CN106846863B

CN106846863B - Accident black spot warning system and method based on augmented reality and cloud intelligent decision

Info

Publication number: CN106846863B
Application number: CN201710022303.7A
Authority: CN
Inventors: 叶昊
Original assignee: Individual
Current assignee: Individual
Priority date: 2017-01-12
Filing date: 2017-01-12
Publication date: 2020-05-05
Anticipated expiration: 2037-01-12
Also published as: CN106846863A

Abstract

The invention discloses a vehicle networking accident black spot warning system and method based on augmented reality and cloud intelligent decision, wherein a vehicle-mounted client system layer realizes the collection of driving data and sends the collected driving data to a cloud control center layer through a cellular communication network layer; receiving a data decision packet sent back from the cloud control center layer, and displaying the decision packet to a driver in an augmented reality mode in combination with a live-action road condition; the cloud control center layer makes decision analysis according to the vehicle driving data and the data in the black point database received in real time and aiming at the vehicle driving state and the current traffic and weather conditions, and sends corresponding warning information to the vehicle-mounted client system layer in a data packet format. According to the invention, various dynamic traffic data sources are fused, and cloud intelligent decision and augmented reality are combined, so that the awareness of a driver to various accident risks at a road black point can be greatly improved, and the safety awareness of the driver at an invisible road black point is enhanced, thereby reducing the accident rate and improving the road black point safety.

Description

Accident black spot warning system and method based on augmented reality and cloud intelligent decision

Technical Field

The invention belongs to the technical field of Internet of vehicles, and particularly relates to an Internet of vehicles accident black spot warning system based on augmented reality display and cloud intelligent decision and a related implementation method thereof.

Background

With the worldwide progress of urbanization accelerating, road traffic safety has become a global problem. According to the research and investigation on traffic accidents in the past, the traffic accidents are not uniformly distributed on a road network, and most of the traffic accidents are concentrated on some frequent road sections on the road network, so that the road sections are generally also called road traffic black spots. The development of the safety technology of the road black spots has important significance and effect on reasonably allocating traffic resources and reducing the rate of traffic accidents.

The formation of the road traffic black spots is complex, and generally caused by the combination of a plurality of accident causes, which can be summarized into four main factors which are easy to cause traffic accidents, mainly including four inducers, namely, automobile factors, road factors, driver factors and environmental factors. Such as automotive factors (e.g., vehicle out of control, scuffing, rear-end collision, side-tipping), road factors (e.g., steep road surface, wet road, excessive grade, excessive bend angle), driver factors (e.g., driver distraction, fatigue driving, drunk driving), environmental factors (e.g., traffic congestion, heavy rain, heavy snow, heavy fog), etc. It is further noted that accidents often occur due to a centralized effect of multiple factors rather than a single factor, and that the factors that cause the accidents are not consistent from place to place. In addition, these factors result in similar traffic accidents that are likely to occur in vehicles or drivers under similar circumstances, with a higher probability of accident repeatability. Therefore, in the dark spot of the road, the driver is reminded of relevant accident risks by analyzing the existing accident history experience and combining the actual situation of the driver, so that the prospect of reducing the dark spot accident rate of the road is great.

The frequent accident road sections defined as the road black spots are various, and typical road black spots comprise urban level crossings, overpasses, T-shaped junctions, curves, wild animal flow regions and the like. The range of the black spot is also different in size under different traffic and environmental conditions, some sizes are no more than a few meters, and some sizes include risk areas within dozens of meters. At present, the mainstream road black spot warning mode mainly depends on arranging a static warning guideboard and reminding a driver of potential road safety risks through warning information. With the advance of science and technology, embedding predefined road safety information in a vehicle navigation system (or a smart phone) becomes another possible warning mode. However, as previously mentioned, road black spot cues are generally combined by a variety of accident causes, which are not only static (e.g., road linetype), but are mostly dynamic (weather, traffic, driver status). These existing methods have little systematic consideration of the role of various dynamic elements of people, vehicles, roads and the environment, and thus have a rather limited role in traffic warning. For example, drivers often fail to notice the relevant guideboard, the driving situation of each driver is very different, the guideboard placement point cannot be adjusted according to the situation, and the like. For embedded systems, the content is only static data, and is not combined with actual data and large environmental traffic data, and the warning information is unreliable and cannot meet the actual safety requirements.

A number of examples of existing car networking technology-based security systems were collected and studied as follows:

the patent [ vehicle-mounted system of smart phone based on car networking, 201210485142.2 ] uses smart phone and on-vehicle diagnostic system OBD to combine car CAN bus collection vehicle-mounted sensor data to send to the car networking control center of a remote end through cell-phone communication module and store. And the vehicle networking center provides data storage and displays the vehicle-mounted state information to the user through related post-processing service. Most of the current commercial vehicle networking diagnosis systems are similar to the invention and are mainly used for detecting and displaying corresponding sensors of automobiles, such as engines, doors and the like.

A calibration system and method for lane departure warning based on internet of vehicles, 201410521225.1, provides a vehicle safety driving assistance system using the internet of vehicles. The invention uses the vehicle-mounted controller unit and the pair of cameras to carry out video identification on the road marking, and provides lane departure early warning by identifying and extracting the marking and then comparing the marking with the real-time position of the vehicle. Similar system functions can only provide short-distance traffic safety applications based on the periphery of the automobile (generally in the range of 1 to 2 meters), such as lane change warning and blind spot warning, which are basic application types of many current vehicle-mounted systems, but the system cannot utilize global information based on a specific environment range, such as traffic road condition and weather information, and then convert the wide-range information into the road safety application.

The patent [ a driving safety early warning method and system based on car networking ], 201510843968.5 ] uses the car networking system to obtain the position of specific vehicle and its peripheral vehicle, realizes real-time monitoring and contrast to each vehicle position through the cloud center to predict the car collision risk according to specific algorithm thereby provides real-time safety warning for the driver. The invention fully considers the geometric spatial position relation between automobiles, but the invention can not consider the influence of various real-time environmental factors, such as traffic road conditions, weather conditions and the like, on the safety and driving behaviors of the automobiles. Thus, such systems can only model the spatial location between cars in ideal conditions and then provide warnings. The warning of the system has great uncertainty, the warning is inaccurate under the condition close to reality, and the advanced and preventive warning function based on the road condition environment cannot be provided.

The invention discloses a driving analysis system based on an internet of vehicles (201410767104.5). The driving data is acquired at a vehicle-mounted terminal and is sent to a remote server. The remote server analyzes the safe driving state of the vehicle according to the risk preset weight of various data in the driving data, sends the safe driving analysis result to the mobile terminal and displays the safe driving analysis result of the vehicle through the mobile terminal. Similar to the similar system, the invention does not consider the influence of various real-time traffic road conditions and weather conditions on driving behaviors and the influence of risk weights of other parameters, and the analysis result of safe driving is unreliable under the action of various traffic environments.

The invention discloses a bidirectional active speed-limiting and overspeed early warning system based on an internet of vehicles, 201310364536.7, and relates to an active speed-limiting and overspeed early warning system based on the internet of vehicles, wherein a driver can actively limit the speed within a specified range by using a remote controller according to own needs and actual conditions, or can actively limit the speed or preset the speed within the specified range of a special line section according to supervision needs. The system only considers speed as a main safety auxiliary feature, and in a specific road section, other safety behaviors besides the speed, such as improving the warning of pedestrian passing and attention to sudden events of traffic environment, and the road section and historical accident black spots needing to consider specific traffic conditions, the system is difficult to effectively play a role in the places due to the action of various factors.

