V2V-based AEB control device and method

Abstract

The invention aims to provide a V2V-based AEB control device and method, and the device comprises a V2V-based AEB control device which comprises a V2V data communication module, a real-time data collection module, a data analysis module, a dynamic braking module, and a voice prompt module. The V2V data communication module is used for carrying out information intercommunication with nearby vehicles and transmitting the information to the data analysis module in real time; the real-time data acquisition module is used for acquiring the speed of the vehicle, the distance between the vehicle and the influence factor and the road condition image, and transmitting to the data analysis module in real time; the data analysis module is used for analyzing the currently acquired image and the vehicle sharing information and outputting a decision strategy to the dynamic braking module; and the dynamic braking module is used for controlling a vehicle braking system to reduce the current vehicle speed to a passing speed suitable for passing through the influence factors according to the received decision strategy, and reminding a driver by voice. According to the invention, road and environment influence factors in the vehicle driving environment can be detected and evaluated in real time, and dynamic control of the vehicle is realized through intelligent decision.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle intelligent control, in particular to an AEB control device and method based on V2V. BACKGROUND

[0002] In recent years, with the development of intelligent driving technology and the improvement of vehicle safety, active safety system has become an important research field of vehicle intelligence. As one of the key functions of active safety technology, automatic emergency braking system (AEB) has been widely used in many advanced driver assistance systems (ADAS). AEB system aims to reduce the possibility of accidents or the severity of accidents by perceiving the vehicle driving environment, actively warning of impending traffic accidents and automatically taking braking measures.

[0003] The existing AEB system is usually based on simple forward radar or camera perception of the environment, and the detection accuracy of the traditional AEB system under bad weather (such as rain, snow and fog weather) or complex road conditions is low. The delay between the perception, decision and braking action of the system may affect the response effect in emergency situations. When the environment perception is imperfect, it is easy to miss detecting pedestrians, obstacles or false threats, which affects the system performance. Therefore, it is necessary to develop an AEB control device and algorithm under V2V condition considering road defects, which combines multi-sensor perception and intelligent decision-making technology, and effectively improves the safety, stability and adaptability of the vehicle. SUMMARY

[0004] The present application aims to provide an AEB control device and method based on V2V, which can detect and evaluate the road and environmental factors in the vehicle driving environment in real time, and realize dynamic control of the vehicle through intelligent decision-making to ensure the driving safety of the driver.

[0005] The technical solutions of the present application are as follows:

[0006] The AEB control device based on V2V comprises a V2V data communication module, a real-time data acquisition module, a data analysis module, a dynamic braking module and a voice prompt module.

[0007] The V2V data communication module is used for information exchange with nearby vehicles and real-time transmission to the data analysis module.

[0008] The real-time data acquisition module is used for acquiring the speed of the vehicle, the distance between the vehicle and the influencing factors, and the road condition image, and transmitting them to the data analysis module in real time.

[0009] The data analysis module is used to analyze the current acquired image and vehicle sharing information, judge whether there is a factor affecting the driving behavior of the driver, if there is an influencing factor, the clustering analysis is performed to distinguish the influence category, and the decision strategy is output to the dynamic braking module, otherwise the detection is continued;

[0010] The dynamic braking module is used to control the vehicle braking system to reduce the current vehicle speed to a passing speed suitable for passing through the influencing factor according to the received decision strategy, and to prewarn the driver through the voice prompt module.

[0011] The real-time data acquisition module includes a vehicle-mounted ranging camera, a vehicle speed sensor, and a driver state detection device; the vehicle-mounted camera is used to acquire the environmental image in front of the vehicle; the vehicle speed sensor acquires the current vehicle speed; the driver state detection device is used to prewarn the driver, and judge whether the driver participates in braking based on the detection data of the vehicle speed sensor through the data analysis module;

[0012] The data analysis module adopts a single-chip microcomputer as the hardware, and the dynamic braking module adopts a PID controller as the hardware.

[0013] The data analysis module includes a neural network module, a driving scene classification module, and a vehicle control strategy formulation module;

[0014] The neural network module is used to classify the real-time image data acquired by the real-time data acquisition module through the neural network model in the neural network module, acquire the road defect type, position, size, and road traffic data, and input the risk level evaluation module and the vehicle control strategy formulation module;

[0015] The risk level evaluation module comprehensively evaluates and determines the driving scene risk level of the current driving area according to the received road defect type, position, and road traffic data, in combination with the vehicle speed, acceleration, brake state of the nearby vehicle transmitted by the V2V data communication module, and the driving scene risk level of the vehicle, and transmits the driving scene risk level to the vehicle control strategy formulation module;

[0016] The vehicle control strategy formulation module selects a corresponding strategy mode according to the driving scene risk level, formulates an AEM control strategy of the vehicle based on the road defect type, position, size, and road traffic data, in combination with the vehicle speed, acceleration, and brake state information of the nearby vehicle transmitted by the V2V data communication module, and transmits the AEM control strategy to the dynamic braking module, which controls the vehicle braking system according to the AEM control strategy and prewarns.

[0017] The processing process in the neural network model is as follows:

[0018] The input image is sequentially processed by convolution, ReLU function, maximum pooling and full connection layer to obtain an output result.

[0019] The full connection layer is composed of a plurality of sequentially connected units, each unit is composed of a convolution layer and a ReLU function, and the output of the convolution layer is taken as the input of the ReLU function.

[0020] In the risk level evaluation module, the driving scene risk levels are classified into low-risk scene, medium-risk scene, high-risk scene and dangerous scene; the judgment standards and strategy modes corresponding to each driving scene risk level are as follows:

[0021]

[0022]

[0023] The driving scene risk level judgment standard is as follows:

[0024] From the aspects of road defect type, road defect size, distance between defect and vehicle and nearby vehicle information, a linear weighted mathematical model is used to construct the evaluation model of the four indexes as shown in formula 1:

[0025] f=f1c1+f2c2+f3c3+f4c4 (1)

[0026] Wherein f represents the comprehensive evaluation result of road risk, f1 represents the road defect type, f2 represents the road defect size, f3 represents the distance between the vehicle and the defect, and f4 represents the overall brake rate of the nearby vehicle. c1, c2, c3, c4 represent the weights corresponding to the road defect type, the size of the road defect, the distance between the vehicle and the defect, and the overall brake rate of the nearby vehicle;

[0027] The type f1 of road defect is divided into 0, 1, 2 and 3 according to the severity, and the specific value division is as follows:

