[go: up one dir, main page]

CN118538053A - Method for determining critical warning distance and critical braking distance, and vehicle cooperative safety interaction system integrating 4D millimeter wave radar and monocular vision camera - Google Patents

Method for determining critical warning distance and critical braking distance, and vehicle cooperative safety interaction system integrating 4D millimeter wave radar and monocular vision camera Download PDF

Info

Publication number
CN118538053A
CN118538053A CN202410520987.3A CN202410520987A CN118538053A CN 118538053 A CN118538053 A CN 118538053A CN 202410520987 A CN202410520987 A CN 202410520987A CN 118538053 A CN118538053 A CN 118538053A
Authority
CN
China
Prior art keywords
vehicle
distance
critical
braking
warning
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202410520987.3A
Other languages
Chinese (zh)
Inventor
胡晓伟
姜琳
安实
王子禾
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Harbin Institute of Technology Shenzhen
Original Assignee
Harbin Institute of Technology Shenzhen
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Harbin Institute of Technology Shenzhen filed Critical Harbin Institute of Technology Shenzhen
Priority to CN202410520987.3A priority Critical patent/CN118538053A/en
Publication of CN118538053A publication Critical patent/CN118538053A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/166—Anti-collision systems for active traffic, e.g. moving vehicles, pedestrians, bikes
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/017—Detecting movement of traffic to be counted or controlled identifying vehicles
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/04—Detecting movement of traffic to be counted or controlled using optical or ultrasonic detectors
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/048—Detecting movement of traffic to be counted or controlled with provision for compensation of environmental or other condition, e.g. snow, vehicle stopped at detector
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/052—Detecting movement of traffic to be counted or controlled with provision for determining speed or overspeed

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Traffic Control Systems (AREA)

Abstract

预警临界距离、制动临界距离确定方法及融合4D毫米波雷达和单目视觉摄像头的车辆协同安全交互系统，属于车路协同安全交互技术领域。解决现有的预警临界距离、制动临界距离确定方法存在与实际误差较大的问题，以及目前缺少一种车辆协同安全交互系统的问题。本发明基于伯克利模型基于前车行驶状态分为三种情况，针对前车匀速、减速、静止情况分别计算得到预警临界距离与制动临界距离；本发明的系统首先对前车进行识别并获取前车与本车的行驶速度信息，监测实时本车和前车车距，判断是否超过激活临界车速进而确定预警信息计算单元处于静息或活跃状态，活跃时计算预警临界距离与制动临界距离并进行相应的预警，若发生事故则对事故相关信息进行共享。

The method for determining the critical warning distance and the critical braking distance, and the vehicle cooperative safety interaction system integrating 4D millimeter wave radar and monocular vision camera belong to the field of vehicle-road cooperative safety interaction technology. The problem that the existing methods for determining the critical warning distance and the critical braking distance have large errors with the actual situation, as well as the problem of the current lack of a vehicle cooperative safety interaction system, is solved. The present invention is based on the Berkeley model and is divided into three situations based on the driving state of the front vehicle. The critical warning distance and the critical braking distance are calculated for the uniform speed, deceleration, and stationary conditions of the front vehicle, respectively. The system of the present invention first identifies the front vehicle and obtains the driving speed information of the front vehicle and the vehicle, monitors the real-time distance between the vehicle and the front vehicle, and determines whether it exceeds the activation critical speed to determine whether the warning information calculation unit is in a static or active state. When active, the critical warning distance and the critical braking distance are calculated and corresponding warnings are issued. If an accident occurs, the accident-related information is shared.