In summary, most of the existing automobile safety systems on the market have single function and lack intelligence in the aspect of processing road risk warnings, most of the systems can only develop safety warnings with relatively short visual fields, and cannot effectively integrate various rich historical data and real-time data, so that various accident causes are comprehensively considered to sense a specific road section and provide road black point safety warnings. In addition, most of basic safety systems do not consider the combination of augmented reality and real-scene road conditions to provide visual safety information, so that a vehicle-mounted black spot safety warning system which can comprehensively consider various accident factors, historical data and real-time data and can provide warning with intuition and better user friendliness according to the driving state of a driver needs to be designed and developed.

Disclosure of Invention

1. Technical problem to be solved by the invention

In order to solve the problem of lack of effectiveness of driving safety warning of corresponding safety systems for accident black spots, the invention provides an Internet of vehicles accident black spot warning system and method based on augmented reality and cloud intelligent decision. The invention collects various driving safety data (related to four inducements of automobile factors, road factors, driver factors and environmental factors), transmits the driving safety data to the cloud end through the wireless communication network, and carries out comprehensive decision in the cloud end control center, the comprehensive decision process not only considers historical accident experience, but also considers safety information unavailable in the prior system, such as human factors, road factors, weather factors and the like, and transmits the comprehensive warning information back to the driver, thereby providing reliable accident black spot warning for the driver, and combining the warning information with road scenes to enhance reality and display to the driver. The system can greatly improve the perception of a driver to various accident risks at the road black points, and enhance the safety awareness of the driver at the invisible road black points, thereby reducing the accident rate and improving the safety of the road black points.

2. Technical scheme

In order to achieve the purpose, the technical scheme provided by the invention is as follows:

the invention provides an intelligent driving black spot safety early warning system and method based on internet of vehicles and augmented reality for solving the driving safety problem at an accident black road section.

The invention relates to an Internet of vehicles accident black spot warning system based on augmented reality and cloud intelligent decision, which comprises a vehicle-mounted client system layer, a cellular communication network layer and a cloud control center layer, wherein:

the vehicle-mounted client system layer realizes the collection of driving data and sends the collected driving data to the cloud control center layer through the cellular communication network layer; meanwhile, a data decision packet sent back from the cloud control center layer is received and visually displayed to a driver in an augmented reality mode in combination with real road conditions;

the cellular communication network layer realizes the two-way communication between the vehicle-mounted client system layer and the cloud control center layer;

the cloud control center layer makes decision analysis according to the driving data received in real time and the data in the black spot database aiming at the driving state and the current situation of the vehicle, and issues corresponding warning information to the vehicle-mounted client system layer in a data packet format; the data in the black spot database comprises traffic road condition data acquired from a traffic management center, historical accident black spot data and weather data acquired from a meteorological station.

Furthermore, the vehicle-mounted client system layer comprises at least one vehicle-mounted client system, the host of the vehicle-mounted client system comprises a vehicle-mounted front-mounted system and a vehicle-mounted rear-mounted system, and the vehicle-mounted front-mounted system comprises a temperature and humidity sensor, an illumination pressure sensor, a tire pressure friction sensor, a windshield wiper sensor, an automobile steering lamp indicating system, an automobile suspension system and an automobile braking system; the vehicle-mounted after-loading system comprises a satellite positioning navigation module, a cellular communication module, a digital map module, a driver monitoring camera module and an augmented reality display module; and the vehicle-mounted client system integrates various data through a vehicle-mounted system display module.

Furthermore, the vehicle-mounted system display module is a vehicle-mounted traveling computer, an external smart phone or a head-up display instrument.

Furthermore, the cellular communication network layer adopts a TCP/IP communication protocol based on cellular communication, adopts a socket as a communication mediator, and provides communication and data transmission for communication between the cloud control center and the application software through the cellular base station.

Furthermore, the cloud control center layer comprises 13 functional modules, and the functional modules are connected in series through data stream logic; wherein: the data communication module receives a communication data packet transmitted by the vehicle-mounted client system, the data decoding module and the video data decoding module respectively decode the received communication data packet, and relevant decoding information is sent to the data analysis module; the accident data mining module is used for mining historical traffic accident data online or offline and storing corresponding accident black point positions, accident causes and results thereof in an accident black point database; the data integration module is responsible for receiving the decoded data, integrating the data into a specified format and sending the format to the intelligent decision module for operation; the external database is used for storing data streams acquired from the outside, including traffic data acquired from the real-time traffic database and weather data acquired from the weather database, the data streams are finally transmitted to the intelligent decision module, and the results processed by the intelligent decision module are output after being coded by the data coding module; in addition, the traffic management personnel can manage the vehicle position, the accident black spot and the traffic information through the geographic information system of the visual platform.

The invention discloses an Internet of vehicles accident black spot warning method based on augmented reality and cloud intelligent decision, which comprises the following steps:

the method comprises the following steps that firstly, an accident black spot warning system is started, a vehicle-mounted client system automatically acquires connection with ports of an internal system and an external system, if the connection is normal, the vehicle-mounted client system collects required data according to a black spot data information collection rule, and the collected data are stored in the vehicle-mounted client system for caching for later use;

step two, in the vehicle driving process, a vehicle-mounted client system opens a communication link and is connected with a cloud control center in real time, basic safety data are sent to the cloud control center at a vehicle-mounted client of a non-accident black spot road section to keep the monitoring of the cloud control center on the basic state of the vehicle, and the position of the vehicle is matched with the position judgment relation of accident black spots in a corresponding area in real time;

step three, when detecting that the automobile enters an accident black spot area, the cloud control center starts an accident black spot mode, at the moment, the vehicle-mounted client system collects the expanded safety data, encodes the collected data and sends the encoded data to the cloud control center, and the data are decoded in the cloud control center and then sent to each functional module for analysis;

the cloud control center analyzes and calculates the acquired data, if the driving data is calculated to be matched with the accident risk obtained by deduction due to accident black points, black point warning data is generated according to the coding rule, a safety warning mechanism is started, and the black point warning data is pushed to a vehicle-mounted client system;

and step five, the vehicle-mounted client system receives the black spot warning data, decodes the black spot warning data and transmits the decoded black spot warning data to the head-up display instrument through Bluetooth, the head-up display instrument displays the black spot information on the head-up display instrument in a form of enhancing reality visualization in a form of combining with a road network by establishing a matched road network model, and reminds a driver by assisting with a sound light and vibration mode.