[0028] Defect type Road defect risk value (f1) Longitudinal crack (D00) 0 Transverse crack (D10) 0 Oblique crack (D20) 1 Dimple (D40) 2

[0029] The size f2 of road defect is obtained by replacing the actual size of the defect with the pixel area of the pre-made frame in image detection, and the formula is:

[0030] f2=(x min -x max )×(y min -y max ) (2)

[0031] Wherein x max represents the relative position of the left edge of the pre-made frame on the x-axis of the image, x mid represents the relative position of the right edge of the pre-made frame on the x-axis of the image, ymin represents the relative position of the upper edge of the frame on the y-axis of the image, y max represents the relative position of the lower edge of the frame on the y-axis of the image;

[0032] The distance f3 between the vehicle and the road defect is obtained by focusing on the road defect part by the vehicle-mounted ranging camera, so as to obtain the distance between the vehicle and the road defect;

[0033] The overall brake rate f4 of the nearby vehicle is obtained by calculating the proportion of the vehicle that brakes in the total amount of vehicle linkage, and the overall brake rate formula is:

[0034]

[0035] Wherein η represents the vehicle that brakes, and β represents the total vehicle that accesses.

[0036] In the vehicle control strategy module, in the strategy mode, the process of formulating the AEM control strategy of the vehicle includes the following steps:

[0037] A. First, calculate the shortest safe distance of vehicle braking, and the current vehicle speed v0 can be obtained through the vehicle speed sensor, and the shortest safe distance S

[0038] The road defect distance recognized by the neural network module is calculated by the safety distance algorithm to obtain the shortest braking distance S max required for full braking and the shortest braking distance S mid required for partial braking of the vehicle.

[0039] B. The TTC algorithm is used to calculate when the vehicle starts braking.

[0040] The shortest braking distance S max required for full braking is calculated as follows:

[0041] The data processing and reminding time is t1, and the vehicle does not generate braking force at the driver reaction time t2 and the brake coordination time t3, and the distance traveled in this process is:

[0042] S1=(t1+t2+t3)v0 (4)

[0043] Let the full braking acceleration of the vehicle during the continuous braking stage be a bmax , and the partial braking acceleration be a bmid The brake braking force increases linearly within t3 time;

[0044]

[0045]

[0046] Integrating both sides of formula (5), (6) gives:

[0047]

[0048] The vehicle speed at t3" is:

[0049]

[0050] The distance traveled by the vehicle during t3is:

[0051]

[0052] During the sustained braking phase, when full braking is applied, the vehicle moves a distance S3:

[0053]

[0054]

[0055] S max = S1+S2+S3 (19) ;

[0056] Substituting formula (4), formula (12), and formula (18) into formula (19) gives

[0057]

[0058] The shortest braking distance S required for partial braking of the vehicle is: mid The calculation process is as follows

[0059] Based on the above formula (4)-(14) ;

[0060] The vehicle moves a distance S3' :

[0061]

[0062] S mid = S1+S2'+S3' (22)

[0063] Substituting formula (4), formula (14), and formula (21) into formula (22) gives:

[0064]

[0065] The process of calculating when the vehicle starts braking using the TTC algorithm is:

[0066] The TTC formula is used to determine when to start braking, and is as follows:

[0067]

[0068] In the formula, vx2 , v y2 are the longitudinal and lateral velocities of the dynamic factor, respectively, obtained by the sensor, wherein m is the mass of the vehicle; v x , v y are the longitudinal and lateral velocities of the vehicle, respectively; are the longitudinal and lateral accelerations, respectively; ω z is the yaw rate; δ represents the front wheel steering angle; F yij represents the lateral force on the wheel; F xij is the longitudinal force on the wheel, wherein 11 represents the left front wheel, 12 represents the right front wheel, 21 represents the left rear wheel, 22 represents the right rear wheel, and Δt is the time for the vehicle to pass through the defect area at the current speed.

[0069] The derivation process of the TTC formula is as follows:

[0070] Based on the basic formula (24)

[0071]

[0072] wherein Δs is the relative distance between the ego vehicle and the target object, v r is the relative speed between the ego vehicle and the target object, which, in the present application, refers to the driving speed of the ego vehicle;

[0073] In order to cope with more complex working conditions, the relative acceleration a r is introduced to cope with more complex working conditions, and the calculation model thereof is:

[0074]

[0075] which can be further expressed as:

[0076]

[0077] a r is the relative acceleration between the ego vehicle and the target object;

[0078] However, after introducing a r , there are still situations in which the TTC fails, i.e., v r ≥ 0, a ≥ 0 and v r 2 - 2Δsa r < 0, the above formula has no solution. At this time, the TTC risk assessment model is likely to misjudge, causing false alarms and missed alarms, etc.; in addition, the model ignores the reaction time of different drivers and the characteristics of the vehicle, such as brake response delay, etc.

[0079] Based on the traditional TTC algorithm, the AEB algorithm is optimized as follows in combination with the vehicle performance analysis of the traditional AEB algorithm and the functional requirements of the AEB optimization algorithm:

[0080] ①Optimize the relative distance in TTC algorithm, introduce the field of view angle of sensor to calculate more accurate relative distance between vehicle and dynamic factor;

[0081] ②Optimize the relative speed in TTC algorithm, use the variables such as longitudinal relative speed and transverse relative speed in dynamic model to optimize the relative speed;

[0082] ③Increase the transverse TTC2 algorithm as a dangerous criterion.

[0083] θ is the azimuth angle of the dynamic factor relative to the straight driving direction of the vehicle, which is used to calculate the longitudinal relative displacement and transverse relative displacement of the dynamic factor and the vehicle, and X is the relative distance between the vehicle and the dynamic factor. Considering the vehicle characteristics such as vehicle speed, braking performance and braking reaction time of the vehicle, and the influence of unstructured road surface on the vehicle, the safety redundant distance is considered, and the longitudinal relative safety distance X of the dynamic factor relative to the front road defect area of the vehicle is obtained. r

[0084] X r =X·cosθ-d x (28)

[0085] The transverse relative safety distance of the road defect factor relative to the front collision area of the vehicle is:

[0086]

[0087] In the formula: d x is the longitudinal safety redundant distance, d y is the transverse safety redundant distance, and w is the vehicle width.

[0088] Substitute formula (28) and (29) into formula (25) to obtain the TTC formula of formula (24).

[0089] The application also discloses an AEB control method based on V2V, which comprises the following steps:

[0090] A, the real-time data acquisition module acquires the vehicle speed, the distance between the vehicle and the influencing factor and the road condition image in real time, and transmits them to the data analysis module in real time. At the same time, the V2V data communication module is used for information exchange with nearby vehicles, and transmits them to the data analysis module in real time.