Description

Early warning critical distance and braking critical distance determining method and vehicle cooperative safety interaction system fused with 4D millimeter wave radar and monocular vision camera
Technical Field
The invention belongs to the technical field of cooperative safety interaction of vehicles and roads, and particularly relates to a method for determining an early warning critical distance and a braking critical distance and a cooperative safety interaction system of vehicles.
Background
And analyzing and displaying traffic accidents according to the related data and the public data: rear-end collisions may account for 30% to 40% of the total amount of traffic accidents. Authoritative publications indicate that people are the most uncertain factors in traffic safety, and specific data indicate that the human factors of drivers are the largest in the statistical traffic accident cases, and the influence rate is about 81%. It is noted that, the accidents caused by the subjective judgment error in the driving process of the driver account for 79% of the total number of accident cases because the driver cannot timely and accurately perform the perception recognition and prevention on the potential danger in the driving process. In summary, the main reason why the automobile is involved in the rear-end collision is that the driver cannot respond accurately in time, so that timely and effective rear-end collision early warning is a key to reduce the rear-end collision accident.
Up to now, there are still many problems with the ranging sensors commonly used in intelligent network-connected car sensing systems, and different sensors have advantages, disadvantages and applicable conditions. If the conventional millimeter wave radar has higher difficulty in identifying targets, the accuracy is lower in high-speed and long-distance ranging; the ultrasonic Lei Dayi is affected by weather and temperature, and the maximum measurement distance is only a few meters, so that the remote monitoring of the vehicle can not be realized; the laser radar sensor has poor penetrability and high cost, can not work all the time and can not work in the weather such as dense fog, rain and snow; the infrared sensor is sensitive to temperature, is easily affected by weather, and can only measure low-speed objects; the optical imaging sensor cannot ascertain the relevant distance, angle, speed, etc. Aiming at the problems, the monocular camera is a current mainstream sensor, the cost is low, the technology is mature, and the application of the monocular camera is wider along with the rapid improvement of the deep learning and image matching technology.
Most of existing vehicle rear-end collision early warning systems are based on vehicle monitoring and vehicle distance alarming of the vehicle, such as a self-adaptive cruise system, but the system can only realize the monitoring of the vehicle, can not realize the information interaction of multiple vehicles, and can not timely share the vehicle information so as to reduce the serial rear-end collision accident. Most of the current vehicle rear-end collision systems concentrate data processing before collision occurs, but ignore processing of linkage effect after accident occurs, such as vehicle congestion and creep, vehicle rescue and accident responsibility judgment, and the like, and have poor cooperative effect with road side facilities. Timely handling of vehicles after accidents is particularly important in daily driving situations, so that regional cooperative early warning of roads and information sharing with related traffic departments are needed to be realized. The surrounding environment can be subjected to full-view angle pre-sensing through effective fusion of various different sensors, so that the accident rate is effectively reduced.
Disclosure of Invention
The invention aims to solve the problems that the existing method for determining the pre-warning critical distance and the braking critical distance has larger actual error and the existing method for determining the braking critical distance lacks a cooperative safety interaction system of a vehicle.
A method for determining an early warning critical distance and a braking critical distance is characterized in that speed information of a front vehicle and a vehicle is input into an improved Berkeley model, and the early warning critical distance and the braking critical distance are calculated based on the Berkeley model;
In the process of calculating the early warning critical distance and the braking critical distance based on the Berkeley model, the running state of the front vehicle is divided into three conditions, and the early warning critical distance and the braking critical distance are calculated according to different conditions:
(1) When the front vehicle runs at a constant speed:
Early warning critical distance D w1:
braking critical distance D b1:
Wherein: d L is the distance travelled by the vehicle during the take over time of the driver; v 1 is the speed of the vehicle; v 2 is the front vehicle speed; a 1 is the brake deceleration of the vehicle; t 1 is the braking system reaction time; t 2 is the braking system coordination time; d 0 is the safety distance after parking;
(2) When the front vehicle runs at a reduced speed:
early warning critical distance D w2:
Braking critical distance D b2:
(3) When the front vehicle is stationary:
early warning critical distance D w3:
braking critical distance D b3:
The front vehicle driving state corresponds to one of three conditions, D w1、Dw2 or D w3 is used as an early warning critical distance D w, and D b1、Db2 or D b3 is used as a braking critical distance D b;
The braking deceleration a 1 of the vehicle is comprehensively determined by collecting the road surface condition of the vehicle and the position information data of the brake pedal:
Firstly, determining a road adhesion coefficient mu based on the road type, and further determining the maximum braking deceleration according to a max =mu g, wherein g is gravity acceleration; simultaneously determining the travel depth of a brake pedal of the vehicle and braking deceleration, converting the change of the travel of the brake pedal into a proportion, and performing relation curve smoothing to obtain a relation curve graph of the travel proportion of the brake pedal and the braking deceleration of the vehicle;
Then, the brake pedal position is divided into a plurality of sections, one brake deceleration a 1 is corresponding to each section corresponding to the brake pedal position, and the brake deceleration a 1 corresponding to a section is expressed as a 1＝ki*amax,ki to represent the proportionality coefficient corresponding to the ith section according to the relation between the brake deceleration a 1 and a max corresponding to each section.
Further, the driving distance of the vehicle in the taking over time of the driver is d L＝v1t0; wherein t 0 is the driver take over time and v 1 is the speed of the vehicle; when the host vehicle is a non-autonomous vehicle, t 0 =0, i.e., d L =0.
Preferably, the brake pedal position is divided into six interval segments: the brake pedal stroke is defined as 3.5% of the first stage of braking, and then 10%, 25%, 45%, 60%, 75% of the brake pedal stroke is defined as the second to sixth stages of braking.
Preferably, the braking decelerations a 1 corresponding to the six segments of the brake pedal position division are 0, 0.20a max、0.4amax、0.6amax、0.8amax and a max, respectively.
A vehicle collaborative security interaction system that fuses a 4D millimeter wave radar and a monocular vision camera, comprising:
Identifying and searching unit: character recognition is carried out on the front license plate information, and the license plate number of the front car is obtained according to character recognition and extraction; the license plate number of the front car is utilized to carry out information matching in a system database, and if the license plate number of the front car is searched in the system database, the system is also used by the front car; if the license plate number of the front car is not searched, the front car does not use the system;
Vehicle speed information acquisition unit: acquiring the running speed information of a front vehicle and a host vehicle;
If the front vehicle information is retrieved, acquiring the vehicle state information of the front vehicle provided by the Beidou satellite navigation system in an information sharing mode, and acquiring the vehicle state information of the own vehicle on the basis of the Beidou satellite navigation system; the vehicle state information comprises position coordinates and vehicle speed information; then extracting speed information from the vehicle state information;
If the information of the front vehicle is not retrieved or the vehicle state information of the front vehicle provided by the Beidou satellite navigation system cannot be obtained in an information sharing mode, obtaining the speed information of the front vehicle by adopting a 4D millimeter wave radar sensor of the vehicle;
Real-time vehicle distance determining unit: when the distance between two vehicles reaches the pre-warning critical distance or the vehicle does not reach the activation critical speed, and when the monocular vision camera fails, a 4D millimeter wave radar sensor is selected to monitor the real-time vehicle distance, otherwise, the monocular vision camera is utilized to monitor the real-time vehicle distance;
An early warning information activating unit: based on the speed of the vehicle monitored in real time, judging whether the speed exceeds an activation critical speed, and further determining that the early warning information calculating unit is in a resting or active state;
Early warning information calculation unit: calculating to obtain an early warning critical distance D w and a braking critical distance D b by adopting the early warning critical distance and braking critical distance determining method;
Early warning judging and prompting unit: comparing the distance D between the front vehicle and the vehicle with the pre-warning critical distance D w;
if the distance between the two vehicles is larger than the pre-warning critical distance, continuously monitoring;
If the distance between the two vehicles is not greater than the pre-warning critical distance, acquiring a license plate number of the front vehicle and a result of whether the system is used or not by the identification and retrieval unit, acquiring registration information of the front vehicle and the self vehicle based on a system database, and determining whether the front vehicle and the self vehicle are automatic driving vehicles or not; carrying out early warning prompt and judging whether a driver takes effective measures or not;
Braking decision and emergency early warning unit: determining a vehicle distance change state in real time based on the distance D between the front vehicle and the own vehicle, and comparing the distance D between the two vehicles with a braking critical distance D b;
If the distance between two vehicles is continuously increased, the system monitors that the distance between two vehicles is larger than the pre-warning critical distance, and then the dangerous state is relieved;
if the vehicle distance between the two vehicles is continuously reduced, the system enters an emergency early warning state; if the distance between two vehicles is continuously reduced to the braking critical distance, aiming at the vehicle being an automatic driving vehicle, the system sends an instruction to the vehicle to automatically take emergency braking; aiming at the fact that the vehicle is a non-automatic driving vehicle, the system sends out vehicle collision early warning, and meanwhile continuously monitors until collision occurs; if the front vehicle also uses the system, the front vehicle is also subjected to vehicle collision early warning;
An accident information sharing unit: after collision, based on the vehicle position information during collision acquired by the Beidou satellite navigation system, the database retrieves the automatic driving grade of the vehicle, the vehicle running speed and collision early-warning condition data and uploads the data to the database, and then based on the vehicle position information during collision, the early-warning range of the accident site is calculated and determined, and the accident reminding is sent to the vehicle using the system in the early-warning range.
Further, the early warning range D A is as follows:
For vehicles that reach an activation threshold vehicle speed:
DA＝DW+De
Wherein D e is the early warning distance, and D W is the early warning critical distance;
for vehicles which do not reach the activation critical speed, the 4D millimeter wave radar is utilized for collision monitoring, D W is omitted, and D A＝De is enabled.
Further, when the monocular vision camera is used for monitoring the real-time vehicle distance, firstly, the monocular vision camera is used for collecting license plate position information of the front vehicle in real time, the license plate position information is used for representing the position information of the front vehicle, and then the real-time vehicle distance between the vehicle and the front vehicle is calculated based on the imaging relation between the license plate position information of the front vehicle and the monocular vision camera.
Further, the process of calculating the real-time vehicle distance between the vehicle and the front vehicle based on the imaging relation between the license plate position information of the front vehicle and the monocular vision camera comprises the following steps:
Determining the projection relation between a license plate plane and an image plane of a front vehicle based on the imaging characteristics of a monocular vision camera; according to the imaging principle of the camera, if the height of the camera of the vehicle running on the horizontal road surface is unchanged, the bottom edge line of the front vehicle has a corresponding relationship with the distance between the two vehicles, as shown in fig. 3;