Furthermore, in the above warning process, the specific process of determining the driving-in and driving-out accident black point is as follows:

step one, collecting current GPS position data Location (x, y) based on an automobile, and converting the current GPS position data Location (x, y) into Local coordinates Location (x, y) under the Local coordinates of a current digital map according to a coordinate format, which is called LLposition (x, y) for short;

step two, sequentially circulating the geometric center coordinates of each accident black point in the accident black point database, searching the accident black point coordinate set closest to the current local coordinate LLocation (x, y), and extracting the vertex coordinate set C (C) of the black point polygon₁，C₂，C₃...C_N)；

Step three, calculating coordinate cross lines PL of the LLocation (x, y), crossing and intersecting the polygon through developing the cross lines PL, and calculating the number of cross points of the polygon and each sideline position; if the cross line intersects with the side line formed by the polygon, the cross point counter automatically increases the corresponding number of cross points; otherwise, the count of the cross point counter is not increased;

step four, making an entrance decision judgment by a cross point counter according to the number of cross points, and if the number of the cross points is an odd number, judging that the current position of the automobile is in an accident black point; if the number of the transverse intersection points is an even number, the automobile is outside the accident black point; when the judgment of the entering and exiting is changed, the moment when the automobile enters the accident black point is the moment, the changing moment is recorded as the time when the automobile enters the accident black point, and the moment is also the judgment condition for starting the warning or starting the data acquisition algorithm.

Furthermore, in the warning process, the specific process of historical accident black point mining is as follows:

acquiring historical accident original data of a road in a specific traffic area at a traffic management department, sorting the accident data, cleaning incomplete data and deleting redundant data;

mining traffic accident black points by adopting various data statistics and a space-time statistics technology based on a geographic information system, and marking possible main accident causes of each accident black point;

thirdly, geographic information editing is carried out on historical data of traffic accident black points according to geographic positions of the traffic accident black points, a corresponding accident black point relation database is designed, and a spatial geographic information accident black point vector layer is established and stored in a geographic information system database as attributes related to black point data;

and step four, formatting the data of the traffic accident black points according to a navigation data format, combining the accident black point data with a navigation system according to the navigation data format, importing the accident black point data into a vehicle-mounted client system, integrating the accident black point data with a navigation map and the navigation system, developing into a vehicle-mounted black point-based database, and displaying the black points on the vehicle-mounted client system and a cloud control center.

Furthermore, in the warning process, the specific process of reasoning and decision making by the cloud control center is as follows:

the method comprises the steps that firstly, under an on-line condition, a cloud control center determines main accident black points in a specific geographic area range through a data mining technology, wherein the main accident black points comprise geographic positions of the accident black points and main accident causes of the accident black points, the main accident causes are classified and priority levels are determined, a rule matching and resolving template based on intelligent decision is customized on the basis, and black point risk matching rules are determined;

step two, in a specific implementation state, the cloud control center acquires driving data of the vehicle-mounted client system in a specific geographic area range, stores required information in a classified manner, and meanwhile, acquires traffic data, weather station and weather data of a traffic management department in real time and stores the data in a cloud database;

after the collected data are decoded, firstly decoding positioning information in an intelligent decision module, executing judgment of entrance and exit of accident black points, if a vehicle enters a specific black point, acquiring various accident cause data of the black point, and extracting real-time state data related to causes for intelligent decision;

step four, executing a digestion algorithm based on the rule, firstly judging a single accident cause, then sequentially judging a plurality of causes, obtaining an accident black point risk index according to a designed weight formula, and determining whether a driver gives a specific driving warning; the weight of the weight formula is given after comprehensive judgment is made according to the historical records and expert opinions of the specific road section;

step five, when the cloud control center determines to give driving warning to the driver, warning information needs to be matched with a specific risk reason and is sent to the vehicle-mounted client;

and step six, the vehicle-mounted client receives the warning coding information and decodes the information to the vehicle-mounted navigation system and the augmented reality module, and the augmented reality module visually displays the information to the driver in a display mode of combining augmented reality and a real road section.

Furthermore, in the warning process, the road sections are split into accident black point road sections and non-accident black point road sections, and different information acquisition is performed on different road sections, so that the acquisition efficiency is improved, and the bandwidth consumption is reduced; basic safety data are collected on a non-accident black spot road section, and the basic safety data collect driving data according to a bandwidth minimum-saving principle; acquiring expanded safety data on the accident black point road section, wherein the expanded safety data are high-density driving data with abundant information; and sending black spot warning data to the vehicle entering the black spot of the accident and deducing the accident risk, wherein the black spot warning data provides potential risk causes of the black spot, corresponding safe operation directions and the like.

3. Advantageous effects

Compared with the prior art, the technical scheme provided by the invention has the following remarkable effects:

(1) the invention relates to an Internet of vehicles accident black spot warning system based on augmented reality and cloud intelligent decision, which considers multiple accident causes and comprehensively judges the accident black spot risk of multiple associated data sources, the technology simultaneously considers real-time driving state data and historical accident recording data, fully considers human, vehicle, road and environment data as the accident causes, uses a rule-based cloud algorithm decision to dynamically judge the risk and possibility of accident occurrence, has higher reliability compared with a common warning mode based on a single data source and a static state, and can start automatic alarm after the risk is deduced;

(2) according to the vehicle networking accident black spot warning system based on augmented reality and cloud intelligent decision, a vehicle-mounted data acquisition system is implemented at a client side to acquire data of a front-mounted system, and meanwhile, data acquisition of the front-mounted system is also realized, such as positioning data and video identification data of a GPS (global positioning system); at the cloud end, the system acquires and realizes data access to corresponding weather and traffic data through an Application Program Interface (API); the system has the advantages that the acquired related data include people, vehicles, roads and environments, the data type is comprehensive, the content is rich, the related data can be stored in the vehicle local machine and can also be transmitted to the cloud data center, the cloud data center can acquire and classify various acquired data, and the data are used for developing various intelligent vehicle-mounted safety applications through an open data interface;

(3) according to the vehicle networking accident black spot warning system based on augmented reality and cloud intelligent decision, the acquisition of the spatial position of a road network black spot and the acquisition of an accident cause are realized by mining background historical accident data, and the historical black spot data and the spatial road network data are integrated; the vehicle-mounted client system and the cloud data center can realize the judgment of the position of the accident black spot and the acquisition of the black spot cause information by integrating the accident black spot map;

(4) according to the vehicle networking accident black spot warning system based on augmented reality and cloud intelligent decision, an available head-up display or other corresponding augmented reality displays are used, a road scene is combined with safety information, road accident black spots are visually presented by using a display technology based on augmented reality, real-time risk avoidance guidance is provided, and the system has positive effects on improving driving safety awareness of a driver and improving road safety;

(5) the vehicle networking accident blackspot warning system based on augmented reality and cloud intelligent decision is simple in equipment requirement, convenient to install, high in expandability, complete in system function, capable of continuously updating and iterating on the basis, capable of developing novel application and capable of meeting requirements for driving safety improvement and road traffic safety supervision.

Drawings

FIG. 1 is a diagram of a vehicle networking system layer architecture of the present invention;

FIG. 2 is a block diagram of a host architecture of the in-vehicle system of the present invention;

FIG. 3 is a block diagram of a cloud control center layer of the Internet of vehicles system of the present invention;

FIG. 4 is a diagram of a TCP/IP communication module of the Internet of vehicles system according to the present invention;

FIG. 5 is a flow chart of the accident blackspot database generation of the present invention;

fig. 6 is a black point information encoding rule and a sampling chart according to the present invention.

Detailed Description

For a further understanding of the invention, reference should be made to the following detailed description taken in conjunction with the accompanying drawings and examples.

Example 1

With reference to fig. 1, the vehicle networking accident black spot warning system of the present embodiment is based on a vehicle networking system architecture, and the whole system is completed by three system function layers, including a vehicle-mounted client system layer 101, a cellular communication network layer 102, and a cloud control center layer 103.