[0091] B, the data analysis module analyzes the current acquired image and vehicle shared information, judges whether there is a factor affecting the driving behavior of the driver, if there is an influencing factor, carries out clustering analysis to distinguish the influence category, outputs the decision strategy to the dynamic braking module, otherwise continues to detect.

[0092] ​C, the dynamic braking module is used for controlling the vehicle braking system to reduce the current speed to a passing speed suitable for passing through the influencing factors according to the received decision strategy, and warning the driver through the voice prompt module.

[0093] Preferably, the V2V-based AEB control method comprises the following steps:

[0094] A, the real-time data acquisition module comprises a vehicle-mounted ranging camera, a vehicle speed sensor and a driver state detection device; the vehicle-mounted camera is used for acquiring environmental images in front of the vehicle; the vehicle speed sensor acquires the current vehicle speed and transmits it to the data analysis module in real time; the driver state detection device is used for warning and reminding the driver and judging whether the driver participates in braking; at the same time, the V2V data communication module is used for information exchange with nearby vehicles and transmits it to the data analysis module in real time;

[0095] B, the data analysis module comprises a neural network module, a driving scene classification module and a vehicle control strategy formulation module;

[0096] The neural network module is used for classifying the real-time image data acquired by the real-time data acquisition module through the neural network model in the neural network module, acquiring the road defect type, position, size and road traffic data, and inputting them into the risk level evaluation module and the vehicle control strategy formulation module;

[0097] The risk level evaluation module comprehensively evaluates and determines the driving scene risk level of the current driving area according to the received road defect type, position and road traffic data, in combination with the speed, acceleration, brake state of the nearby vehicles transmitted by the V2V data communication module and the driving scene risk level of the vehicle, and transmits the driving scene risk level to the vehicle control strategy formulation module;

[0098] The vehicle control strategy formulation module selects the corresponding strategy mode according to the driving scene risk level, formulates the AEM control strategy of the vehicle based on the road defect type, position, size and road traffic data, in combination with the speed, acceleration and brake state information of the nearby vehicles transmitted by the V2V data communication module, and transmits it to the dynamic braking module, which controls the vehicle braking system according to the AEM control strategy and performs warning and reminding.

[0099] The processing process in the neural network model is as follows:

[0100] The input image is sequentially processed by convolution, ReLU function, maximum pooling and full connection layer to obtain the output result;

[0101] The full connection layer is composed of a plurality of units connected in sequence, each unit is composed of a convolution layer and a ReLU function, and the output of the convolution layer is used as the input of the ReLU function.

[0102] In the risk level evaluation module, the driving scene risk levels are classified into low-risk scene, medium-risk scene, high-risk scene and dangerous scene; the judgment standards and strategy modes corresponding to each driving scene risk level are as follows:

[0103]

[0104] The driving scene risk level judgment standard is as follows:

[0105] From the road defect type, road defect size, defect distance from the vehicle and nearby vehicle information, a linear weighted mathematical model is used to construct the evaluation model of the four indicators as shown in formula 1:

[0106] f=f1c1+f2c2+f3c3+f4c4 (1)

[0107] Wherein f represents the comprehensive evaluation result of road risk, f1 represents the road defect type, f2 represents the road defect size, f3 represents the distance between the vehicle and the defect, and f4 represents the overall brake rate of the nearby vehicle. c1, c2, c3, c4 represent the weights corresponding to the road defect type, the size of the road defect, the distance between the vehicle and the defect, and the overall brake rate of the nearby vehicle;

[0108] The type f1 of road defect is divided into 0, 1, 2 and 3 according to the severity, and the specific value division is as follows:

[0109] Defect type Road defect risk value (f1) Longitudinal crack (D00) 0 Transverse crack (D10) 0 Oblique crack (D20) 1 Dimple (D40) 2

[0110] The road defect size f2 is obtained by replacing the actual size of the defect with the pixel area occupied by the pre-made frame in image detection, and the formula is:

[0111] f2=(x max -x min )×(y max -y min ) (2)

[0112] Wherein x min represents the relative position of the left edge of the pre-made frame on the image x axis, x max represents the relative position of the right edge of the pre-made frame on the image x axis, y min represents the relative position of the upper edge of the pre-made frame on the image y axis, and y max represents the relative position of the lower edge of the pre-made frame on the image y axis;

[0113] The distance f3 between the vehicle and the road defect is obtained by focusing on the road defect part through the vehicle-mounted ranging camera, so as to obtain the distance between the vehicle and the road defect.

[0114] The overall brake rate f4 of the nearby vehicles is obtained by calculating the proportion of the vehicles that brake in the total amount of vehicle links, and the overall brake rate formula is:

[0115]

[0116] Wherein η represents the vehicle that brakes, and β represents the total vehicle that accesses.

[0117] In the vehicle control strategy module, in the strategy mode, the process of formulating the AEM control strategy of the vehicle includes the following steps:

[0118] A. First, calculate the shortest safe distance of vehicle braking, and the current vehicle speed v0 can be obtained through the vehicle speed sensor. Based on

[0119] The road defect distance recognized by the neural network module is calculated by the safety distance algorithm to obtain the shortest braking distance S max required for full braking and the shortest braking distance S mid required for partial braking of the vehicle.

[0120] B. The TTC algorithm is used to calculate when the vehicle starts braking.

[0121] The shortest braking distance S max required for full braking is calculated as follows:

[0122] The data processing and reminding time is t1, and the vehicle does not generate braking force at the driver reaction time t2 and the brake coordination time t3. The distance traveled in this process is:

[0123] S1=(t1+t2+t3)v0 (4)

[0124] Let the full braking acceleration of the vehicle during the continuous braking stage be a bmax , and the partial braking acceleration be a bmid The brake braking force increases linearly within t3 time;

[0125]

[0126] Integrating both sides of formulas (5) and (6) gives:

[0127]

[0128] The vehicle speed of the vehicle at t3” is:

[0129]

[0130] The distance traveled by the vehicle during t3 time is:

[0131]

[0132] In the sustained braking phase, when full braking is applied, the vehicle travel distance S3 is:

[0133]

[0134] S max = S1+S2+S3 (19) ;

[0135] Substituting formula (4), formula (12), and formula (18) into formula (19) gives

[0136] The shortest braking distance S required for partial braking of the vehicle is: mid The calculation process is as follows based on the above formulas (4)-(14) ;

[0137] The vehicle travel distance S3' is:

[0138]