Substituting the perceived image plane characteristics of the front vehicle into the following monocular vision camera distance calculation formula to obtain the distance D between the front vehicle and the front vehicle;
a space rectangular coordinate system is established by taking a camera point C as a center, a Z axis is parallel to a main optical axis of the camera, X π is an imaging point of an image plane, the corresponding relation of the license plate plane and the image plane in the figure 4 exists, and the related initial installation angle, focal length, pixel height and observation angle of the vehicle-mounted camera are substituted into a formula, so that the distance D between two vehicles can be obtained;
Wherein D is the distance between the front vehicle and the vehicle; h is the height of the camera from the road surface; θ c is the included angle between the camera and the vertical axis, namely the initial installation angle of the camera; h is the pixel height; h i is the height of the effective imaging surface of the camera; d p is the effective size of the imaging plane pixel; alpha is the observable angle of the camera.
Further, in the process that the early warning judging and prompting unit judges whether the driver takes effective measures, whether the driver takes the effective measures is determined based on a set steering threshold value and a braking prefabrication threshold value, and when the transverse offset of the vehicle exceeds the steering threshold value or the stroke of a braking pedal of the vehicle exceeds the braking prefabrication threshold value, the driver is judged to receive early warning information and takes the effective measures; if the vehicle does not take effective measures, the vehicle is judged not to receive the early warning information, and the system continuously monitors until collision occurs.
Preferably, the activation threshold vehicle speed is 25km/h.
The beneficial effects of the invention are as follows:
The method and the device are based on an improved Berkeley model, and respectively determine the early warning critical distance and the braking critical distance according to different conditions, so that the method and the device are more in line with the actual conditions, and consider the influence of road surface condition conditions and the position of a vehicle brake pedal on the maximum braking deceleration of the vehicle when determining the early warning critical distance and the braking critical distance, so that the actual conditions are more in line with objective change, the degree of coincidence with the actual conditions is improved, and the problem that the early warning is not timely easily caused by determining the braking deceleration as a maximum fixed value in the traditional critical distance calculation can be effectively improved.
Aiming at the rear-end collision accident of the vehicle, the invention realizes multi-vehicle cooperative early warning and regional cooperative danger avoiding, effectively reduces the accident rate and relieves the traffic jam caused by the accident. The invention has comprehensive analysis, strong flexibility and good regional cooperative interaction effect, realizes effective communication between regional joint danger avoidance and vehicles in the vehicle-road interconnection environment, comprehensively considers the advantages and disadvantages of various sensors in different application scenes, and improves the perception performance by combining a 4D millimeter wave radar with a camera sensor. The invention establishes a safe vehicle distance calculation model for integrating the judgment of the speed of the front vehicle and the position of the brake pedal of the vehicle on the basis of a plurality of critical distance calculation models. The vehicle distance judgment is more accurate, and adverse conditions such as false alarm or untimely alarm of a system are effectively reduced; in addition, the improved Berkeley model considers the influence of objective environmental factors such as pavement, weather and the like possibly occurring in daily driving, and the early warning range is set scientifically and reasonably.
In addition, the invention brings the automatic driving vehicle into the system research object and the application range, determines the take-over time, the early warning range and the critical distance applicable to the invention according to different driving modes, and improves the application capability of the vehicle collaborative safety interaction system when meeting the current market condition of the vehicle.
Drawings
FIG. 1 is a detailed flow chart of a system operation implementation of the present invention.
Fig. 2 is an example of license plate extraction and character recognition.
Fig. 3 is a diagram of monocular vision camera imaging geometry.
Fig. 4 is a schematic view of a projection relationship between a license plate plane and an image plane.
Fig. 5 is a schematic diagram of a 4D millimeter wave radar ranging scheme.
Fig. 6 is a detailed process diagram of the braking of the automobile.
Fig. 7 is a graph of brake pedal travel depth versus brake deceleration.
Fig. 8 is a graph of brake pedal travel ratio versus brake deceleration.
Fig. 9 is an automatic driving vehicle class division diagram.
Detailed Description
Aiming at the problems of cooperative early warning and sensor ranging in the background technology, the invention combines the 4D millimeter wave radar sensor with the monocular vision camera sensor, dynamically combines the Beidou satellite navigation system, the roadside intelligent infrastructure and the traffic department based on character recognition and comprehensive sensor environment monitoring after deep learning training, realizes cooperative sharing and interaction of vehicle information in an area, and effectively carries out cooperative safety early warning of vehicles. The following description is made in connection with the specific embodiments.
The first embodiment is as follows:
The embodiment is a method for determining an early warning critical distance and a braking critical distance, wherein speed information of a front vehicle and a vehicle is input into an improved Berkeley model, and the early warning critical distance and the braking critical distance are calculated based on the Berkeley model;
In the process of calculating the early warning critical distance and the braking critical distance based on the Berkeley model, the running state of the front vehicle is divided into three conditions, and the early warning critical distance and the braking critical distance are calculated according to different conditions:
(1) When the front vehicle runs at a constant speed:
Early warning critical distance D w1:
braking critical distance D b1:
Wherein: d L is the distance travelled by the vehicle during the take over time of the driver; v 1 is the speed of the vehicle; v 2 is the front vehicle speed; a 1 is the brake deceleration of the vehicle; t 1 is the braking system reaction time; t 2 is the braking system coordination time; d 0 is the safety distance after parking;
The driving distance of the vehicle in the taking over time of the driver is d L＝v1t0; wherein t 0 is the driver take over time and v 1 is the speed of the vehicle; when the host vehicle is a non-autonomous vehicle, t 0 =0, i.e., d L =0.
(2) When the front vehicle runs at a reduced speed:
early warning critical distance D w2:
Braking critical distance D b2:
(3) When the front vehicle is stationary:
early warning critical distance D w3:
braking critical distance D b3:
The front vehicle driving state corresponds to one of three conditions, D w1、Dw2 or D w3 is used as an early warning critical distance D w, and D b1、Db2 or D b3 is used as a braking critical distance D b;
The braking deceleration a 1 of the vehicle is comprehensively determined by collecting information data such as the road surface condition of the vehicle, the position of a brake pedal and the like:
Firstly, determining a road adhesion coefficient mu based on the road type, and further determining the maximum braking deceleration according to a max =mu g, wherein g is gravity acceleration; simultaneously determining the travel depth of a brake pedal of the vehicle and braking deceleration, converting the change of the travel of the brake pedal into a proportion, and performing relation curve smoothing to obtain a relation curve graph of the travel proportion of the brake pedal and the braking deceleration of the vehicle;
Then, the brake pedal position is divided into a plurality of sections, one brake deceleration a 1 is corresponding to each section corresponding to the brake pedal position, and the brake deceleration a 1 corresponding to a section is expressed as a 1＝ki*amax,ki to represent the proportionality coefficient corresponding to the ith section according to the relation between the brake deceleration a 1 and a max corresponding to each section.
In the present embodiment, the brake pedal position is divided into six sections: the brake pedal stroke is defined as 3.5% of the first stage of braking, and then 10%, 25%, 45%, 60%, 75% of the brake pedal stroke is defined as the second to sixth stages of braking.
In the present embodiment, the braking decelerations a 1 corresponding to the six sections of the brake pedal position division are 0, 0.20a max、0.4amax、0.6amax、0.8amax, and a max, respectively.
The second embodiment is as follows: the present embodiment will be described with reference to figure 1,
The embodiment is a vehicle cooperative security interaction system fusing a 4D millimeter wave radar and a monocular vision camera, comprising:
Identifying and searching unit: character recognition is carried out on the front license plate information, and the license plate number of the front car is obtained according to character recognition and extraction; the license plate number of the front car is utilized to carry out information matching in a system database, and if the license plate number of the front car is searched in the system database, the system is also used by the front car; if the license plate number of the front car is not searched, the front car does not use the system;
the process of character recognition of the front license plate information comprises the following steps:
Firstly, a single-frame picture is extracted from a dynamic video stream of a camera, and a license plate is segmented from the picture shot by the camera. The license plate picture of the data set is required to be preprocessed before license plate recognition, the original image in the color video information extracted by the camera is converted into a gray image, the data size of the image can be greatly reduced, and meanwhile, the interference of the color on the subsequent character recognition is eliminated. And carrying out binarization processing on the preprocessed gray level image, and highlighting a license plate region in the image to determine the outline of the license plate region. Then, denoising the picture, such as: expansion, corrosion and other operations, a relatively accurate license plate outline is obtained, and then OCR is adopted to perform character recognition, and a sample of the recognition process is shown in fig. 2.
The process of character recognition of the front license plate information adopts a deep learning technology, so that the recognition accuracy can be improved, the license plate number recognition accuracy of the system is improved, and the process of recognition by adopting the deep learning technology comprises the following steps:
Based on a license plate data set of CCPD2019, the data set comprises 8 types of pictures of a plurality of scenes such as inclined, fuzzy, rainy days, snowy days and the like, the resolution of each picture is 720×1160×3, the picture comprises more than 25 ten thousand Chinese city license plate images, mainly blue license plate data, and further comprises license plate detection and identification information labels. However, in recent years, the amount of new energy vehicles is greatly increased, so that the CCPD2020 license plate data set is adopted for carrying out supplementary training, and the new energy green license plate data under different brightness, different inclination angles and different weather environments mainly comprise ten thousand total pieces. And respectively selecting 80% of samples in the two data sets as a training set, and the rest samples as a test set to test the training effect of deep learning, wherein the accuracy rate of license plate recognition after training reaches 93.7%.
Vehicle speed information acquisition unit: acquiring the running speed information of a front vehicle and a host vehicle;
In the process, if the front vehicle information is retrieved, acquiring the vehicle state information of the front vehicle provided by the Beidou satellite navigation system in an information sharing mode, and acquiring the vehicle state information of the own vehicle based on the Beidou satellite navigation system; the vehicle state information comprises position coordinates (longitude and latitude), vehicle speed and other relevant information; then extracting speed information from the vehicle state information;
if the information of the front vehicle is not retrieved or the vehicle state information of the front vehicle provided by the Beidou satellite navigation system cannot be obtained in an information sharing mode, the 4D millimeter wave radar sensor of the vehicle is adopted to obtain the speed information of the front vehicle.
Real-time vehicle distance determining unit: monitoring real-time vehicle distance by using a monocular vision camera or a 4D millimeter wave radar installed on the vehicle;
The vehicle distance calculation delay is larger and the accuracy is lower after the vehicle position information is acquired through the Beidou satellite navigation system, so that the real-time vehicle distance is monitored by using the monocular vision camera or the 4D millimeter wave radar arranged on the vehicle.