Each system functional layer realizes the expected function according to the designed car networking structure, and the specific function implementation is described as follows:

the vehicle-mounted client system layer 101 mainly realizes two functions: firstly, the vehicle-mounted client system is used for acquiring driving data, and the acquired driving data is sent to the cloud control center layer 103 through the cellular communication network layer 102. Here, the data collected by the vehicle-mounted system includes four types of data sources related to accident causes, including: driver data, driving state data, local traffic data and environmental data. Since each type of data comes from different data sources, the data can be classified and stored to the vehicle-mounted client based on the data format and the category (for example, a text format or a video format); secondly, as mentioned above, the corresponding driving data is collected to the cloud data center layer 103 and then processed by the cloud decision center, and black spot warning information is inferred and fed back to the vehicle-mounted client system. Therefore, the second function of the vehicle-mounted client is to receive the data decision packet sent back from the cloud control center layer 103, and visually display the data decision packet to the driver after combining with the real road conditions in an augmented reality manner, thereby playing roles of warning and safety guidance.

The main role of the cellular communication network layer 102 is to serve as an intermediate communication layer (function similar to an adhesive) for realizing bidirectional communication between the in-vehicle client system layer 101 and the cloud control center layer 103. The modes of the two-way communication include data transmission from the vehicle-mounted client to the cloud control center (forward communication) and data broadcasting from the cloud control center to the vehicle-mounted client (reverse communication). Here, the cellular communication is selected to be advantageous because it can realize reliable two-way communication over long distances, and at the same time, because the infrastructure for cellular communication is already established in most cities, it can maximize the reduction in cost and increase the communication coverage.

Cloud control center layer 103 is the core functional layer of this embodiment, and it mainly realizes three functions: firstly, decoding the collected and received vehicle-mounted end driving data, reprocessing the data according to the designed system logic, storing the data into a cloud database, and executing information decision of black spot warning through a decision module; secondly, connecting the traffic management center and a corresponding local meteorological station through an application program interface API (application program interface), acquiring traffic conditions (such as road network traffic flow conditions) and weather conditions (such as sudden rainstorm weather) of a specific area, mining historical accident black points through a data mining technology, integrating the accident black points into a spatial map database, and establishing a black point decision database which integrates the traffic conditions, the weather database and an accident hot point database; thirdly, the cloud decision center applies real-time driving data and various data in the black point database according to a developed intelligent decision algorithm, makes decision analysis according to the driving state and the current situation of a specific vehicle, issues corresponding warning information to the vehicle-mounted client system layer 101 in a data packet format, and displays the warning information through the augmented reality display module of the vehicle-mounted client.

The detailed design of each system functional layer in this embodiment is described as follows:

the vehicle-mounted client system host mainly comprises a vehicle-mounted front loading system and a vehicle-mounted rear loading system, as shown in fig. 2:

the vehicle-mounted front-mounted system comprises various sensors already installed in a vehicle system, such as a temperature and humidity sensor 201, a light pressure sensor 202, a tire pressure friction sensor 203, a wiper sensor 204, a vehicle turn light indicating system 205, a vehicle suspension system 206 and a vehicle brake system 207. The vehicle-mounted sensors listed above can provide driving parameters of the driving process of the automobile, and running environment parameters and driving environment parameters of the automobile. For example, the temperature and humidity sensor can be used for indicating the ambient temperature and humidity of the driving environment of the automobile, and the temperature and humidity affect the driving fatigue of the driver so as to be used for evaluating the safety risk; the illumination sensor is used for evaluating the illumination condition around the automobile, so that the illumination sensor is used for evaluating the visual range and the performance of the automobile; the tire pressure friction sensor is used for judging the tire condition and the road surface friction force and evaluating the road surface characteristics; the wiper sensor can collect local rainfall weather and rainfall degree through the opening and closing of the wiper; the steering lamp system is used for acquiring a steering intention in a driving process; the suspension system is used for indicating the flatness of the road surface; the braking system is used for deducing driving behaviors, acceleration, traffic conditions and the like.

The vehicle-mounted after-loading system comprises a driving auxiliary system, such as a GPS navigation system, a vehicle data recorder and the like, which are installed after each automobile manufacturer or user. These systems can be used to provide traffic information contained in addition to front-mounted system sensors. The afterloading system modules may include a satellite positioning navigation module 209, a cellular communication module 210, a digital map module 211, a vehicle event recorder-based driver monitoring camera module 212, a heads-up display-based augmented reality display module 213, and other modules may be flexibly added to the afterloading system to improve the afterloading system functions. Compared with the vehicle-mounted front-mounted system, the vehicle-mounted rear-mounted system can provide more accurate traffic safety auxiliary information in many aspects, for example, the satellite positioning navigation module 209 can provide precise time, high-precision positioning (longitude and latitude), speed, elevation, real-time steering, azimuth angle and the like; the cellular communication module 210 is used for providing bidirectional long-range communication between the vehicle-mounted client and the cloud control center, the cellular communication module 210 establishes bidirectional verifiable communication through a reliable data communication protocol TCP so as to guarantee the communication reliability, and more stable communication quality can be improved by combining a third generation 3G communication technology or a fourth generation 4G cellular communication technology on a hardware layer. The digital map module 211 can provide an integrated display digital navigation map, the digital map module 211 stores map information and map attribute data through the digital map, and provides arrangement, line type, curve, black point data and the like of a road network after integrating with black point data, and the digital map module 211 can be used for vehicle-mounted local decision-making and cloud decision-making; driver monitoring camera module 212 realizes the video road conditions record outside the car and the face and the driving operation control of driver in the car through front and back camera, and accessible technologies such as video identification discern driver's action and driving state, send the data information who obtains back to high in the clouds control center. The augmented reality display module 213 can display the safety warning information sent back from the cloud control center to the driver through a friendly visual interface, and provide visual and concise safety warning by combining the augmented reality scene with the real-time road condition.

The information integration of the vehicle-mounted client system is mainly completed by the vehicle-mounted system display module 208. If the vehicle is equipped with a vehicle computer, the vehicle-mounted system display module 208 can perform rapid information integration and calculation through the vehicle computer. For a vehicle not loaded with a traveling computer, the vehicle can be operated by a single external smart phone or a head-up display, data is processed and integrated by a built-in processor, and relevant information is displayed, but attention needs to be paid that the operation speed of external equipment such as a mobile phone is lower than that of a vehicle-mounted computer, and the data operation efficiency and the timeliness of system information are possibly reduced.

The cloud control center layer 103 is a core for processing all collected data and processing data. All the functional modules of the cloud control center 103 are arranged in the cloud, and the large data storage capacity of the cloud and the high-speed computing capacity of the cloud can be fully utilized. As shown in fig. 3, the main functions of the cloud control center are implemented by 13 modules in total, including: the system comprises a data communication module 301, a data decoding module 302, an accident data mining module 303, an accident black spot database 304, a data analysis module 305, a video data decoding module 306, a data integration module 307, an intelligent decision module 308, a data coding module 309, an external database 310, a geographic information system 311, a real-time traffic database 312 and a weather database 313.

