area of ​​the current driving lane of the vehicle. Then, the image projection point of each second rear target is compared with the area limited by the lane line information. If the projection point of each second rear target is located in the area limited by the lane line information, then each second rear target is determined to be qualified as a candidate for the lane. Then, according to the above formula (1), the reliability of each second rear target qualified as a candidate for the lane is checked, that is, whether the target comprehensive confidence C is greater than or equal to 0.40. Only when the target comprehensive confidence C ≥ 0.40 is it recognized as a valid target and officially determined as a candidate target in the lane. If the overall confidence level of the target is C < 0.40, it is directly determined as an invalid target and removed from the candidate set.

[0127] It should be noted that if the vehicle camera malfunctions and cannot extract valid lane line information, the visual judgment at this level will be automatically discarded. Candidate targets will be determined only based on the results of the first two levels of lateral offset and motion consistency. At the same time, the reliability of the candidate targets will be verified by combining the above formula (1). Only valid candidate targets with a comprehensive confidence level that meets the requirements (such as C≥0.40) will be retained, and non-vehicle interference targets in the adjacent lanes will be eliminated to avoid invalid targets from entering the subsequent risk calculation.

[0128] Calculate the first collision time of multiple candidate targets, and select the candidate target with the smallest first collision time as the rear target in this lane.

[0129] In one possible implementation, the initial first collision time of each candidate target in the lane is calculated one by one according to the above formula (2); and the initial first collision time is dynamically thresholded by combining the vehicle speed, road type, weather conditions and the type of the following vehicle to obtain the first collision time of multiple candidate targets; after the first collision time of all candidate targets is calculated, the candidate targets are sorted from smallest to largest according to the value of the first collision time, and according to the preset screening rules, that is, a maximum of 3 candidate targets in the lane are retained per frame. The candidate target with the smallest first collision time value and the highest risk of rear-end collision is selected from them and determined as the final rear target in the lane and used as the main target for subsequent risk weight calculation, front and rear risk comparison, graded warning and active braking intervention, so as to provide a unique and reliable target data source for subsequent rear risk quantification assessment and vehicle active safety control.

[0130] The vehicle braking control method provided in this application projects the radar spatial coordinates of each second rear target onto the image coordinate system of the rear-view camera. If the target projection point falls within the boundary of the driving area of ​​the lane defined by the lane line information, the second rear target is determined to have passed the visual lane line verification and is identified as a candidate target within the lane. This can effectively filter out irrelevant targets in adjacent lanes and avoid false detection and false warning problems caused by targets outside the lane from the source. Then, the first collision time of each candidate target is solved one by one using the collision time formula. Based on the fixed number of candidate targets in the lane that are retained in each frame, the candidate target with the smallest first collision time value is selected as the finally confirmed rear target within the lane, so as to lock the rear vehicle with the highest risk of rear-end collision.

[0131] Based on the same inventive concept, this application also provides a vehicle braking control device. Since the principle of the device in this application is similar to the vehicle braking control method described above in this application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0132] Figure 5 This is a schematic diagram of the structure of a vehicle braking control device provided in an embodiment of this application. Figure 5 As shown, the vehicle braking control device 500 may include: The acquisition module 501 is used to acquire a first risk parameter representing the risk of rearward collision between the vehicle and a target behind it, and a second risk parameter representing the risk of forward collision between the vehicle and a target in front of it, when the vehicle is in a deceleration condition. Comparison module 502 is used to compare the rear collision risk and the forward collision risk based on the first risk parameter and the second risk parameter; The control module 503 is used to control the vehicle to perform active intervention operations that match the level of rear collision risk if the rear collision risk is greater than the front collision risk.

[0133] In one optional implementation, the control module 503 is specifically configured to: if the rear collision risk level is Level 1, control the vehicle to maintain its current driving condition and not trigger active intervention operations; if the rear collision risk level is Level 2, control the vehicle to perform a warning intervention operation, which involves outputting at least one warning signal to alert the driver to the risk and illuminating the brake lights to convey the intention to slow down to the following vehicle; if the rear collision risk level is Level 3, control the vehicle to perform a warning intervention operation and auxiliary braking, which involves changing the vehicle's braking pressure in stages until the rear emergency braking system is activated; wherein, Level 1 is less than Level 2 and less than Level 3.

[0134] In one optional implementation, the auxiliary braking in the control module 503 is specifically used for: if the first risk parameter is less than or equal to the first preset risk parameter threshold, controlling the vehicle to perform pulse braking at a first preset frequency and illuminating the brake lights; if the first risk parameter is less than or equal to the second preset risk parameter threshold, the second risk parameter is greater than the third preset risk parameter threshold, and the deceleration of the vehicle is less than the first preset deceleration threshold, controlling the vehicle to perform pulse braking at a second preset frequency; if the first risk parameter is less than or equal to the fourth preset risk parameter threshold, controlling the vehicle to activate the rear automatic emergency braking system; the first preset risk parameter threshold is greater than the second preset risk parameter threshold, which is greater than the fourth preset risk parameter threshold; and the first preset frequency is less than the second preset frequency.