A. Acquiring license plate position information of a front vehicle in real time by using a monocular vision camera, representing the position information of the front vehicle by using the license plate position information, and calculating to obtain the real-time vehicle distance between the front vehicle and the own vehicle based on the imaging relation between the license plate position information of the front vehicle and the monocular vision camera;
The process for calculating the real-time vehicle distance between the vehicle and the front vehicle based on the imaging relation between the license plate position information of the front vehicle and the monocular vision camera comprises the following steps:
determining the projection relation between a license plate plane and an image plane of a front vehicle based on the imaging characteristics of a monocular vision camera; according to the imaging principle of the camera, the height of the camera of the vehicle running on the horizontal road surface is unchanged, so that the bottom edge line of the front vehicle has a corresponding relationship with the distance between the two vehicles, as shown in fig. 3.
Substituting the image plane characteristics of the front vehicle obtained through perception into the following monocular vision camera distance calculation formula, and obtaining the distance D between the front vehicle and the front vehicle.
And (3) a space rectangular coordinate system is established by taking the camera point C as the center, so that the Z axis is parallel to the main optical axis of the camera, X π is an imaging point of an image plane, the corresponding relation of the license plate plane and the image plane in the figure 4 exists, and the data of the relevant initial installation angle, focal length, pixel height, observation angle and the like of the vehicle-mounted camera are substituted into a formula, so that the distance D between two vehicles can be obtained.
Wherein D is the distance between the front vehicle and the vehicle; h is the height of the camera from the road surface; θ c is the included angle between the camera and the vertical axis, namely the initial installation angle of the camera; h is the pixel height; h i is the height of the effective imaging surface of the camera; d p is the effective size of the imaging plane pixel; alpha is the observable angle of the camera.
B. Acquiring the real-time vehicle distance of a front vehicle in real time by using a 4D millimeter wave radar;
Because the 4D millimeter wave radar has higher measurement accuracy in short distance and low speed, namely when the distance between two vehicles reaches the pre-warning critical distance or the vehicle does not reach the activation critical speed (namely, the vehicle speed is less than 25 km/h), and when the monocular vision camera fails (such as the conditions of heavy fog, snow burst, heavy rain and the like, the 4D millimeter wave radar can be triggered according to a preset mode when the condition that the monocular vision camera fails is achieved), the 4D millimeter wave radar sensor is selected for ranging, and otherwise, the monocular vision camera is utilized for monitoring the vehicle distance change.
Radar and cameras are common sensors for measuring vehicle distance. They each have advantages and disadvantages. Because the radar sensor is not affected by weather, and the measurement is accurate for the low-speed stage of the vehicle just started, but the radar is difficult to distinguish the details of the target, the radar sensor is not ideal for the application requiring more detailed information, and the monocular vision camera sensor is adopted for monitoring when the vehicle speed is large and the vehicle distance is far. Therefore, the system has the advantages of low cost, low maintenance cost and popularization, and the defect of supplementing the monocular camera sensor by adopting the 4D millimeter wave radar sensor.
The 4D millimeter wave radar installed on the vehicle can provide environment sensing information in four dimensions of distance, horizontal positioning, vertical positioning and speed for the system, the speed measurement accuracy data of the system depend on signal to noise ratio, and the specific measurement principle is shown in fig. 5.
Meanwhile, under the condition that a camera cannot be used for ranging and extracting license plates when a lane is changed, overtaking and the like, the system adopts a 4D millimeter wave radar sensor to detect short-distance vehicle distance information, tangential movement and angle of a vehicle in real time.
Auxiliary information acquisition unit: acquiring the passing condition of the vehicle based on the road monitoring camera, namely judging whether the road is congested and slowly moving;
The congestion and creep of the traffic flow are judged through the existing internet of vehicles technology and a road monitoring camera, the time required for the traffic to pass through the road is obtained based on the Beidou satellite navigation system, information interaction is established between the traffic flow and a system database, the average time required for the traffic to pass through the road is comprehensively calculated, if the required time is larger than the average time, the traffic flow and creep of the road are judged, and the system database shares information to the traffic in the system.
An early warning information activating unit: based on the speed of the vehicle monitored in real time, judging whether the speed exceeds an activation critical speed, and further determining that the early warning information calculating unit is in a resting or active state;
When judging the applicable vehicle speed range of the system, the daily running condition of the vehicle is fully considered, and the average running speed of the commute peak of the urban road is close to the minimum speed of the daily running, so that the system has reference significance, and the commute peak is taken into the consideration range of the system.
The actual speed of the commute peak in 2022 can be found to be greater than 25km/h based on the index reflecting urban traffic conditions obtained by mining and calculating mass traffic travel data, vehicle track data and position service data of the hundred-degree map. The commute peak actual travel speed should be the lowest value of the vehicle travel speed in the city and the average travel speed should be greater than the commute peak actual travel speed in normal smooth travel conditions. If the vehicle runs in the expressway, the lowest speed limit is 60km/h, and the running speed is lower than 60km/h, which belongs to the traffic illegal action, and the running speed of the vehicle on the expressway is necessarily higher than the actual running speed of the commute peak, so that the invention determines the activation critical speed as 25km/h for meeting two application scenes of urban roads and expressways.
When the speed of the vehicle is greater than 25km/h, the probability of collision accidents of the vehicle is increased, so that the early warning information calculation unit and the 4D millimeter wave radar enter an active running state; when the speed of the vehicle is not more than 25km/h, the early warning information calculation unit is in a resting state, and at the moment, the real-time monitoring and collision early warning of the whole distance of the vehicle are only carried out based on the 4D millimeter wave mine, and the situation is needed to be explained, namely, the collision detection is carried out by utilizing the distance information measured by the 4D millimeter wave mine, and the process is only needed to be carried out by adopting the prior art, so that the invention does not carry out excessive explanation.
Early warning information calculation unit: substituting the speed information of the front vehicle and the speed information of the own vehicle into an improved Berkeley model, and calculating to obtain an early warning critical distance and a braking critical distance;
Most of the existing vehicle early warning modes are judged based on the state of the vehicle, so that various factors need to be comprehensively considered when a critical distance calculation model is determined, and adverse conditions such as untimely warning or false warning are avoided. The early warning distance and the braking distance are set in the basic Berkeley model and are matched with the data required by the system, so that the system is improved on the basis of the Berkeley model.
The invention fully considers the technical development of the existing vehicle, considers the application range of the system for the automatic driving automobile, and reflects the influence of the automatic driving automobile in a calculation model of the critical distance. For an autonomous vehicle, the time required for the driver to take over control (i.e., take over time) needs to be considered.
The early warning distance and the braking distance are affected by factors such as a driver, a vehicle and the environment, and according to the analysis of the braking process of the automobile, it can be found that when the driver brakes, the whole process mainly comprises four stages, as shown in fig. 6, respectively: a braking reaction phase, a braking coordination phase, a braking duration phase and a braking stop phase.
The vehicle runs in daily life, the critical distance between the early warning and the braking can be calculated according to the speed and the position information of the front vehicle and the vehicle, and the critical distance between the early warning and the braking can be calculated according to different running states of the front vehicle:
(1) Front vehicle at uniform speed
When the front vehicle runs at a constant speed, if collision is possible, the speed of the front vehicle is longer than that of the front vehicle. When such a situation occurs, in order to avoid collision, the vehicle needs to be decelerated to be not more than the speed of the front vehicle in time, and the running state of the vehicle is analyzed to know that:
Firstly, according to vehicle registration information of a system database, whether the vehicle is an automatic driving vehicle is determined, if the vehicle is the automatic driving vehicle, the driving distance of the vehicle in the taking over time of a driver is as follows: d L＝v1t0; wherein t 0 is the driver take over time; typically 1.5 to 2.0s, here 1.8s; v 1 is the speed of the vehicle;
For the vehicle to be a non-autonomous vehicle, the driver take-over time is not considered during braking, i.e. at this time t 0 =0, so d L =0.
The early warning and braking critical distances at this time are respectively:
Early warning critical distance D w1:
braking critical distance D b1:
Wherein: v 1 is the speed of the vehicle; v 2 is the front vehicle speed; a 1 is the brake deceleration of the vehicle; t 1 is the reaction time of the braking system, and is generally 0.1 to 0.2s, here 0.15s; t 2 is the coordination time of the braking system, and is generally 0.3s; d 0 is the safety distance after parking, and is generally 1m. Wherein a 1 selects the road surface condition needing to comprehensively consider the running of the vehicle, and the speed of the vehicle is larger than the position of a brake pedal of the vehicle when the vehicle is in front.
The road surface condition has a larger influence on the maximum braking deceleration of the vehicle, the traditional critical distance calculation model has fewer consideration on the maximum braking deceleration, the braking deceleration is determined to be a maximum fixed value, and the condition of untimely early warning is easy to occur. Therefore, in the critical distance calculation model, different maximum braking deceleration of the vehicle is adopted according to different road adhesion coefficients of the driving environment, and the specific selection is shown in table 1.
Maximum braking deceleration achievable by the vehicle: a max =μg; mu is the road adhesion coefficient; g=9.8 m/s 2.
TABLE 1 road attachment coefficient and maximum braking deceleration for different road conditions
The relationship between the brake pedal travel depth and the brake deceleration of the vehicle is shown in fig. 7, and the curve is a wave-like rise. The change of the brake pedal stroke is converted into a proportion, and the relation curve is smoothed, so that the relation curve graph of the brake pedal stroke proportion and the vehicle brake deceleration of the figure 8 can be obtained.
The brake pedal position can be divided into six phases according to two diagrams. In the 0-3.5% interval, namely in the initial stage of braking, a certain physical gap exists between the pedal and the friction plate, and the brake can generate obvious deceleration after overcoming the idle stroke of the brake pedal, so that the 3.5% of the stroke of the brake pedal is defined as the first stage of braking, and then the stroke of the brake pedal is divided into four stages of 10%, 25%, 45% and 60%. According to national regulations, when the brake pedal travel is greater than 75%, the brake system is required to provide maximum brake pressure, at which point the vehicle reaches maximum brake deceleration, thus defining 75% of the pedal travel as the sixth stage of braking. The maximum braking deceleration for the different braking phases for the different ratios is shown in table 2.
TABLE 2 brake pedal travel and brake deceleration
When the speed of the vehicle is greater than the speed of the preceding vehicle, the maximum braking deceleration a max of the vehicle is determined according to the road surface condition of the vehicle, and then the braking deceleration a 1 of the vehicle is determined according to the acquired position of the braking pedal of the vehicle, and then the critical distance is calculated according to the table 2.
(2) Front vehicle deceleration