[0139] S mid = S1+S2'+S3' (22)

[0140] Substituting formula (4), formula (14), and formula (21) into formula (22) gives:

[0141]

[0142] The process of calculating when the vehicle starts braking using the TTC algorithm is:

[0143] The TTC formula is as follows:

[0144]

[0145] In the formula, v x2 , v y2 are the longitudinal and lateral speeds of the dynamic factor, respectively, obtained from the sensor In the formula, m is the mass of the vehicle; v x , v y are the longitudinal and lateral speeds of the vehicle, respectively; are the longitudinal and lateral accelerations, respectively; ω z is the yaw rate; δ represents the front wheel steering angle; F yij represents the lateral force on the wheel; F xij is the longitudinal force on the wheel, where 11 represents the left front wheel, 12 represents the right front wheel, 21 represents the left rear wheel, and 22 represents the right rear wheel, and Δt is the time for the vehicle to pass through the defect area at the current speed;

[0146] The data analysis module analyzes the current acquired image and vehicle sharing information, judges whether there is a factor affecting the driving behavior of the driver based on TTC, if there is an influencing factor, carries out clustering analysis to distinguish the influence category, outputs a decision strategy to the dynamic braking module, otherwise continues to detect;

[0147] C, the dynamic braking module is used for controlling the vehicle braking system to reduce the current vehicle speed to a passing speed suitable for passing through the influencing factor according to the received decision strategy, and warning the driver through the voice prompt module.

[0148] The model of the V2V data communication module can be selected as RM500; the RM500 communication device is mounted on the GW105-5G communication module, the frequency range is 2.412-2.4835GHz / 5.180-5.825GHz, the transmission power is 14-22dBm, the receiving sensitivity is-75dBm~ -88dBm, and the channel number is 36-173.

[0149] The single-chip microcomputer model of the data analysis module can be selected as STM32G4 single-chip microcomputer.

[0150] The beneficial effects of the present application are as follows:

[0151] The present application establishes a risk assessment model, which considers four aspects of road defect types, road defect sizes, distances between defects and vehicles, and nearby vehicle information obtained by V2V communication equipment. Among them, when the vehicle matches the nearby vehicle, data sharing is realized through the V2V communication equipment to obtain the surrounding vehicle information, so as to make up for the situation that the front vehicle of the vehicle is too much, the vision of the vehicle-mounted ranging camera is blocked, and the front road surface passability cannot be detected in time. At the same time, the speed, acceleration, brake state of the nearby vehicle and the scene type of the nearby vehicle are shared by V2V, which are used as one of the conditions of the current vehicle driving scene risk level.

[0152] The road environment detection brake system of the present application is suitable for various intelligent vehicles, automatic driving vehicles, and advanced driving assistance systems (ADAS). Whether in urban roads, highways or rural roads, the system can provide accurate environment perception and braking control, improve the safety of the vehicle and the driving experience of the driver. In addition, the design of the system can adapt to various severe weather and complex terrain, and provide more reliable support for the automatic driving and auxiliary driving of the vehicle.

[0153] The present application constructs an AEB control device and algorithm considering road defects combined with V2V, which can effectively improve the safety, stability and adaptability of the vehicle, and has good application prospect. DETAILED DESCRIPTION

[0154] The application will be described in detail below with reference to the embodiments.

[0155] Embodiment 1

[0156] The V2V-based AEB control device comprises a V2V data communication module, a real-time data acquisition module, a data analysis module, a dynamic braking module and a voice prompt module.

[0157] The V2V data communication module is used for information exchange with nearby vehicles and real-time transmission to the data analysis module.

[0158] The real-time data acquisition module is used for acquiring the vehicle speed, the distance between the influencing factors and the road condition image, and real-time transmission to the data analysis module.

[0159] The real-time data acquisition module comprises a vehicle-mounted ranging camera, a vehicle speed sensor and a driver state detection device.

[0160] The data analysis module is used for analyzing the current acquired image and vehicle shared information, judging whether there is an influencing factor affecting the driving behavior of the driver, and if there is an influencing factor, clustering analysis is performed to distinguish the influence types, and the decision strategy is output to the dynamic braking module, otherwise the detection continues.

[0161] The data analysis module comprises a neural network module, a driving scene classification module and a vehicle control strategy formulation module.

[0162] The neural network module is used for classifying the real-time image data acquired by the real-time data acquisition module through the neural network model in the neural network module, acquiring the road defect type, position, size and road traffic data, and inputting the risk level evaluation module and the vehicle control strategy formulation module.

[0163] The risk level evaluation module receives the road defect type, position and road traffic data, combines the vehicle speed, acceleration, brake state of the nearby vehicle transmitted by the V2V data communication module and the driving scene risk level formulated by the vehicle, comprehensively evaluates and determines the driving scene risk level of the current driving area, and transmits the driving scene risk level to the vehicle control strategy formulation module.

[0164] The vehicle control strategy module selects a corresponding strategy mode according to the driving scene risk level, formulates an AEM control strategy of the vehicle based on road defect types, positions, sizes, and road traffic data, in combination with speed, acceleration, and brake state information of nearby vehicles transmitted by the V2V data communication module, and transmits the AEM control strategy to the dynamic braking module, which controls the vehicle braking system and performs early warning according to the AEM control strategy.

[0165] The processing process in the neural network model is as follows:

[0166] The input image is sequentially processed by convolution, a ReLU function, maximum pooling, and a full connection layer to obtain an output result.

[0167] The full connection layer is composed of a plurality of sequentially connected units, each unit is composed of a convolution layer and a ReLU function, and the output of the convolution layer is used as the input of the ReLU function.

[0168] In the risk level evaluation module, the driving scene risk level is divided into a low-risk scene, a medium-risk scene, a high-risk scene, and a dangerous scene, and the corresponding judgment standards and strategy modes of each driving scene risk level are as follows:

[0169]

[0170]

[0171] The driving scene risk level judgment standard is as follows:

[0172] From the road defect type, the road defect size, the distance between the defect and the vehicle, and the nearby vehicle information, a linear weighted mathematical model is used to construct an evaluation model of the four indexes as shown in formula 1:

[0173] f=f1c1+f2c2+f3c3+f4c4 (1)

[0174] Wherein f represents the comprehensive evaluation result of the road risk, f1 represents the road defect type, f2 represents the road defect size, f3 represents the distance between the vehicle and the defect, and f4 represents the overall brake rate of the nearby vehicle. c1, c2, c3, and c4 represent the weights corresponding to the road defect type, the road defect size, the distance between the vehicle and the defect, and the overall brake rate of the nearby vehicle.