[0135] In an optional embodiment, the vehicle braking control device 500 is further configured to: enter an exit process if the first risk parameter is greater than the fifth preset risk parameter threshold, or if the deceleration of the vehicle is less than the second preset deceleration threshold and continues for a first preset duration, wherein the fifth preset risk parameter threshold is greater than the fourth preset risk parameter threshold; the first preset deceleration threshold is greater than the second preset deceleration threshold; and during the exit process, control the vehicle to reduce braking pressure according to a preset exit rule and close the warning output according to a preset exit sequence.

[0136] In one optional implementation, the vehicle braking control device 500 is specifically used to: control the braking pressure of the vehicle to decrease by a preset time constant, and the rate of decrease of the braking pressure is less than a preset maximum rate threshold; and in the first stage, control the flashing mode of the vehicle's brake lights to switch to a standard mode; in the second stage, control the voice prompt to terminate after the current playback cycle is completed; and in the third stage, control the rear proximity icon of the human-machine interface to switch to a hidden state; wherein the execution start time of the first stage is earlier than that of the second stage, and the execution start time of the second stage is earlier than that of the third stage.

[0137] In an optional implementation, the acquisition module 501 is further configured to: calculate the lateral offset of each rear object relative to the vehicle based on a first distance and a first relative velocity of multiple rear objects relative to the vehicle, wherein the lateral offset is the vertical distance of each rear object from the center line of the vehicle's driving lane; perform recognition processing on the acquired image data to obtain the type and lane line information of each rear object, wherein the type is used to distinguish between vehicles and non-vehicle obstacles, and the lane line information is used to define the boundary of the vehicle's driving area; and determine the rear targets belonging to the current lane based on the lateral offset, the type of each rear object, and the lane line information.

[0138] In one optional implementation, the acquisition module 501 is specifically used to: calculate a first risk parameter representing the risk of rearward collision between the vehicle and the rear target based on a first distance and a first relative speed of the rear target; In one optional implementation, the acquisition module 501 is specifically used to: determine the vehicle type among the multiple rear objects based on the type of each rear object, so as to obtain multiple first rear targets; determine multiple second rear targets from the multiple first rear targets based on the lateral offset; and determine the rear target belonging to the current lane from the multiple second rear targets based on lane line information.

[0139] In one optional implementation, the acquisition module 501 is specifically used to: determine that the lateral offset of each first rear target is less than or equal to the sum of half of the preset standard lane width and the preset margin; if so, determine that each first rear target is in the candidate area of ​​this lane, and determine each first rear target in the candidate area of ​​this lane as a second rear target.

[0140] In an optional implementation, the acquisition module 501 is further configured to: acquire the lateral velocity of each second rear target within a preset time window; if the lateral velocity of each second rear target is less than a first preset speed threshold, determine that each second rear target is in lane-keeping mode; and project each second rear target in lane-keeping mode onto the coordinate system of the image data to obtain the projection point of each second rear target.

[0141] In an optional implementation, the acquisition module 501 is specifically used to: if the projection point of each second rear target is located within the area defined by the lane line information, then each second rear target is determined as a candidate target belonging to the lane; calculate the first collision time of multiple candidate targets, and select the candidate target with the smallest first collision time as the rear target in the lane.

[0142] It should be noted that for details not disclosed in the vehicle braking control device of this application embodiment, please refer to the details disclosed in the vehicle braking control method of this application embodiment, which will not be repeated here.

[0143] These modules can be configured as one or more integrated circuits implementing the above methods, such as one or more Application Specific Integrated Circuits (ASICs), one or more microprocessors, or one or more Field Programmable Gate Arrays (FPGAs). Alternatively, when a module is implemented using processing element scheduler code, the processing element can be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. Furthermore, these modules can be integrated together as a system-on-a-chip (SOC).

[0144] Optionally, embodiments of this application also provide a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor performs the steps of the vehicle braking control method in the above embodiments. The specific implementation and technical effects are similar and will not be repeated here.

[0145] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. The division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the functional units in the various embodiments of this application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The integrated unit described above can be implemented in hardware or in the form of hardware plus software functional units.

[0146] Optionally, this embodiment also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned related steps to implement a vehicle braking control method provided in the above embodiment.

[0147] In this embodiment, the device, computer-readable storage medium, computer program product, or chip are all used to execute the corresponding methods provided above. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0148] Through the above description of the embodiments, those skilled in the art will understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0149] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. The division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed can be implemented through some interfaces. The indirect coupling or communication connection of devices or units can be electrical, mechanical, or other forms.

[0150] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A vehicle braking control method, characterized in that, include: When the vehicle is decelerating, a first risk parameter representing the risk of a rearward collision between the vehicle and a target behind it, and a second risk parameter representing the risk of a forward collision between the vehicle and a target in front of it are obtained. Based on the first risk parameter and the second risk parameter, the rearward collision risk and the forward collision risk are compared; If the rearward collision risk is greater than the forward collision risk, then the vehicle is controlled to perform an active intervention operation that matches the level of the rearward collision risk.