When the vehicle is driven along with the vehicle, if the front vehicle suddenly and emergently decelerates, the driver does not react timely, and the rear-end collision is most likely to happen. Therefore, the system calculates and monitors the distance between the front vehicle and the own vehicle in real time, and if the distance between the front vehicle and the own vehicle is recognized to reach the calculated early warning critical distance, the system needs to perform early warning. When two vehicles are braked and parked, under the same road surface condition, the maximum emergency braking deceleration of the two vehicles is the same, and according to the analysis of the speed and the braking deceleration of the two vehicles, the early warning and the braking critical distance at the moment are respectively as follows:
early warning critical distance D w2:
Braking critical distance D b2:
(3) Stationary front vehicle
When the front vehicle is stationary, namely v 2 is equal to 0 at the moment, in order to avoid rear-end collision accidents, the vehicle needs to brake in time, and then the early warning and the critical distance calculation of the braking only need to consider v 1, which are respectively:
early warning critical distance D w3:
braking critical distance D b3:
Since the preceding vehicle is one of the three conditions, D w1、Dw2 or D w3 is taken as the pre-warning critical distance D w, and D b1、Db2 or D b3 is taken as the braking critical distance D b;
Information data such as road surface condition of the vehicle, brake pedal position and the like are collected, and the brake deceleration a 1 of the vehicle is comprehensively determined.
Early warning judging and prompting unit: comparing the distance D between the front vehicle and the vehicle with the pre-warning critical distance D w;
if the distance between the two vehicles is larger than the pre-warning critical distance, continuously monitoring;
If the distance between the two vehicles is not greater than the pre-warning critical distance, acquiring the license plate number of the front vehicle and the result of whether the system is used or not by the identification and retrieval unit, acquiring registration information of the front vehicle and the self vehicle based on a system database, and determining whether the front vehicle and the self vehicle are automatic driving vehicles or not; carrying out early warning prompt and judging whether a driver takes effective measures or not;
Current autopilot technology is rapidly evolving and has increasingly been used, so the system takes autopilot vehicles into account. The process of determining whether the vehicle is an autonomous vehicle is accomplished based on registration information stored in a system database, and if the vehicle uses the present system, initial information of the vehicle needs to be registered in the system, including whether the vehicle is an autonomous car and an autonomous class. According to the automatic driving classification shown in the following fig. 9, the L2 still needs the human driver to monitor the surrounding environment, and the design concept of cooperative communication early warning with the vehicle road of the system is not consistent, so that the automatic driving vehicles of the L3 level and above are considered, and the vehicles can avoid risks in advance by means of the omnibearing perception of the system.
If the front vehicle also uses the system, the front vehicle and the vehicle are jointly subjected to early warning prompt;
if the front vehicle does not use the system, the vehicle performs self-early warning;
After early warning, the attitude data of the vehicle steering, braking and the like of the vehicle or two vehicles (a front vehicle and the vehicle) are monitored in real time, whether a driver takes effective measures or not is judged, and the set threshold value of whether the driver takes the effective measures or not respectively considers two conditions of the steering and the braking of the vehicle:
(1) Turning: if the transverse offset threshold value of the vehicle exceeds 0.2m, the system judges that the driver definitely carries out the current early warning prompt and then consciously carries out steering operation;
(2) Braking: if the travel threshold of the brake pedal of the vehicle reaches 10% of the maximum value, judging that the driver knows the current early warning prompt and then consciously performing braking operation;
If one of the two conditions of the vehicle posture is monitored, judging that the driver receives the early warning information and timely taking effective measures; if the vehicle does not take effective measures, the vehicle is judged not to receive the early warning information, and the system continuously monitors until collision occurs.
Braking decision and emergency early warning unit: determining a vehicle distance change state in real time based on the distance D between the front vehicle and the own vehicle, and comparing the distance D between the two vehicles with a braking critical distance D b;
If the distance between two vehicles is continuously increased, the system monitors that the distance between two vehicles is larger than the pre-warning critical distance, and then the dangerous state is relieved;
If the vehicle distance between the two vehicles is continuously reduced, the system enters an emergency early warning state; if the distance between two vehicles is continuously reduced to the braking critical distance, aiming at the vehicle being an automatic driving vehicle, the system sends an instruction to the vehicle to automatically take emergency braking; aiming at the fact that the vehicle is a non-automatic driving vehicle, the system sends out vehicle collision early warning, and meanwhile continuously monitors until collision occurs.
If the front vehicle also uses the system, the front vehicle is also subjected to vehicle collision early warning;
An accident information sharing unit: after collision, based on vehicle longitude and latitude information acquired by the Beidou satellite navigation system during collision (when the data during collision is missing, the data is taken as the time of collision immediately before collision), the database searches relevant data such as the automatic driving grade of the vehicle, the vehicle running speed, the collision early warning condition and the like, and uploads the relevant data to the database, and then based on the vehicle longitude and latitude information during collision, the early warning range of the accident site is calculated and determined, accident reminding is sent to vehicles using the system in the range, and surrounding vehicles are guided to avoid.
Because the traffic accident causes the crowded influence scope is great, in order to ensure the travelling comfort of other vehicles, the travelling direction of the specified accident site is the affected area De of accident in 100m, therefore the early warning scope D A of this system is:
For vehicles that reach an activation threshold vehicle speed:
DA＝DW+De
Wherein, D e is the early warning distance, which is 100m in the present embodiment; dw is the critical distance of early warning;
For vehicles which do not reach the activation critical speed, the collision is monitored only by using a 4D millimeter wave radar, and the situation of D W cannot be obtained, then D A＝De =100deg.C;
Meanwhile, the related accident data is synchronized with the traffic departments, the related departments are prompted to process and solve, timely information is provided for vehicle evacuation in the following accident vehicle rescue and early warning range, and the auxiliary information acquisition unit can be used for carrying out collaborative processing on accidents through road side traffic control and guidance facilities such as intersection traffic lights, electronic guideboard information display and the like based on whether traffic flow states monitored by the road monitoring cameras are congested and slow; after the accident is solved, the traffic department can issue accident release information through the system and restore the road side facilities to the state before the accident.
In the process of sending accident reminding to the vehicles using the system in the range and guiding surrounding vehicles to avoid, the prompt receiving condition of the vehicles of the system can be monitored, the feedback signal of the vehicles of the system in the early warning range is obtained, and if the vehicles timely feed back and effectively avoid the accident area, the system monitors the vehicles to leave the early warning range and then releases the early warning; if the vehicle does not provide feedback information in time within the limited feedback time, the vehicle is subjected to early warning prompt on the vehicle which is not fed back.
For non-autonomous vehicles, i.e. vehicles operated by the driver, the response speed of the driver to the early warning signal is limited during driving, and the response time is about 0.5s under the general condition. The selective judging time required to carry out complex analysis can reach 2s, and the effective feedback time of the driver is preliminarily determined to be 3s according to the individual differences of the driver.
For an autonomous vehicle, the reaction time of the system is far less than that of the driver, but after receiving the early warning, the driver selects to take over driving right, and the taking over time needs to be considered in a limited feedback time. To sum up, to ensure the consistency of the system, the limited feedback duration is determined to be 3s.
If the feedback time is within the limited feedback time, the system receives the feedback signal of the system vehicle in the early warning range, and then the system monitors the vehicle to leave the early warning range in real time and releases the early warning; if the system does not receive the feedback signal, the vehicle early warning prompt is carried out on the vehicle which is not fed back.
According to the invention, the 4D millimeter wave radar is combined with the monocular vision camera, so that the problem of low current vehicle distance measurement precision is effectively solved, and the vehicle surrounding systems are reminded of avoiding vehicles through cooperative early warning and effective feedback of the system, so that traffic jam caused by accidents is greatly relieved. The working process of the system comprises the following steps:
Acquiring relevant state information such as vehicle speed, longitude and latitude of a system by using a Beidou satellite navigation system, acquiring vehicle passing conditions by using intelligent infrastructures such as road monitoring, monitoring the vehicle speed in real time by using the system, judging the vehicle speed in advance, and entering an active state by using the system if the vehicle speed reaches the system condition;
And secondly, collecting auxiliary information data of the system vehicle, substituting the information of the speed of the front vehicle and the speed of the vehicle meeting the preset speed condition of the system into an improved Berkeley model, and determining the critical distance between early warning and braking of the system vehicle through calculation.
Step three, using a monocular vision camera to collect and capture the information of the license plate of the front vehicle in real time, calculating the distance between the vehicles, determining the distance between the two vehicles, and comparing the calculation result with the pre-warning critical distance;
Step four, if the distance between the two vehicles is larger than the pre-warning critical distance, the system monitors in real time; otherwise, the front license plate is extracted based on the character recognition of the deep learning of the camera sensor, and the system service condition of the front license is judged;
Fifthly, if the front vehicle uses the system, the front vehicle and the rear vehicle carry out early warning prompt, and monitor whether the two vehicles take effective measures such as braking, steering and the like; if the front vehicle does not use the system, the vehicle pre-warns itself;
Step six, judging whether the distance between the two vehicles reaches a braking critical distance, checking whether the vehicle is an automatic driving vehicle, and if the distance is continuously reduced, performing self-adaptive forced braking on the automatic driving vehicle and performing early warning on the non-automatic driving vehicle; if the vehicle distance is increased, the two vehicles are released from the alarm state;
Step seven, measuring distance of the driving environment in real time based on a 4D millimeter wave radar, monitoring the surrounding environment of the vehicle body, uploading accident data such as vehicle position information, vehicle damage degree and the like to a system if the dangerous distance is reached, determining an early warning range by the system through backstage scheduling, carrying out regional collaborative accident early warning of the system vehicle based on intelligent traffic infrastructure, otherwise, monitoring the vehicle to a safe vehicle distance state in real time by the system;
And step eight, monitoring feedback signals of the system vehicles in the area. If the vehicle does not provide feedback information in time within the effective feedback time, the vehicle early warning prompt is carried out; aiming at timely avoiding of effective feedback vehicles, the system monitors the vehicles to the extent that the vehicles exit the early warning range.
The above examples of the present invention are only for describing the calculation model and calculation flow of the present invention in detail, and are not limiting of the embodiments of the present invention. Other variations and modifications of the above description will be apparent to those of ordinary skill in the art, and it is not intended to be exhaustive of all embodiments, all of which are within the scope of the invention.