[0175] The road defect type f1 is divided into 0, 1, 2, and 3 according to the severity, and the specific value division is as follows:

[0176] Defect type Road defect risk value (f1) Longitudinal crack (D00) 0 Transverse crack (D10) 0 Oblique crack (D20) 1 Dimple (D40) 2

[0177] The road defect size f2 is obtained by replacing the actual size of the defect with the pixel area occupied by the image detection pre-made frame, and the formula is:

[0178] f2 = (x max -x min ) × (y max -y min ) (2)

[0179] wherein x min represents the relative position of the left edge of the pre-made frame on the image x-axis, x max represents the relative position of the right edge of the pre-made frame on the image x-axis, y min represents the relative position of the upper edge of the pre-made frame on the image y-axis, and y max represents the relative position of the lower edge of the pre-made frame on the image y-axis.

[0180] The distance f3 of the vehicle from the road defect is obtained by focusing on the road defect part by the vehicle-mounted ranging camera, thereby obtaining the distance between the vehicle and the road defect;

[0181] The overall brake rate f4 of the nearby vehicle is obtained by calculating the proportion of the vehicles that brake in the total number of vehicle links, and the formula is:

[0182]

[0183] wherein η represents the vehicles that brake, and β represents the total vehicles.

[0184] In the vehicle control strategy formulation module, in the strategy mode, the process of formulating the AEM control strategy of the vehicle includes the following steps:

[0185] A. First, calculate the shortest safe distance of vehicle braking, and the current speed v0 can be obtained by the vehicle speed sensor, based on

[0186] The road defect distance recognized by the neural network module is calculated by the safety distance algorithm to obtain the shortest braking distance S max required for full braking and the shortest braking distance S mid required for partial braking of the vehicle.

[0187] B. The TTC algorithm is used to calculate when the vehicle starts braking.

[0188] The shortest braking distance S max required for full braking is calculated as follows:

[0189] The data processing and reminding time is t1, and the vehicle does not generate braking force at the driver reaction time t2 and the brake coordination time t3, and the distance traveled in this process is:

[0190] S1 = (t1 + t2 + t3)v0 (4)

[0191] Let the full braking acceleration of the car during the continuous braking phase be a bmax , and the partial braking acceleration be a bmid During t3, the brake braking force increases linearly;

[0192]

[0193] Integrating both sides of formula (5) and (6) gives:

[0194]

[0195] The vehicle speed at t3" is:

[0196]

[0197] The distance traveled by the car during t3 is:

[0198]

[0199] During the continuous braking phase, the vehicle moving distance S3 is:

[0200] S max = S1 + S2 + S3 (19); Substituting formula (4), formula (12), and formula (18) into formula (19) gives

[0201]

[0202] The shortest braking distance S mid The calculation process is as follows

[0203] Based on the above formula (4)-(14);

[0204] The vehicle moving distance S3' is:

[0205]

[0206] S mid = S1 + S2' + S3' (22)

[0207] Substituting formula (4), formula (14), and formula (21) into formula (22) gives:

[0208]

[0209] The TTC algorithm is used to calculate when the vehicle starts braking, and the process is as follows:

[0210] The TTC formula is used to determine the start braking, and the TTC formula is as follows:

[0211]

[0212] In the formula, v x2 , v y2 are the longitudinal and lateral velocities of the dynamic factors, respectively, obtained by the sensor, where m is the mass of the vehicle; v x , v y are the longitudinal and lateral velocities of the vehicle, respectively; are the longitudinal and lateral accelerations, respectively; ω z is the yaw rate; δ represents the front wheel angle; F yij represents the lateral force on the wheel; F xij is the longitudinal force on the wheel, where 11 represents the left front wheel, 12 represents the right front wheel, 21 represents the left rear wheel, and 22 represents the right rear wheel, and Δt is the time for the vehicle to pass through the defect area at the current speed.

[0213] The dynamic braking module is used to control the vehicle braking system to reduce the current vehicle speed to a suitable passing speed through the influence factors according to the received decision strategy, and to prewarn the driver through the voice prompt module.

[0214] Embodiment 2

[0215] The V2V-based AEB control method includes the following steps:

[0216] A, the real-time data acquisition module includes a vehicle-mounted ranging camera, a vehicle speed sensor, and a driver state detection device; the vehicle-mounted camera is used to acquire the environmental image in front of the vehicle; the vehicle speed sensor acquires the current vehicle speed; and real-time data is transmitted to the data analysis module; the driver state detection device is used to prewarn the driver and determine whether the driver participates in braking; at the same time, the V2V data communication module is used to communicate information with nearby vehicles and real-time data is transmitted to the data analysis module;

[0217] B, the data analysis module includes a neural network module, a driving scene classification module, and a vehicle control strategy development module;

[0218] The neural network module is used to classify the real-time image data acquired by the real-time data acquisition module through the neural network model in the neural network module, to acquire the road defect type, position, size, and road traffic data, and to input the risk level evaluation module and the vehicle control strategy development module;

[0219] The risk level evaluation module evaluates and determines the driving scene risk level of the current driving area according to the received road defect type, position, and road traffic flow data, the speed, acceleration, brake state of the nearby vehicle transmitted by the V2V data communication module, and the driving scene risk level of the vehicle, and transmits the driving scene risk level to the vehicle control strategy formulation module.

[0220] The vehicle control strategy formulation module selects the corresponding strategy mode according to the driving scene risk level, formulates the AEM control strategy of the vehicle based on the road defect type, position, size, and road traffic flow data, and the speed, acceleration, and brake state information of the nearby vehicle transmitted by the V2V data communication module, and transmits the AEM control strategy to the dynamic braking module.

[0221] The processing process in the neural network model is as follows:

[0222] The input image is sequentially processed by convolution, ReLU function, maximum pooling, and full connection layer to obtain an output result.

[0223] The full connection layer is composed of a plurality of sequentially connected units, each unit is composed of a convolution layer and a ReLU function, and the output of the convolution layer is used as the input of the ReLU function.