All the functional modules are connected in series through data stream logic to form a cloud control center. The function of each functional module is organized as follows: the data communication module 301 is responsible for receiving communication data packets transmitted by the in-vehicle client system. After the data packet reception is completed, the data decoding module 302 and the video data decoding module 306 respectively decode the vehicle-mounted communication data packet (general text data) and the video data packet (video data) according to the communication protocol and the encoding specification. The data decoding module 302 is responsible for decoding vehicle data packets, including two types of data, the first type of data includes vehicle action parameters, such as vehicle speed, location, etc.; the second category includes vehicle local environmental status data such as weather conditions (e.g., windshield wiper determination), road surface anomalies (e.g., friction sensor determination). After the data decoding is completed, the data decoding module 302 and the video data decoding module 306 send the relevant decoding information to the data analysis module 305. Because the data formats are greatly different, the corresponding video information can be transmitted to the video data decoding module 306 and decoded to generate driver monitoring data, and the fatigue judgment is realized by algorithms such as capturing the facial features of the driver and the like. The accident data mining module 303 mainly provides a data mining technology for mining historical traffic accident data online or offline, and stores the corresponding accident black point position and accident cause and the result thereof in the accident black point database 304. The accident black point database 304 is a spatial database that can store spatial vector data. The incident black point database 304 may store geospatial vectors of the incident and may also store corresponding incident black point data attributes. The data integration module 307 is responsible for receiving the decoded vehicle condition information and the decoded driver state information, and then integrating the data into a specified format according to an algorithm rule and sending the format to the intelligent decision module 308 for operation. The external database 310 is mainly used for storing data streams obtained by being linked with the outside, for example, the data streams may be weather data accessed to the weather database 313 through the internet, or real-time traffic data (for example, obtaining traffic flow and congestion distribution, etc.) linked to the traffic management center traffic database 312. After data is accessed, the data is transmitted to an external data decoding module, and then the data is decoded and transmitted to a core module, namely an intelligent decision module 308. The intelligent decision module 308 is primarily responsible for processing two-dimensional data information and data sources, including decoded data from the vehicle data decoding module and data information from the external traffic database, and at the same time, is capable of processing data from multiple data sources, including historical traffic accident black spot data, real-time weather, road condition data, and the like. The result processed by the intelligent decision module 308 is encoded by the data encoding module 309 and then output. In addition, in the cloud control center, traffic management personnel can visually manage the vehicle position, the accident black spot and other corresponding traffic information through the geographic information system 311 of the visual platform.

The vehicle networking accident black spot warning system of this embodiment must satisfy the security demand of car, therefore the communication of cloud control center and vehicle client system must be based on reliable safety protocol, guarantees data communication's reliability and the security of the whole framework system of vehicle networking. As shown in fig. 4, in the present embodiment, a TCP/IP communication protocol based on cellular communication is adopted, a Socket (Socket)402 is adopted as a communication mediator, and a communication and data transmission function is provided for communication between the cloud control center 401 and the application software 405 through the cellular base station 404.

The specific communication system functions and communication flow are designed as follows:

the method comprises the following steps that firstly, a cloud control center 401 keeps a monitoring state through a communication module, after a vehicle equipped with a vehicle-mounted client enters a coverage range of a connectable cellular base station 404, an application end of a vehicle-mounted client system starts a socket 402 to request data docking with a cellular communication port, and the cloud control center 401 distributes a corresponding communication data port to a specific vehicle-mounted client system after monitoring a data docking request;

secondly, the vehicle-mounted client system performs data coding on the acquired data according to a specified format, packages the data and sends the data to a data communication module, is connected to a cellular base station 404 through a socket 402, establishes a duplex data communication mode through a port distributed by a TCP/IP communication network 403 and a data transmission port, and establishes and opens a duplex data communication mode;

step three, the vehicle-mounted client system uniformly encodes the data collected by the vehicle-mounted front loading system and the vehicle-mounted rear loading system, and transmits the data to the cloud control center 401 by means of a communication protocol and a communication system;

step four, the cellular base station 404 sends the data to a socket 402 of the cloud control center 401, the socket 402 sends the data to a data decoding module of the cloud application software, the data decoding module decodes the data and checks the data type, the data integrity and the like, the data sending success verification information is synchronously provided for the application software 405, and meanwhile, the data decoding module in the system sends the data to each module for further cloud decision analysis;

after the decision making by the cloud control center 401 is completed, encoding decision information and sending data to the cellular base station 404 through the socket 402, and sending a data packet to an application socket and an application through the socket by the cellular base station 404 and displaying the data packet to a driver;

the core algorithm of the embodiment is a series of cloud-based function and algorithm modules, and is formed by combining a functional module to realize data sorting, knowledge-based reasoning and a decision module. The decision and implementation of the road black spot comprises three main stages of preparation, decision and implementation. In the preparation stage, the early-stage data mining and knowledge and experience arrangement, analysis and induction are carried out to form a series of judgment conditions. And the decision stage is implemented by an inference decision machine, and different accident causes are integrated to finally form a hot spot inference decision mechanism depending on a specific road section and a specific driver behavior. And in the implementation stage, real-time derivation is carried out, corresponding hotspot warning information is generated finally, and safety warning is provided for the driver in an augmented reality display mode after the information, the actual traffic condition and the road section are combined.

The algorithm of the embodiment can be divided into four steps of accident black point mining, black point decision reasoning, black point information coding and black point data sampling. The detailed algorithm is described as follows:

an accident black point data mining process:

the accident black spot module performs data sorting and cleaning in the early stage by collecting historical traffic accident data, including cleaning incomplete data and deleting redundant data. And importing accident data into a geographic information system, and performing spatial and temporal statistical analysis and identification on the accident black points of the road network by using black point statistics (spatial data statistics technology) and black point identification technology. The black spot technology is to mine traffic accident black spots according to traffic historical data of a specific area and various data statistical modes (time, space and space time), and encode and store corresponding information in a space road network database.

The black point database attribute should include three categories of information, namely time information, geographical location information and attribute information of the black point. The time information represents a black spot frequent time period, such as a high-incidence day period and a week period, a seasonal time period, and the like; the geographic information comprises accident black spot frequent region (a series of longitude and latitude collections), such as slope road sections and cross road sections; the attribute information represents attributes corresponding to the accident black points, such as the number of accidents, main causes, black point risks and the like.

As shown in fig. 5, the main steps of the accident black spot mining are described as follows:

acquiring historical road accident original data in a specific traffic area at a traffic management department, determining a data format (corresponding to 501), sorting the accident data according to requirements, including cleaning incomplete data and deleting redundant data (corresponding to 502), and importing the accident data into statistical software and a geographic information system to perform data statistical analysis and spatial statistical analysis (corresponding to 503) respectively;

step two, mining the traffic accident black points by adopting various data statistics and a space-time statistics technology based on a geographic information system, and marking possible main accident causes (time, place, vehicle condition, driving condition, traffic condition, weather condition and the like) of each accident black point (corresponding to 504);

thirdly, geographic information editing is carried out on historical data of traffic accident black points according to geographic positions of the traffic accident black points, a corresponding accident black point data relation database (corresponding 505) is designed, and a spatial geographic information accident black point vector layer is established and stored in a geographic information system database (corresponding 506) as attributes related to black point data;

and step four, formatting the data of the traffic accident black points according to a navigation data format, combining the accident black point data with a navigation system according to the navigation data format, importing the accident black point data into a vehicle-mounted client system, integrating the accident black point data with a navigation map and the navigation system (corresponding to 507), developing the accident black point data into a vehicle-mounted black point database, and displaying the traffic accident black points on the client and the cloud (corresponding to 508 and 509).