Claims (10)

1.一种预警临界距离、制动临界距离确定方法，其特征在于，将前车和本车的车速信息输入改进的伯克利模型，基于伯克利模型计算得到预警临界距离与制动临界距离；1. A method for determining a critical warning distance and a critical braking distance, characterized in that the speed information of the preceding vehicle and the vehicle is input into an improved Berkeley model, and the critical warning distance and the critical braking distance are calculated based on the Berkeley model; 基于伯克利模型计算得到预警临界距离与制动临界距离过程中，将前车行驶状态分为三种情况，针对不同情况计算得到预警临界距离与制动临界距离：In the process of calculating the warning critical distance and braking critical distance based on the Berkeley model, the driving state of the front vehicle is divided into three situations, and the warning critical distance and braking critical distance are calculated for different situations: (1)前车匀速行驶时：(1) When the vehicle ahead is traveling at a constant speed: 预警临界距离Dw1：Warning critical distance Dw1 : 制动临界距离Db1：Critical braking distance D b1 : 式中：dL为驾驶员接管时间内车辆的行驶距离；v1为本车速度；v1为前车速度；a1为本车制动减速度；t1为制动系统反应时间；t2为制动系统协调时间；d0为停车后的安全车距；Where: d L is the distance traveled by the vehicle during the driver's takeover time; v 1 is the speed of the vehicle; v 1 is the speed of the preceding vehicle; a 1 is the braking deceleration of the vehicle; t 1 is the braking system reaction time; t 2 is the braking system coordination time; d 0 is the safe distance after parking; (2)前车减速行驶时：(2) When the vehicle ahead is slowing down: 预警临界距离Dw2：Warning critical distance Dw2 : 制动临界距离Db2：Critical braking distance D b2 : (3)前车静止时：(3) When the vehicle ahead is stationary: 预警临界距离Dw3：Warning critical distance Dw3 : 制动临界距离Db3：Critical braking distance D b3 : 前车行驶状态分对应为三种情况之一，将Dw1、Dw2或Dw3作为预警临界距离Dw，将Db1、Db2或Db3作为制动临界距离Db；The driving state of the preceding vehicle corresponds to one of the three situations, and D w1 , D w2 or D w3 is used as the warning critical distance D w , and D b1 , D b2 or D b3 is used as the braking critical distance D b ; 所述的本车的制动减速度a1通过采集本车行驶的路面状况条件、制动踏板位置信息数据综合确定：The braking deceleration a1 of the vehicle is determined by collecting the road conditions and brake pedal position information data of the vehicle: 首先基于路面类型确定路面附着系数μ，进而根据amax＝μg确定最大制动减速度，g为重力加速度；同时确定车辆的制动踏板行程深度与制动减速，将制动踏板行程的变化转化为比例，并进行关系曲线平滑处理，得到制动踏板行程比例与车辆制动减速度的关系曲线图；First, the road adhesion coefficient μ is determined based on the road type, and then the maximum braking deceleration is determined according to a max = μg, where g is the acceleration of gravity; at the same time, the vehicle's brake pedal travel depth and braking deceleration are determined, the change in brake pedal travel is converted into a ratio, and the relationship curve is smoothed to obtain a curve diagram of the relationship between the brake pedal travel ratio and the vehicle's braking deceleration; 然后将制动踏板位置划分为多个区间段，并为制动踏板位置对应的每个区间段对应一个制动减速度a1，根据每个区间段对应的制动减速度a1与amax的关系，将某区间段对应的制动减速度a1表示为a1＝ki*amax，ki表示第i个区间所对应的比例系数。Then the brake pedal position is divided into multiple intervals, and a braking deceleration a 1 is assigned to each interval corresponding to the brake pedal position. According to the relationship between the braking deceleration a 1 and a max corresponding to each interval, the braking deceleration a 1 corresponding to a certain interval is expressed as a 1 = k i * a max , where k i represents the proportional coefficient corresponding to the i-th interval. 2.根据权利要求1所述的一种预警临界距离、制动临界距离确定方法，其特征在于，所述的驾驶员接管时间内车辆的行驶距离为dL＝v1t0；其中，t0为驾驶员接管时间，v1为本车速度；当本车为非自动驾驶车辆时，t0＝0，即dL＝0。2. A method for determining a critical warning distance and a critical braking distance according to claim 1, characterized in that the distance traveled by the vehicle during the driver takeover time is d L =v 1 t 0 ; wherein t 0 is the driver takeover time, and v 1 is the vehicle speed; when the vehicle is a non-automatic driving vehicle, t 0 =0, that is, d L =0. 3.根据权利要求1后2所述的一种预警临界距离、制动临界距离确定方法，其特征在于，将制动踏板位置划分为六个区间段：将制动踏板行程3.5％定为制动第一阶段，而后将制动踏板行程的10％、25％、45％、60％、75％定为制动第二至第六阶段。3. A method for determining a warning critical distance and a braking critical distance according to claim 1 or 2, characterized in that the brake pedal position is divided into six intervals: 3.5% of the brake pedal stroke is set as the first stage of braking, and then 10%, 25%, 45%, 60%, and 75% of the brake pedal stroke are set as the second to sixth stages of braking. 4.根据权利要求3所述的一种预警临界距离、制动临界距离确定方法，其特征在于，制动踏板位置划分的六个区间段对应的制动减速度a1分别为0、0.20amax、0.4amax、0.6amax、0.8amax和amax。4. A method for determining a critical warning distance and a critical braking distance according to claim 3, characterized in that the six intervals divided by the brake pedal position correspond to braking decelerations a1 of 0 , 0.20a max , 0.4a max , 0.6a max , 0.8a max and a max respectively. 5.一种融合4D毫米波雷达和单目视觉摄像头的车辆协同安全交互系统，其特征在于，包括：5. A vehicle cooperative safety interaction system integrating 4D millimeter wave radar and monocular vision camera, characterized by comprising: 识别检索单元：对前车牌照信息进行字符识别，根据字符识别提取得到前车的车牌号码；利用前车的车牌号码在系统数据库进行信息的匹配，若在系统数据库中检索到前车牌照号码，即前车也使用本系统；若未检索到前车牌照号码，则前车未使用本系统；Identification and retrieval unit: perform character recognition on the license plate information of the preceding vehicle, and extract the license plate number of the preceding vehicle based on the character recognition; use the license plate number of the preceding vehicle to match the information in the system database. If the license plate number of the preceding vehicle is retrieved in the system database, it means that the preceding vehicle also uses this system; if the license plate number of the preceding vehicle is not retrieved, it means that the preceding vehicle does not use this system; 车辆速度信息获取单元：获取前车与本车的行驶速度信息；Vehicle speed information acquisition unit: acquires the driving speed information of the preceding vehicle and the vehicle itself; 如果检索到前车信息，则以信息共享的方式获取基于北斗卫星导航系统提供的前车的车辆状态信息，并基于本身的北斗卫星导航系统获取本车的车辆状态信息；所述的车辆状态信息包括位置坐标、车辆速度信息；然后在车辆状态信息中提取速度信息；If the preceding vehicle information is retrieved, the vehicle status information of the preceding vehicle provided by the Beidou satellite navigation system is obtained in an information sharing manner, and the vehicle status information of the vehicle is obtained based on the Beidou satellite navigation system of the vehicle itself; the vehicle status information includes the position coordinates and the vehicle speed information; and then the speed information is extracted from the vehicle status information; 如果未检索到前车信息，或者无法通过信息共享的方式获取基于北斗卫星导航系统提供的前车的车辆状态信息，则采用本车的4D毫米波雷达传感器获取前车的速度信息；If the preceding vehicle information is not retrieved, or the vehicle status information of the preceding vehicle provided by the Beidou satellite navigation system cannot be obtained through information sharing, the vehicle's 4D millimeter-wave radar sensor is used to obtain the preceding vehicle's speed information; 实时车距确定单元：当两车间距达到预警临界距离后或车辆未达到激活临界车速时，以及当单目视觉摄像头失效时，选用4D毫米波雷达传感器监测实时车距，反之利用单目视觉摄像头监测实时车距；Real-time vehicle distance determination unit: When the distance between two vehicles reaches the critical warning distance or the vehicle does not reach the critical activation speed, or when the monocular vision camera fails, the 4D millimeter wave radar sensor is used to monitor the real-time vehicle distance, otherwise the monocular vision camera is used to monitor the real-time vehicle distance; 预警信息激活单元：基于实时监控的本车速度，判断是否超过激活临界车速，进而确定预警信息计算单元处于静息或活跃状态；Warning information activation unit: Based on the real-time monitored vehicle speed, it determines whether it exceeds the critical activation speed, and then determines whether the warning information calculation unit is in a dormant or active state; 预警信息计算单元：采用权利要求1至4任意一项所述的一种预警临界距离、制动临界距离确定方法计算得到预警临界距离Dw与制动临界距离Db；Warning information calculation unit: using a warning critical distance and a braking critical distance determination method according to any one of claims 1 to 4 to calculate a warning critical distance Dw and a braking critical distance Db ; 预警判断与提示单元：将前车与本车的间距D与预警临界距离Dw进行比较；Warning judgment and prompt unit: compares the distance D between the front vehicle and the vehicle with the warning critical distance Dw ; 若两车间距大于预警临界距离，进行持续监控；If the distance between the two vehicles is greater than the critical warning distance, continuous monitoring will be performed; 若两车间距不大于预警临界距离，获取识别检索单元得到前车的车牌号码及是否使用本系统的结果，基于系统数据库获取前车和本车的登记信息，确定前车和本车是否为自动驾驶车辆；进行预警提示，同时判断驾驶员是否采取有效措施；If the distance between the two vehicles is not greater than the critical warning