[0224] In the risk level evaluation module, the driving scene risk level is divided into low-risk scene, medium-risk scene, high-risk scene, and dangerous scene, and the corresponding judgment standards and strategy modes of each driving scene risk level are as follows:

[0225]

[0226]

[0227] The driving scene risk level judgment standard is as follows:

[0228] From the road defect type, road defect size, distance between the defect and the vehicle, and nearby vehicle information, a linear weighted mathematical model is used to construct an evaluation model of the four indexes as shown in formula 1:

[0229] f = f1c1 + f2c2 + f3c3 + f4c4 (1)

[0230] Wherein f represents the comprehensive evaluation result of road risk, f1 represents the road defect type, f2 represents the road defect size, f3 represents the distance between the vehicle and the defect, and f4 represents the overall brake rate of nearby vehicles. c1, c2, c3, and c4 represent the weights corresponding to the road defect type, the size of the road defect, the distance between the vehicle and the defect, and the overall brake rate of nearby vehicles;

[0231] The value of the road defect type f1 is divided into 0, 1, 2, and 3 according to the severity, and the specific value division is as follows:

[0232] Defect type Road defect risk value (f1) Longitudinal crack (D00) 0 Transverse crack (D10) 0 Oblique crack (D20) 1 Dimple (D40) 2

[0233] The road defect size f2 is obtained by replacing the actual size of the defect with the pixel area occupied by the pre-made frame in image detection, and the formula is:

[0234] f2=(x max -x min )×(y max -y min ) (2)

[0235] Wherein x min represents the relative position of the left edge of the pre-made frame on the image x-axis, x max represents the relative position of the right edge of the pre-made frame on the image x-axis, y min represents the relative position of the upper edge of the pre-made frame on the image y-axis, and y max represents the relative position of the lower edge of the pre-made frame on the image y-axis.

[0236] The distance f3 between the vehicle and the road defect is obtained by focusing on the road defect part through the vehicle-mounted ranging camera, thereby obtaining the distance between the vehicle and the road defect;

[0237] The overall brake rate f4 of nearby vehicles is obtained by calculating the proportion of vehicles that brake in the total number of vehicle links, and the overall brake rate formula is:

[0238]

[0239] Wherein η represents the vehicle that brakes, and β represents the total number of vehicles connected.

[0240] In the vehicle control strategy formulation module, in the strategy mode, the process of formulating the AEM control strategy of the vehicle includes the following steps:

[0241] A. First, calculate the shortest safe distance for vehicle braking. The current speed v0 can be obtained through the speed sensor, and based on

[0242] The road defect distance recognized by the neural network module is calculated by the safety distance algorithm to obtain the shortest braking distance S required for full brakingmax and the shortest braking distance S required for full braking of the vehicle mid ;

[0243] B, TTC algorithm is used to calculate when the vehicle starts braking.

[0244] The shortest braking distance S required for full braking of the vehicle max The calculation process is as follows:

[0245] The data processing and reminding time is t1, the vehicle does not generate braking force at the driver reaction time t2 and the brake coordination time t3, and the distance traveled during this process is:

[0246] S1 = (t1 + t2 + t3)v0 (4)

[0247] Let the full braking acceleration of the vehicle during the sustained braking phase be a bmax , and the partial braking acceleration be a bmid During t3, the brake braking force increases linearly;

[0248]

[0249] Integrating both sides of equations (5) and (6) gives:

[0250]

[0251] The vehicle speed at t3" is:

[0252]

[0253] The distance traveled by the vehicle during t3 is:

[0254]

[0255] During the sustained braking phase, the vehicle moves a distance S3 when full braking:

[0256] S max = S1 + S2 + S3 (19);

[0257] Substituting equations (4), (12), and (18) into equation (19) gives

[0258]

[0259] The shortest braking distance S required for partial braking of the vehicle mid The calculation process is as follows

[0260] Based on the above equations (4)-(14);

[0261] The vehicle moving distance S3' is:

[0262]

[0263] S mid = S1 + S2' + S3' (22)

[0264] Substituting formula (4), formula (14) and formula (21) into formula (22) obtains:

[0265]

[0266] The process of calculating when the vehicle starts braking by using the TTC algorithm is:

[0267] The TTC formula is used to determine the start braking, and the TTC formula is:

[0268]

[0269] In the formula, v x2 , v y2 are the longitudinal and lateral velocities of the dynamic factor, which are obtained by the sensor, wherein m is the mass of the vehicle; v x , v y are the longitudinal and lateral velocities of the vehicle; are the longitudinal and lateral accelerations, respectively; ω z is the yaw rate; δ represents the front wheel steering angle; F yij represents the lateral force on the wheel; F xij is the longitudinal force on the wheel, wherein 11 represents the left front wheel, 12 represents the right front wheel, 21 represents the left rear wheel, 22 represents the right rear wheel, and Δt is the time for the vehicle to pass through the defect area at the current speed;

[0270] The data analysis module analyzes the current acquired image and the vehicle shared information, determines whether there is a factor affecting the driving behavior of the driver based on the TTC, and if there is an influencing factor, performs clustering analysis to distinguish the influence category, and outputs a decision strategy to the dynamic braking module, otherwise continues to detect;

[0271] C, the dynamic braking module is used to control the vehicle braking system to reduce the current vehicle speed to a passing speed suitable for passing through the influencing factor according to the received decision strategy, and prewarning is reminded to the driver through the voice prompt module.

Claims

1. A V2V-based AEB control device, comprising a V2V data communication module, a real-time data acquisition module, a data analysis module, a dynamic braking module, and a voice prompt module, characterized in that: The V2V data communication module is used to exchange information with nearby vehicles and transmit it to the data analysis module in real time; The real-time data acquisition module is used to obtain the vehicle's speed, the distance between the vehicle and the influencing factors, and the road condition image, and transmit it to the data analysis module in real time; The data analysis module is used to analyze the currently acquired image and vehicle shared information to determine whether there are factors that affect the driver's driving behavior. If there are influencing factors, cluster analysis is performed to distinguish the types of influencing factors and output the decision strategy to the dynamic braking module. Otherwise, detection continues; The dynamic braking module is used to control the vehicle braking system to reduce the current vehicle speed to a passing speed suitable for passing such influencing factors according to the received decision strategy, and to warn the driver through the voice prompt module.

2. The V2V-based AEB control device according to claim 1, wherein: The real-time data acquisition module includes an on-board ranging camera, a speed sensor, and a driver status detection device; the on-board camera is used to capture images of the environment in front of the vehicle; the speed sensor obtains the current vehicle speed; the driver status detection device is used to provide early warnings to the driver, and based on the speed sensor detection data, the data analysis module determines whether the driver is engaged in braking; The hardware used in the data analysis module is a single chip microcomputer, and the hardware used in the dynamic braking module is a PID controller.

3. The V2V-based AEB control device according to claim 1, wherein: The data analysis module includes a neural network module, a driving scenario classification module, and a vehicle control strategy formulation module; The neural network module is used to classify the real-time image data obtained by the real-time data acquisition module through the neural network model within it, obtain the type, location, size of road defects, and road traffic flow data, and input them into the risk level assessment module and the vehicle control strategy formulation module; The risk level assessment module uses the received road defect type and location, as well as road traffic volume data, combined with the speed, acceleration, braking status of nearby vehicles transmitted by the V2V data communication module, and the vehicle's proposed driving scenario risk level to comprehensively assess and determine the driving scenario risk level of the current driving area. The driving scenario risk level is then transmitted to the vehicle control strategy formulation module. The vehicle control strategy formulation module selects the corresponding strategy mode according to the risk level of the driving scenario. Based on the type, location, size of road defects, and road traffic data, combined with the speed, acceleration, and braking status information of nearby vehicles transmitted by the V2V data communication module, it formulates the vehicle's AEM control strategy and transmits it to the dynamic braking module. The dynamic braking module controls the vehicle's braking system according to the AEM control strategy and issues early warning reminders.

4. The V2V-based AEB control device according to claim 3, wherein: The processing process in the neural network model is as follows: The input image is processed in sequence by convolution, ReLU function, maximum pooling, and fully connected layer to obtain the output result; Among them, the fully connected layer consists of multiple sequentially connected units, each unit consists of a convolutional layer and a ReLU function, and the output of the convolutional layer serves as the input of the ReLU function.

5. The V2V-based AEB control device according to claim 3, wherein: In the risk level assessment module, the risk levels of exercise scenarios are divided into low-risk scenarios, medium-risk scenarios, high-risk scenarios, and dangerous scenarios. The judgment criteria and strategy models corresponding to each exercise scenario risk level are as follows: Preferably, the risk level judgment criteria for the exercise scenario are as follows: Considering the road defect type, road defect size, distance between the defect and the vehicle, and nearby vehicle information, a linear weighted mathematical model is used to construct an evaluation model of four major indicators as shown in Formula 1: f=f1c1+f2c2+f3c3+f4c4 (1) Where f represents the comprehensive assessment of road risk, f1 represents the road defect type, f2 represents the road defect size, f3 represents the distance between the vehicle and the defect, and f4 represents the overall braking rate of nearby vehicles. c1, c2, c3, and c4 represent the weights corresponding to the road defect type, road defect size, distance between the vehicle and the defect, and overall braking rate of nearby vehicles. The type of road defect f1 is divided into 0, 1, 2, and 3 according to the severity. The specific values ​​are as follows: To obtain the road defect size f2, the pixel area occupied by the prefabricated frame of the image detection is used to replace the actual size of the defect. The formula is: f2=(x max -x min )×(and max -and min ) (2) where x min Indicates the relative position of the left edge of the prefabricated frame on the x-axis of the image, x max Indicates the relative position of the right edge of the prefabricated frame on the x-axis of the image, y min Indicates the relative position of the upper edge of the frame on the y-axis of the image, y max Indicates the relative position of the lower edge of the frame on the y-axis of the image; The distance f3 between the vehicle and the road defect is obtained by focusing the road defect part with the on-board ranging camera; The overall braking rate f4 of nearby vehicles is calculated by calculating the proportion of vehicles that brake in the total number of vehicles in the vehicle chain, so the overall braking rate formula is: Where η represents the vehicle that brakes, and β represents the total number of vehicles connected.

6. The V2V-based AEB control device according to claim 3, wherein: In the vehicle control strategy formulation module, in strategy mode, the process of formulating the vehicle's AEM control strategy includes the following steps: A. First calculate the shortest safe distance for vehicle braking. The current vehicle speed v0 can be obtained through the vehicle speed sensor. Based on The neural network module identifies the road defect distance and uses the safety distance algorithm to calculate the shortest braking distance S required for full braking. max The shortest braking distance S required for partial braking of the vehicle mid ; B. Use the TTC algorithm to calculate when the vehicle starts braking.

7. The V2V-based AEB control device according to claim 6, wherein: The shortest braking distance S required for full braking max The calculation process is as follows: The data processing and reminder time is t1. During the driver's reaction time t2 and the brake coordination time t3, the vehicle does not generate braking force. The distance traveled during this process is: S1=(t1+t2+t3)v0 (4) Assume that the full braking acceleration of the car during the continuous braking phase is a bmax , part of the braking acceleration is a bmid During the time t3, the braking force increases linearly; Integrating both sides of formula (5) and (6) we can get: The speed of the car at t3″ is: The distance traveled by the car during t3 is: In the continuous braking stage, when braking with full force, the vehicle moves a distance S3 of: S max =S1+S2+S3 (19); Substituting formula (4), formula (12), and formula (18) into formula (19), we can obtain The shortest braking distance S required for partial braking of the vehicle mid The calculation process is as follows Based on the above formulas (4)-(14); The vehicle moving distance S3' is: S mid =S1+S2'+S3' (22) Substituting formula (4), formula (14), and formula (21) into formula (22), we obtain:

8. The V2V-based AEB control device according to claim 6, wherein: The process of using the TTC algorithm to calculate when the vehicle starts braking is as follows: The start-up braking judgment is based on the following TTC formula: Where, v x2 、v y2 They are the longitudinal velocity and lateral velocity of the dynamic factors, obtained by the sensor, where m is the vehicle mass; v x 、v y are the longitudinal speed and lateral speed of the vehicle respectively; are longitudinal acceleration and lateral acceleration respectively; ω z is the yaw rate; δ represents the front wheel turning angle; F yij Indicates the lateral force on the wheel; F xij is the longitudinal force on the wheel, where 11 represents the left front wheel, 12 represents the right front wheel, 21 represents the left rear wheel, 22 represents the right rear wheel, and Δt is the time it takes for the vehicle to pass through the defect area at the current speed.

9. The V2V-based AEB control method according to claim 1, wherein: The following steps are involved: A. The real-time data acquisition module acquires the vehicle's speed, distance to influencing factors, and road condition images in real time, and transmits them to the data analysis module in real time. Simultaneously, the V2V data communication module is used to exchange information with nearby vehicles and transmits it to the data analysis module in real time. B. The data analysis module analyzes the currently acquired images and vehicle shared information to determine whether there are factors that affect the driver's driving behavior. If there are influencing factors, cluster analysis is performed to distinguish the types of influencing factors and output the decision strategy to the dynamic braking module. Otherwise, the detection continues; C. The dynamic braking module is used to control the vehicle's braking system to reduce the current vehicle speed to a speed suitable for passing through such influencing factors based on the received decision strategy, and to warn the driver through the voice prompt module.

10. The V2V-based AEB control method according to claim 9, wherein: The following steps are involved: A. Real-time data acquisition module includes vehicle-mounted ranging camera, vehicle speed sensor, and driver status detection equipment; The onboard camera is used to collect images of the environment in front of the vehicle; the speed sensor obtains the current vehicle speed and transmits it to the data analysis module in real time; The driver status detection device provides early warnings to the driver and determines whether the driver is engaging in braking. Meanwhile, the V2V data communication module is used to exchange information with nearby vehicles and transmit it to the data analysis module in real time. B. Data analysis module includes neural network module, driving scenario classification module, and vehicle control strategy formulation module; The neural network module is used to classify the real-time image data obtained by the real-time data acquisition module through the neural network model within it, obtain the type, location, size of road defects, and road traffic flow data, and input them into the risk level assessment module and the vehicle control strategy formulation module; The risk level assessment module uses the received road defect type and location, as well as road traffic volume data, combined with the speed, acceleration, braking status of nearby vehicles transmitted by the V2V data communication module, and the vehicle's proposed driving scenario risk level to comprehensively assess and determine the driving scenario risk level of the current driving area. The driving scenario risk level is then transmitted to the vehicle control strategy formulation module. The vehicle control strategy formulation module selects the corresponding strategy mode according to the risk level of the driving scenario. Based on the type, location, size of road defects, and road traffic data, combined with the speed, acceleration, and braking status information of nearby vehicles transmitted by the V2V data communication module, it formulates the vehicle's AEM control strategy and transmits it to the dynamic braking module. The dynamic braking module controls the vehicle's braking system according to the AEM control strategy and issues early warning reminders. The processing process in the neural network model is as follows: The input image is processed in sequence by convolution, ReLU function, maximum pooling, and fully connected layer to obtain the output result; Among them, the fully connected layer consists of multiple sequentially connected units, each unit consists of a convolutional layer and a ReLU function, and the output of the convolutional layer serves as the input of the ReLU function. In the risk level assessment module, the risk levels of exercise scenarios are divided into low-risk scenarios, medium-risk scenarios, high-risk scenarios, and dangerous scenarios. The judgment criteria and strategy models corresponding to each exercise scenario risk level are as follows: The risk level judgment criteria for the exercise scenario are as follows: Considering the road defect type, road defect size, distance between the defect and the vehicle, and nearby vehicle information, a linear weighted mathematical model is used to construct an evaluation model of four major indicators as shown in Formula 1: f=f1c1+f2c2+f3c3+f4c4 (1) Where f represents the comprehensive assessment of road risk, f1 represents the road defect type, f2 represents the road defect size, f3 represents the distance between the vehicle and the defect, and f4 represents the overall braking rate of nearby vehicles. c1, c2, c3, and c4 represent the weights corresponding to the road defect type, road defect size, distance between the vehicle and the defect, and overall braking rate of nearby vehicles. The type of road defect f1 is divided into 0, 1, 2, and 3 according to the severity. The specific values ​​are as follows: To obtain the road defect size f2, the pixel area occupied by the prefabricated frame of the image detection is used to replace the actual size of the defect. The formula is: f2=(x max -x min )×(and max -and min ) (2) where x min Indicates the relative position of the left edge of the prefabricated frame on the x-axis of the image, x max Indicates the relative position of the right edge of the prefabricated frame on the x-axis of the image, y min Indicates the relative position of the upper edge of the frame on the y-axis of the image, y max Indicates the relative position of the lower edge of the frame on the y-axis of the image; The distance f3 between the vehicle and the road defect is obtained by focusing the road defect part with the on-board ranging camera; The overall braking rate f4 of nearby vehicles is calculated by calculating the proportion of vehicles that brake in the total number of vehicles in the vehicle chain, so the overall braking rate formula is: Where η represents the vehicle that brakes, and β represents the total number of vehicles connected. In the vehicle control strategy formulation module, in strategy mode, the process of formulating the vehicle's AEM control strategy includes the following steps: A. First calculate the shortest safe distance for vehicle braking. The current vehicle speed v0 can be obtained through the vehicle speed sensor. Based on The neural network module identifies the road defect distance and uses the safety distance algorithm to calculate the shortest braking distance S required for full braking. max The shortest braking distance S required for partial braking of the vehicle mid ; B. Use the TTC algorithm to calculate when the vehicle starts braking. The shortest braking distance S required for full braking max The calculation process is as follows: The data processing and reminder time is t1. During the driver's reaction time t2 and the brake coordination time t3, the vehicle does not generate braking force. The distance traveled during this process is: S1=(t1+t2+t3)v0 (4) Assume that the full braking acceleration of the car during the continuous braking phase is a bmax , part of the braking acceleration is a bmid During the time t3, the braking force increases linearly; Integrating both sides of formula (5) and (6) we can get: The speed of the car at t3″ is: The distance traveled by the car during t3 is: In the continuous braking stage, when braking with full force, the vehicle moves a distance S3 of: S max =S1+S2+S3 (19); Substituting formula (4), formula (12), and formula (18) into formula (19), we can obtain The shortest braking distance S required for partial braking of the vehicle mid The calculation process is as follows Based on the above formulas (4)-(14); The vehicle moving distance S3' is: S mid =S1+S2'+S3' (22) Substituting formula (4), formula (14), and formula (21) into formula (22), we obtain: The process of using the TTC algorithm to calculate when the vehicle starts braking is as follows: The start-up braking judgment is based on the following TTC formula: Where, v x2 、v y2 They are the longitudinal velocity and lateral velocity of the dynamic factors, obtained by the sensor, where m is the vehicle mass; v x 、v y are the longitudinal speed and lateral speed of the vehicle respectively; are longitudinal acceleration and lateral acceleration respectively; ω z is the yaw rate; δ represents the front wheel turning angle; F yij Indicates the lateral force on the wheel; F xij is the longitudinal force on the wheel, where 11 represents the left front wheel, 12 represents the right front wheel, 21 represents the left rear wheel, 22 represents the right rear wheel, and Δt is the time it takes for the vehicle to pass through the defect area at the current speed; The data analysis module analyzes the currently acquired images and vehicle shared information, and determines whether there are factors that affect the driver's driving behavior based on TTC. If there are influencing factors, cluster analysis is performed to distinguish the types of influencing factors, and the decision strategy is output to the dynamic braking module. Otherwise, the detection continues; C. The dynamic braking module is used to control the vehicle's braking system to reduce the current vehicle speed to a speed suitable for passing through such influencing factors based on the received decision strategy, and to warn the driver through the voice prompt module.

Images