Black spot entry and exit decision algorithm:

the position proximity relation is a precondition for triggering the corresponding warning of the black dot part. The black dot entry/exit decision algorithm of the present embodiment is based on a point-polygon relationship algorithm in computer graphics. In the operation aspect, the driving-in and the driving-out of the automobile at the black spot of the road are mainly determined by the geographical position Proximity relationship (Proximity) of the automobile and the black spot of the accident. The logic steps of the black spot entry and exit decision algorithm are described as follows:

the current position coordinates of the car are considered as the Location (x, y) of the recording point. According to the change of the automobile position, the automobile position coordinates form a set of position point changes, which is L₁，L₂，L₃...L_t. Thus, the current position of the vehicle is related to the set of routes traveled by the vehicle by L_i∈L_N. Considering any accident black point as a polygon according to the spatial geographic position, and defining the vertex of each polygon as a polygon set C (C)₁，C₂，C₃...C_N) For each vertex, a number ID is defined, corresponding to the geographical coordinates C (x, y), and the set of polygons forms the spatial extent of the accident black point. The whole algorithm is executed as follows:

step one, acquiring current GPS position data Location (x, y) based on an automobile, and converting the current GPS position data Location (x, y) into Local coordinates Location (x, y) under the Local coordinates of a current digital map according to a coordinate format of the current GPS position data Location (x, y), wherein the Local coordinates Location (x, y) is hereinafter referred to as Location (x, y);

Step three, calculating coordinate cross lines PL of the LLocation (x, y), crossing and intersecting the polygon through developing the cross lines PL, and calculating the number of cross points of the polygon and each sideline position; if the cross line intersects with the side line formed by the polygon, the cross point counter automatically increases the corresponding cross number; otherwise, the count of the cross point counter is not increased;

and step four, making an in-out decision judgment by the cross point counter according to the number of the cross points. Judging the relation between the automobile position and the accident black point according to the number of the cross points: if the number of the crossing points is odd, the current position of the automobile is judged to be in the accident black point; otherwise, if the number of the transverse intersection points is an even number, the automobile is outside the accident black point; if the judgment of the entering and exiting is changed, the time when the automobile enters the accident black point is the moment, and the changed time can be defined as the corresponding time when the automobile enters the accident black point and is also the judgment condition for starting the warning or starting the data acquisition algorithm.

An intelligent decision algorithm:

the intelligent decision algorithm of the embodiment generates the warning content of the accident black spot through a series of inference decision mechanisms in the cloud control center through the intelligent algorithm inference. The algorithm can be assembled in the early period according to different black spot characteristics, and finally a black spot warning reasoning decision mechanism based on a specific road section and a specific driver behavior is formed through continuous integration of different accident cause weights.

The decision algorithm is a deduction mechanism of 'IF-CONDITION (IF-CONDITION) THEN-ACTION (THEN-execute)', the judgment of accident black points and the corresponding accident black point cause come from the statistics and analysis of past historical traffic accident data, and a series of rule inference decision machines based on various accident causes are generated according to professional traffic knowledge and historical accident data mining. In the accident black spot warning decision process, different accident cause reasoning methods can be combined according to specific conditions through a reasoning decision machine, and finally a hot spot reasoning decision mechanism depending on specific road sections and specific driver behaviors is formed. For example, a black spot in a flat intersection accident may prioritize the vehicle speed variable based on its historical data analysis, while a black spot in a curve accident may prioritize vehicle speed and weather for decision making. Therefore, the algorithm mechanism is malleable, and can break up or recombine to form different reasoning conditions and form warning reasoning based on specific black points.

The decision process using the inference decision machine is as follows:

firstly, under an online condition, the cloud control center determines main accident black points in a specific geographic area range through a data mining technology, wherein the main accident black points comprise main geographic positions of the main accident black points and main accident causes of the accident black points; for example, the data mining technology classifies main accident reasons and determines priority levels by analyzing historical accident records, for example, the main accident reasons are caused by poor road linearity, extreme weather or traffic jam, and a rule matching resolution template (such as a road condition judgment set, a driving behavior judgment set and a weather judgment set) based on an intelligent decision is customized on the basis. The decision machine can judge under the condition that real-time data enters according to a set judgment template, for example, a main accident characteristic of a specific road section is obtained by analyzing according to historical accident records and is caused by overspeed behaviors of bad weather, and intelligent judgment is performed by matching speed and weather mainly when a decision rule is designed.

And step two, in a specific implementation state, the cloud control center acquires driving data of the vehicle-mounted client system in a specific geographic area range through the data communication module, and stores the required information in a classified manner. Meanwhile, the cloud external database is accessed to a traffic management department or a meteorological station through application program interface API (application program interface) data to acquire current traffic condition data and weather data of a local area and stores the data in the cloud database;

and step three, after data acquisition and decoding, sending the data to an intelligent decision module, firstly decoding positioning information, and executing a black spot access decision algorithm. The algorithm determines whether the automobile enters a specific black point in the black point database, and simultaneously, if the automobile enters the specific black point, attribute data acquisition (causes of various accidents) of the corresponding black point is started, and simultaneously, real-time state data related to the causes are extracted and sent to a decision machine for standby;

and step four, executing a rule-based digestion algorithm, judging a single cause, for example, if the real-time driving speed of a driver exceeds the accident risk speed of an accident black point, executing a specific speed risk warning to the driver, or if the accident black point causes the cause to be complex and the accident black point is easy to cause traffic risks under various possible conditions due to multiple causes (as mentioned above, the accident black point is judged by a data mining algorithm), judging multiple causes in sequence, obtaining an accident black point risk index according to a designed weight formula, and determining whether to give the driver a specific driving warning. The weight of the weight formula is given after comprehensive judgment and expert opinions are made according to the historical records of the specific road section;

and fifthly, paying attention to that when the cloud control center determines to give driving warning to a driver, warning information needs to be matched with a specific risk reason to the driver, so that the driver pays attention to the specific accident risk factor, for example, the accident causes that the speed exceeds the standard or the road surface is wet and slippery on the rainy day. The warning information needs to be sent to the vehicle-mounted client after passing through an agreed coding rule;

and step six, the vehicle-mounted client receives the warning coding information and decodes the information to the vehicle-mounted navigation system and the augmented reality module. Here, the display module visually displays to the driver in a display manner of combining the augmented reality and the real-scene section. Here, the augmented reality is used for visually displaying to the driver after being combined with the road scene, for example, the black spot part renders the road scene in red and displays the road scene to a head-up display instrument or projects the road scene to a windshield of an automobile in an augmented reality mode;

the coding rules of the black spot data acquisition information and the black spot warning information are as follows:

the driving data sampling technology is mainly used for reducing the use of communication bandwidth and improving the data collection efficiency in a black spot warning system. The driving data sampling mainly comprises a sampling rate and sampling content. By adjusting the data sampling rate, the vehicle-mounted client system can collect detailed safety information at important positions (such as the interior of a black spot) so as to improve the reliability of warning, and only collects necessary data information at non-important positions (such as a non-accident black spot road section), so that the communication bandwidth is reduced, and the efficiency of the black spot warning system is improved.

The data structure collected by the driving data sampling technology mainly comprises three blocks: basic security data 601, extended security data 602, and black dot warning data 603. The sampling technology divides road sections of a road network into accident black point road sections and non-accident black point road sections, and different information acquisition methods are executed on different road sections, so that the acquisition efficiency is improved, and the bandwidth consumption is reduced.

(1) Basic safety data-driving data required to provide non-critical locations (non-accident black spot road segments). The primary purpose of the basic safety data design is to reduce bandwidth but collect underlying traffic data (e.g., location). The basic safety data packet is small, the sampling frequency is low, the periodic driving data monitoring packet (for example, 5-second acquisition or 10-second acquisition) can be set, and the driving data is acquired according to the principle of the least bandwidth. The structure of the basic safety data packet comprises necessary automobile safety data, such as driving position, driving speed and the like;

(2) safety data is expanded, and high-density and information-rich driving data is provided for important positions (accident black spot road sections). The collected information is mainly used for judging causes in all aspects, so that the information quantity requirement is high. The information format not only contains basic safety information (such as data contained in the basic safety data), but also contains other aspects of automobile safety data, and also contains various extended information needed by decision analysis, such as navigation information, vehicle-mounted sensor data information, driver information and the like;

(3) the black spot warning data-black spot warning information is used to provide a safety warning for the driver at the black spot. The warning information mainly provides risk causes of black spots and corresponding warning recommendations, and provides corresponding danger information marks for displaying and reminding the driver. The information format not only comprises basic black point information, such as black point position information, the content of a black point warning, such as specific cause content and a vehicle speed warning, but also comprises corresponding safety recommendation warning information; after the black spot warning information is inferred according to the data in the cloud intelligent center, sending corresponding warning information and recommendation information of safe driving to a driver;

the installation process of the vehicle networking accident black spot warning system of the embodiment is as follows:

(1) relevant parts of the vehicle-mounted client system are installed at a proper position of the vehicle according to the actual condition of the vehicle, and comprise a head-up display instrument, a driving camera, a vehicle-mounted mobile phone (or vehicle-mounted front-mounted computer software for installation and adjustment), an OBD data collector, a satellite positioning system antenna, a 3G communication antenna and the like, so that the relevant instruments are guaranteed not to shield the sight line.

(2) Connecting a lead from a vehicle storage battery with a vehicle-mounted system host to ensure power supply, installing an OBD collector on an OBD interface of the vehicle, and connecting an OBD data interface with a vehicle-mounted head-up display through Bluetooth; install suitable position with driver's surveillance camera module to driving the camera and being connected through the bluetooth with the OBD data, reverse side camera can just be just to shooing driver face, and positive camera is just to the street view.

(3) The external antenna required by the external system is arranged on the outer part of the automobile, the 3G communication antenna is adjusted repeatedly, the satellite positioning system is arranged at a position which is close to the automobile shell and has good signal receiving performance, the system can receive signals with low noise influence, and the signals are stable and have no interference.

(4) After the installation and the setting of each external system are finished, the power supply is started, and the normal work of the data acquisition module (including smooth GPS data connection, 3G communication, vehicle-mounted sensing data and the like) is ensured. The APP software of the system is manually started, the system is started through the software, the data link condition of each sub-module of the client is detected through the system, the setting of the corresponding vehicle-mounted system is adjusted, and the normal work of each module is guaranteed. And opening a vehicle-mounted remote data communication port, connecting the vehicle-mounted remote data communication port with a cloud control center through 3G, testing the link condition of a communication system of the vehicle-mounted system and a cloud system, and debugging whether the real-time data coding condition is consistent at the client and the cloud.

The operation process of the vehicle networking accident black spot warning system of the embodiment is as follows:

(1) when the system is started, the vehicle-mounted client automatically detects that the automobile engine is started, the vehicle-mounted system automatically acquires the port connection between the vehicle-mounted client system and each internal and external system, if all the connections are normal, the vehicle-mounted system automatically acquires required data according to black point data acquisition information (as described above), and the data is stored in the vehicle-mounted system for standby after being acquired;

(2) in the driving process, the vehicle-mounted system opens a communication link and is connected with the cloud control center in real time, basic safety data (driving monitoring data) are sent to the cloud control center at a vehicle-mounted client of a non-accident black spot road section to keep the cloud control center monitoring the basic state of the automobile, and the position of the automobile is matched with the position judgment relation of the accident black spot of a corresponding area in real time;

(3) when the automobile is detected to enter an accident black spot area, the cloud control center starts an accident black spot mode, the vehicle-mounted client automatically changes the mode, switches to the black spot acquisition mode, acquires the expanded safety data, performs data encoding on the sensing data and sends the sensing data to the cloud control center, and after the data is received by the cloud control center, the data is decoded by the decoding module and sent to each cloud module for analysis;

(4) the cloud control center analyzes and calculates the collected vehicle-mounted data flow through the decision module according to a decision program, if it is calculated that various state data of a driver are matched with accident risks derived due to accident black points, black point warning data are generated according to a coding rule, a safety warning mechanism is started, and the black point warning data are pushed to a vehicle-mounted client system.

(5) The vehicle-mounted system receives the black spot warning data, decodes the black spot warning data, transmits the decoded black spot warning data to the head-up display instrument through Bluetooth, displays the black spot information in a visual form through establishing a matched road network model in a form of combining with a road network, and reminds a driver by means of a sound light and vibration mode.

The present invention and its embodiments have been described above schematically, without limitation, and what is shown in the drawings is only one of the embodiments of the present invention, and the actual structure is not limited thereto. Therefore, if the person skilled in the art receives the teaching, without departing from the spirit of the invention, the person skilled in the art shall not inventively design the similar structural modes and embodiments to the technical solution, but shall fall within the scope of the invention.

Claims

1. The utility model provides an accident black spot warning system based on augmented reality and high in the clouds intelligence decision-making which characterized in that: including on-vehicle client system layer (101), cellular communication network layer (102) and cloud control center layer (103), wherein:

the vehicle-mounted client system layer (101) collects driving data and sends the collected driving data to the cloud control center layer (103) through the cellular communication network layer (102); meanwhile, a data decision packet sent back from the cloud control center layer (103) is received and visually displayed to a driver in an augmented reality mode in combination with real road conditions;

the cellular communication network layer (102) realizes bidirectional communication between the vehicle-mounted client system layer (101) and the cloud control center layer (103);

the cloud control center layer (103) makes decision analysis according to the driving data received in real time and the data in the black spot database aiming at the driving state and the current situation of the vehicle, and issues corresponding warning information to the vehicle-mounted client system layer (101) in a data packet format; the data in the black spot database comprises traffic road condition data acquired from a traffic management center, historical accident black spot data and weather data acquired from a meteorological station;

the cloud control center layer (103) comprises 13 functional modules, and the functional modules are connected in series through data stream logic; wherein: the data communication module (301) receives a communication data packet transmitted by the vehicle-mounted client system, the data decoding module (302) and the video data decoding module (306) respectively decode the received communication data packet, and relevant decoding information is sent to the data analysis module (305); the accident data mining module (303) is used for mining historical traffic accident data on line or off line and storing the corresponding accident black point position, the accident cause and the result thereof in an accident black point database (304); the data integration module (307) is responsible for receiving the decoded data, integrating the data into a specified format and sending the format to the intelligent decision module (308) for operation; the external database (310) is used for storing data streams acquired from the outside, including traffic data acquired from the real-time traffic database (312) and weather data acquired from the weather database (313), and the data streams are finally transmitted to the intelligent decision module (308), the intelligent decision module calculates whether the driving data is matched with accident risks acquired by deduction of accident black points, and the results processed by the intelligent decision module (308) are encoded by the data encoding module (309) and then output; in addition, traffic management personnel can manage vehicle position, accident black spots and traffic information through the visual platform geographic information system (311).