distance, the recognition and retrieval unit obtains the license plate number of the preceding vehicle and the result of whether the system is used, obtains the registration information of the preceding vehicle and the present vehicle based on the system database, and determines whether the preceding vehicle and the present vehicle are autonomous driving vehicles; issues a warning prompt and determines whether the driver has taken effective measures; 制动决策及紧急预警单元：基于前车与本车的间距D实时确定车距变化状态，并将两车间距D与制动临界距离Db进行比较；Braking decision and emergency warning unit: determines the distance change state in real time based on the distance D between the preceding vehicle and the vehicle, and compares the distance D between the two vehicles with the critical braking distance Db ; 如果两车间距持续增大，则系统监控至两车间距大于预警临界距离后解除危险状态；If the distance between the two vehicles continues to increase, the system will release the dangerous state after monitoring until the distance between the two vehicles exceeds the warning critical distance; 如果两车车距持续减小，则系统进入紧急预警状态；若两车间距持续缩小至制动临界距离，针对本车为自动驾驶车辆，系统向车辆发送指令自行采取紧急制动；针对本车为非自动驾驶车辆，系统发出车辆碰撞预警，同时持续进行监控至发生碰撞；若前车也使用本系统，则对前车也进行车辆碰撞预警；If the distance between the two vehicles continues to decrease, the system will enter an emergency warning state; if the distance between the two vehicles continues to decrease to the critical braking distance, if the vehicle is an autonomous vehicle, the system will send a command to the vehicle to perform emergency braking on its own; if the vehicle is not an autonomous vehicle, the system will issue a vehicle collision warning and continue to monitor until a collision occurs; if the vehicle in front also uses this system, the vehicle collision warning will also be issued to the vehicle in front; 事故信息共享单元：碰撞发生后，基于北斗卫星导航系统采集的碰撞时车辆位置信息，数据库检索车辆的自动驾驶等级、车辆行驶速度、碰撞预警情况数据上传至数据库，然后基于碰撞发生时车辆位置信息，计算确定事故地点的预警范围，对预警范围内的使用本系统车辆发送事故提醒。Accident information sharing unit: After a collision occurs, based on the vehicle position information at the time of collision collected by the Beidou satellite navigation system, the database retrieves the vehicle's automatic driving level, vehicle speed, and collision warning data and uploads them to the database. Then, based on the vehicle position information at the time of collision, the warning range of the accident location is calculated and determined, and accident reminders are sent to vehicles using this system within the warning range. 6.根据权利要求5所述的一种融合4D毫米波雷达和单目视觉摄像头的车辆协同安全交互系统，其特征在于，所述的预警范围DA如下：6. According to the vehicle cooperative safety interaction system integrating 4D millimeter wave radar and monocular vision camera in claim 5, it is characterized in that the warning range D A is as follows: 针对达到激活临界车速的车辆：For vehicles reaching the activation threshold speed: DA＝DW+De D A = D W + D e 其中，De为预警距离，DW为预警临界距离；Among them, De is the warning distance, and D W is the warning critical distance; 针对未达到激活临界车速的车辆，利用4D毫米波雷达进行碰撞监测，忽略DW，令DA＝De。For vehicles that have not reached the critical activation speed, 4D millimeter-wave radar is used for collision monitoring, D W is ignored, and D A =D e is set. 7.根据权利要求5所述的一种融合4D毫米波雷达和单目视觉摄像头的车辆协同安全交互系统，其特征在于，利用单目视觉摄像头监测实时车距时，首先利用单目视觉摄像头实时采集前车的车牌位置信息，以车牌位置信息表示前车的位置信息，然后基于前车的车牌位置信息与单目视觉摄像头的成像关系计算得到本车与前车的实时车距。7. According to claim 5, a vehicle collaborative safety interaction system integrating 4D millimeter-wave radar and monocular vision camera is characterized in that when using the monocular vision camera to monitor the real-time vehicle distance, the monocular vision camera is first used to collect the license plate position information of the front vehicle in real time, and the license plate position information is used to represent the position information of the front vehicle. Then, the real-time vehicle distance between the vehicle and the front vehicle is calculated based on the imaging relationship between the license plate position information of the front vehicle and the monocular vision camera. 8.根据权利要求7所述的一种融合4D毫米波雷达和单目视觉摄像头的车辆协同安全交互系统，其特征在于，基于前车的车牌位置信息与单目视觉摄像头的成像关系计算得到本车与前车的实时车距的过程包括以下步骤：8. According to the vehicle cooperative safety interaction system integrating 4D millimeter wave radar and monocular vision camera as described in claim 7, it is characterized in that the process of calculating the real-time vehicle distance between the vehicle and the front vehicle based on the license plate position information of the front vehicle and the imaging relationship of the monocular vision camera comprises the following steps: 基于单目视觉摄像头的成像特点，确定前车车牌平面与像平面投影关系；根据摄像头的成像原理可知，在水平路面上行驶的车辆摄像头高度不变，则前车的底边线与两车距离存在对应关系，如图3所示；Based on the imaging characteristics of the monocular vision camera, the projection relationship between the license plate plane of the front vehicle and the image plane is determined; according to the imaging principle of the camera, if the camera height of the vehicle traveling on a horizontal road remains unchanged, then there is a corresponding relationship between the bottom line of the front vehicle and the distance between the two vehicles, as shown in Figure 3; 将感知所得前车的像平面特征代入下述单目视觉摄像头车距计算公式，即可得到前车与本车的两车间距D；Substituting the perceived image plane features of the front vehicle into the following monocular vision camera distance calculation formula, the distance D between the front vehicle and the vehicle can be obtained; 以摄像头点C为中心，建立空间直角坐标系，令Z轴与摄像头主光轴平行，Xπ为像平面的成像点，则车牌平面与像平面存在下图4的对应关系，将车载摄像头的相关初始安装角度、焦距、像素高度、观测角度代入公式，即可求得两车间距D；With the camera point C as the center, a spatial rectangular coordinate system is established, and the Z axis is parallel to the main optical axis of the camera. Xπ is the imaging point of the image plane. Then the license plate plane and the image plane have a corresponding relationship as shown in Figure 4 below. Substituting the relevant initial installation angle, focal length, pixel height, and observation angle of the vehicle-mounted camera into the formula, the distance D between the two vehicles can be obtained; 其中，D为前车与本车间距；H为摄像头距路面的高度；θc为摄像头与竖轴夹角，即摄像头的初始安装角度；h为像素高度；hi为摄像机有效成像面的高度；dp为成像平面像素的有效尺寸；α为摄像头可观测的角度。Where D is the distance between the front vehicle and the vehicle itself; H is the height of the camera from the road; θc is the angle between the camera and the vertical axis, that is, the initial installation angle of the camera; h is the pixel height; hi is the height of the effective imaging surface of the camera; dp is the effective size of the pixel on the imaging plane; and α is the observable angle of the camera. 9.根据权利要求5所述的一种融合4D毫米波雷达和单目视觉摄像头的车辆协同安全交互系统，其特征在于，预警判断与提示单元判断驾驶员是否采取有效措施的过程中，判断驾驶员是否采取有效措施是基于设定的转向阈值和制动预制阈值确定，当车辆的横向偏移量超过转向阈值，或者车辆的制动踏板行程超过制动预制阈值，判定驾驶员接收到预警信息并采取了有效措施；若车辆未采取有效措施，则判定车辆未接收到预警信息，系统持续进行监控至发生碰撞。9. According to claim 5, a vehicle collaborative safety interaction system integrating 4D millimeter-wave radar and monocular vision camera is characterized in that, in the process of the early warning judgment and prompting unit judging whether the driver has taken effective measures, the judgment of whether the driver has taken effective measures is determined based on the set steering threshold and braking pre-threshold. When the lateral offset of the vehicle exceeds the steering threshold, or the brake pedal stroke of the vehicle exceeds the braking pre-threshold, it is determined that the driver has received the early warning information and taken effective measures; if the vehicle has not taken effective measures, it is determined that the vehicle has not received the early warning information, and the system continues to monitor until a collision occurs. 10.根据权利要求5至9任意一项所述的一种融合4D毫米波雷达和单目视觉摄像头的车辆协同安全交互系统，其特征在于，所述激活临界车速为25km/h。10. A vehicle collaborative safety interaction system integrating a 4D millimeter-wave radar and a monocular vision camera according to any one of claims 5 to 9, characterized in that the critical activation speed is 25 km/h.
CN202410520987.3A 2024-04-28 2024-04-28 Method for determining critical warning distance and critical braking distance, and vehicle cooperative safety interaction system integrating 4D millimeter wave radar and monocular vision camera Pending CN118538053A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410520987.3A CN118538053A (en) 2024-04-28 2024-04-28 Method for determining critical warning distance and critical braking distance, and vehicle cooperative safety interaction system integrating 4D millimeter wave radar and monocular vision camera