2. The method according to claim 1, characterized in that, The control of the vehicle to perform active intervention operations commensurate with the level of rear-collision risk includes: If the rear collision risk level is Level 1, then the vehicle is controlled to maintain its current driving condition and the active intervention operation is not triggered. If the rear collision risk level is level two, then the vehicle is controlled to perform a warning intervention operation, which is to output at least one warning signal to indicate the driving risk and to illuminate the brake lights to convey the intention to slow down to the following vehicle. If the rear collision risk level is level three, then the vehicle is controlled to perform the warning intervention operation and auxiliary braking. The auxiliary braking is achieved by changing the braking pressure of the vehicle in stages until the rear emergency braking system is activated. Wherein, the first level is less than the second level, which is less than the third level.

3. The method according to claim 2, characterized in that, The auxiliary braking includes: If the first risk parameter is less than or equal to the first preset risk parameter threshold, then the vehicle is controlled to perform pulse braking at the first preset frequency and the brake lights are turned on. If the first risk parameter is less than or equal to the second preset risk parameter threshold, the second risk parameter is greater than the third preset risk parameter threshold, and the deceleration of the vehicle is less than the first preset deceleration threshold, then the vehicle is controlled to perform pulse braking at the second preset frequency. If the first risk parameter is less than or equal to the fourth preset risk parameter threshold, then the vehicle is controlled to activate the rear automatic emergency braking system; the first preset risk parameter threshold is greater than the second preset risk parameter threshold, which is greater than the fourth preset risk parameter threshold; the first preset frequency is less than the second preset frequency.

4. The method according to claim 1, characterized in that, The method further includes: If the first risk parameter is greater than the fifth preset risk parameter threshold, or if the deceleration of the vehicle is less than the second preset deceleration threshold and continues for a first preset duration, then the exit process is initiated. The fifth preset risk parameter threshold is greater than the fourth preset risk parameter threshold; the first preset deceleration threshold is greater than the second preset deceleration threshold. During the exit process, the vehicle is controlled to reduce braking pressure according to a preset exit pattern and to shut off warning outputs according to a preset exit sequence.

5. The method according to claim 4, characterized in that, The control of the vehicle to reduce braking pressure according to a preset withdrawal pattern and to shut down the warning output according to a preset withdrawal sequence includes: The braking pressure of the vehicle is controlled to decrease by a preset time constant, and the rate of decrease of the braking pressure is less than a preset maximum rate threshold. In the first stage, the flashing mode of the vehicle's brake lights is switched to the standard mode. In the second stage, the voice prompts will terminate after the current playback cycle is completed; In the third stage, the icons that appear behind the human-computer interaction interface are switched to a hidden state; The execution start time of the first stage is earlier than that of the second stage, and the execution start time of the second stage is earlier than that of the third stage.

6. The method according to claim 1, characterized in that, Before obtaining the first risk parameter representing the rearward collision risk between the vehicle and a rearward target, the method further includes: Based on the first distance and first relative velocity of multiple rear objects relative to the vehicle, the lateral offset of each rear object relative to the vehicle is calculated, wherein the lateral offset is the vertical distance of each rear object from the center line of the vehicle's driving lane. The acquired image data is processed to identify the type and lane line information of each object behind it. The type is used to distinguish between vehicles and non-vehicle obstacles, and the lane line information is used to define the boundary of the vehicle's driving area. Based on the lateral offset, the type of each rear object, and the lane line information, rear targets belonging to the current lane are determined.

7. The method according to claim 6, characterized in that, The acquisition of the first risk parameter, used to represent the risk of a rear-end collision between the vehicle and a rearward target, includes: Based on the first distance and the first relative speed of the rear target, a first risk parameter is calculated to represent the risk of a rearward collision between the vehicle and the rear target.

8. The method according to claim 6, characterized in that, The step of determining rear targets belonging to the current lane based on the lateral offset, the type of each rear object, and the lane line information includes: Based on the type of each rear object, determine the vehicle type among the plurality of rear objects to obtain a plurality of first rear targets; Based on the lateral offset, a plurality of second rear targets are determined from the plurality of first rear targets; Based on the lane line information, the rear target belonging to the current lane is determined from the plurality of second rear targets.

9. The method according to claim 8, characterized in that, Before determining the rear target belonging to the current lane from the plurality of second rear targets based on the lane line information, the method further includes: Obtain the lateral velocity of each of the second rear targets within a preset time window; If the lateral speed of each of the second rear targets is less than the first preset speed threshold, then it is determined that each of the second rear targets is in lane keeping mode. Project each of the second rear targets in lane-keeping mode onto the coordinate system of the image data to obtain the projection point of each second rear target.

10. An electronic device, characterized in that, The electronic device includes: Memory, used to store executable program code; A processor for calling and running the executable program code from the memory, causing the electronic device to perform the method as described in any one of claims 1 to 9.

Images