2. The system of claim 1, wherein the system comprises: the vehicle-mounted client system layer (101) comprises at least one vehicle-mounted client system, the host of the vehicle-mounted client system comprises a vehicle-mounted front-mounted system and a vehicle-mounted rear-mounted system, the vehicle-mounted front-mounted system comprises a temperature and humidity sensor (201), an illumination pressure sensor (202), a tire pressure friction sensor (203), a wiper sensor (204), an automobile steering lamp indicating system (205), an automobile suspension system (206) and an automobile braking system (207); the vehicle-mounted after-loading system comprises a satellite positioning navigation module (209), a cellular communication module (210), a digital map module (211), a driver monitoring camera module (212) and an augmented reality display module (213); the vehicle-mounted client system integrates various data through a vehicle-mounted system display module (208).

3. The system of claim 2, wherein the system comprises: the vehicle-mounted system display module (208) is a vehicle-mounted traveling computer, an external smart phone or a head-up display instrument.

4. The system of claim 2, wherein the system comprises: the cellular communication network layer (102) adopts a TCP/IP communication protocol based on cellular communication, adopts a socket as a communication mediator, and provides communication and data transmission for communication between the cloud control center and application software through the cellular base station.

5. An accident black spot warning method based on augmented reality and cloud intelligent decision-making comprises the following steps:

the method comprises the following steps that firstly, an accident black spot warning system is started, a vehicle-mounted client system automatically acquires connection with ports of an internal system and an external system, if the connection is normal, the vehicle-mounted client system acquires required data according to black spot data information acquisition rules, namely, a road section is divided into an accident black spot road section and a non-accident black spot road section, and different information acquisition is carried out on different road sections, so that the acquisition efficiency is improved and the bandwidth consumption is reduced; basic safety data are collected on a non-accident black spot road section, and the basic safety data collect driving data according to a bandwidth minimum-saving principle; acquiring expanded safety data on the accident black point road section, wherein the expanded safety data are high-density driving data with abundant information; after data acquisition is finished, storing the data in a vehicle-mounted client system for caching for later use;

the cloud control center analyzes and calculates the acquired data, if the driving data is calculated to be matched with the accident risk obtained by deduction due to accident black points, black point warning data is generated according to the coding rule, a safety warning mechanism is started, and the black point warning data is pushed to a vehicle-mounted client system; the black spot warning data provides potential risk causes of black spots and corresponding safe operation guidance;

6. The accident blackspot warning method based on augmented reality and cloud intelligence decisions of claim 5, wherein: in the third step, the specific process of determining the black point of the driving-in accident and the driving-out accident is as follows:

step 1, collecting current GPS position data Location (x, y) based on an automobile, and converting the current GPS position data Location (x, y) into Local coordinates Location (x, y) under the Local coordinates of a current digital map according to a coordinate format, wherein the Local coordinates Location (x, y) is called LLposition (x, y) for short;

step 2, sequentially circulating the geometric center coordinates of each accident black point in the accident black point database, searching the accident black point coordinate set closest to the current local coordinate LLocation (x, y), and extracting the vertex coordinate set C (C) of the black point polygon₁，c₂，c₃…c_N)；

Step 3, calculating coordinate cross lines PL of the LLocation (x, y), crossing and intersecting the polygon through developing the cross lines PL, and calculating the number of cross points of the polygon and each sideline position; if the cross line intersects with the side line formed by the polygon, the cross point counter automatically increases the corresponding number of cross points; otherwise, the count of the cross point counter is not increased;

step 4, making an entrance decision judgment by a cross point counter according to the number of cross points, and judging that the current position of the automobile is in the accident black point if the number of the cross points is an odd number; if the number of the transverse intersection points is an even number, the automobile is outside the accident black point; when the judgment of the entering and exiting is changed, the moment when the automobile enters the accident black point is the moment, the changing moment is recorded as the time when the automobile enters the accident black point, and the moment is also the judgment condition for starting the warning or starting the data acquisition algorithm.

7. The accident blackspot warning method based on augmented reality and cloud intelligence decisions of claim 5, wherein: in the first step, historical accident black point excavation is required, and the specific process is as follows:

acquiring historical accident original data of a road in a specific traffic area in a traffic management department, sorting the accident data, cleaning incomplete data and deleting redundant data;

step (3), performing geographic information editing on historical data of traffic accident black points according to geographic positions of the traffic accident black points, designing a corresponding accident black point relation database, and establishing a spatial geographic information accident black point vector map layer which is stored in a geographic information system database as attributes related to black point data;

and (4) formatting the data of the traffic accident black points according to a navigation data format, combining the accident black point data with a navigation system according to the navigation data format, importing the data into a vehicle-mounted client system, integrating the data with a navigation map and the system, developing the data into a vehicle-mounted black point-based database, and displaying the black points on the vehicle-mounted client system and a cloud control center.

8. The method for warning the accident black spot based on the augmented reality and cloud intelligence decision as claimed in any one of claims 5-7, wherein: in the fourth step, the specific process of reasoning and decision making by the cloud control center is as follows:

step A, under an on-line condition, the cloud control center determines main accident black points in a specific geographic area range through a data mining technology, wherein the main accident black points comprise geographic positions of the accident black points and main accident causes of the accident black points, the main accident causes are classified and priority levels are determined, a rule matching and resolving template based on intelligent decision is customized on the basis, and black point risk matching rules are determined;

step B, in a specific implementation state, the cloud control center acquires driving data of the vehicle-mounted client system in a specific geographic area range, stores required information in a classified mode, meanwhile, acquires traffic data of a traffic control department and weather data of a weather station in real time, and stores the data in a cloud database;

step C, after the collected data are decoded, firstly decoding positioning information in an intelligent decision module, executing judgment of entrance and exit of accident black points, if a vehicle enters a specific black point, acquiring various accident cause data of the black point, and extracting real-time state data related to causes for intelligent decision;

step D, executing a digestion algorithm based on the rule, firstly judging a single accident cause, then sequentially judging a plurality of causes, obtaining an accident black point risk index according to a designed weight formula, and determining whether a driver gives a specific driving warning; the weight of the weight formula is given after comprehensive judgment is made according to the historical records and expert opinions of the specific road section;

e, when the cloud control center determines to give driving warning to the driver, warning information needs to be matched with a specific risk reason and is sent to the vehicle-mounted client;

and step F, the vehicle-mounted client receives the warning coding information and decodes the information to the vehicle-mounted navigation system and the augmented reality module, and the augmented reality module visually displays the information to the driver in a display mode of combining augmented reality and a real road section.