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410520987.3A CN118538053A (en) 2024-04-28 2024-04-28 Method for determining critical warning distance and critical braking distance, and vehicle cooperative safety interaction system integrating 4D millimeter wave radar and monocular vision camera

Publications (1)

Publication Number Publication Date
CN118538053A true CN118538053A (en) 2024-08-23

Family

ID=92381828

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410520987.3A Pending CN118538053A (en) 2024-04-28 2024-04-28 Method for determining critical warning distance and critical braking distance, and vehicle cooperative safety interaction system integrating 4D millimeter wave radar and monocular vision camera

Country Status (1)

Country Link
CN (1) CN118538053A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN119942473A (en) * 2025-01-16 2025-05-06 扬州宝科信息技术服务有限公司 Safety warning signal transmission control method applied to intelligent monitoring equipment
CN120756442A (en) * 2025-09-11 2025-10-10 安车智行(北京)科技有限公司 A passive safety distance control system based on monocular perception

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN119942473A (en) * 2025-01-16 2025-05-06 扬州宝科信息技术服务有限公司 Safety warning signal transmission control method applied to intelligent monitoring equipment
CN120756442A (en) * 2025-09-11 2025-10-10 安车智行(北京)科技有限公司 A passive safety distance control system based on monocular perception

Similar Documents

Publication Publication Date Title
CN113276769B (en) Vehicle blind area anti-collision early warning system and method
CN111223302B (en) External coordinate real-time three-dimensional road condition auxiliary device for mobile carrier and system
CN113147733B (en) Intelligent speed limiting system and method for automobile in rain, fog and sand dust weather
CN111469838A (en) Collaborative ACC/AEB decision management system based on Internet of vehicles and vehicle
CN114559933B (en) Unmanned vehicle emergency risk avoidance sensing and decision-making system based on vehicle-road cooperation
US20180173236A1 (en) Vehicle control system, vehicle control method, and vehicle control program
JP7119365B2 (en) Driving behavior data generator, driving behavior database
CN118538053A (en) Method for determining critical warning distance and critical braking distance, and vehicle cooperative safety interaction system integrating 4D millimeter wave radar and monocular vision camera
CN107204128B (en) Automobile anti-collision early warning system and method based on ZigBee communication
CN111402626B (en) A safe vehicle following distance control system and control method based on vehicle-road coordination
CN108639059A (en) Driver based on least action principle manipulates behavior quantization method and device
KR102505867B1 (en) Multi-vehicle data analysis and traffic signal control method and system using LiDar sensor
CN110097783A (en) Vehicle early warning method and system
US11753013B2 (en) Method for operating a driving assistance system, and driving assistance system
US20240046787A1 (en) Method And System For Traffic Clearance At Signalized Intersections Based On Lidar And Trajectory Prediction
JP4093026B2 (en) Road environment information notification device, in-vehicle notification device, information center device, and road environment information notification program
CN112489383A (en) Early warning system and method for preventing red light running accident based on machine vision
CN116798222A (en) Highway visibility detection and intelligent induction methods, systems and devices
CN112927514B (en) Prediction method and system for motor vehicle yellow light running behavior based on 3D lidar
CN115871717A (en) Method and device for emergency braking of automatic driving vehicle
JP3639196B2 (en) Vehicle identification device
CN114942450A (en) Method for detecting vehicle weighing behavior by laser radar
CN118220249A (en) Rail vehicle anti-collision control system, method, device and medium
CN118351496A (en) Traffic abnormal event detection method and system based on radar fusion
CN116165655A (en) Global vehicle track tracking system based on millimeter wave radar group

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination