CN113963538A

CN113963538A - Intelligent highway information physical system based on data self-flowing

Info

Publication number: CN113963538A
Application number: CN202111216233.1A
Authority: CN
Inventors: 岑晏青; 刘博�; 宋向辉; 高欢; 刘宏本
Original assignee: Research Institute of Highway Ministry of Transport
Current assignee: Research Institute of Highway Ministry of Transport
Priority date: 2021-10-19
Filing date: 2021-10-19
Publication date: 2022-01-21
Anticipated expiration: 2041-10-19
Also published as: CN113963538B

Abstract

The invention provides an intelligent highway information physical system based on data self-flow, which determines the functional requirements of a physical space-a connecting channel-an information space in the information physical system, the system operation logic based on data automatic flow and an intelligent control decision and optimization mechanism based on scheme rehearsal and knowledge accumulation, can provide accurate, timely and personalized travel auxiliary information for travelers, supports the landing application of technologies such as vehicle-road cooperation and automatic driving, improves the operation efficiency of a highway network, and can effectively promote the standardized work of infrastructure in the process of establishing a landing scheme of the information physical system in the traffic field and promoting the development of the vehicle-road cooperation automatic driving technology.

Description

Intelligent highway information physical system based on data self-flowing

Technical Field

The invention relates to the technical field of intelligent traffic, in particular to an intelligent highway information physical system based on data self-flow.

Background

In terms of guaranteeing efficient management and operation of highways, a traffic operation monitoring and scheduling Center (TOCC) is designed in the existing intelligent Transportation industry and is already arranged along the highways. The use of the TOCC realizes the collection, transmission, processing and release of traffic information, simultaneously monitors the running state of an electromechanical system of the highway and the traffic state of each road section in the highway network, unifies and coordinates resources based on traffic accident detection information and construction maintenance information, realizes the information management of emergency rescue, and ensures the safe, reliable and efficient operation of the highway.

With the rapid development of technologies such as intelligent networked automobile, vehicle-road cooperation and automatic driving, the landing application of the technologies urgently requires that the information perception, information transmission and monitoring control of the highway are more accurate, real-time and efficient. In addition, with the further development of intelligent transportation construction, the requirements of travelers on intelligent travel service are further upgraded, so that efficient and safe travel is facilitated. Accordingly, the existing TOCC technology architecture and management system face significant challenges, which mainly appear as follows:

firstly, the accurate real-time information service is not sufficient. The traffic data flow mode of the existing TOCC technical architecture mostly adopts a layer-by-layer reporting mode, information calculation processing is mainly deployed on the upper-layer TOCC, information issuing is layer-by-layer issuing, the whole traffic data flow process, the lower-layer TOCC only plays a transition role, and the information calculation capacity of the lower-layer TOCC cannot be fully exerted. With the increasing scale of highway networks, the turnover of passengers and goods increases year by year, the service objects of corresponding highway management systems increase day by day, and the data transfer mode is difficult to ensure that information for travelers and managers on different road sections is uploaded and issued in real time.

And secondly, intelligentizing travel service is insufficient. In order to better serve efficient and safe trips of travelers, intelligent trips such as quasi-all-weather trips, accompanying information services, lane-level management and control services, intelligent service area induction services and the like need to be added to the existing highway management system step by step. The intelligent travel service provides new requirements for the existing traffic information sensing system, such as lane-level vehicle positioning information, high-precision map information, service area state information and the like.

And thirdly, the intelligent decision control efficiency is low. As a core capability of intelligent transportation, the level of intelligent decision control determines the efficiency of highway operation management and the associated quality of service provided to travelers. On one hand, with the new adjustment of the perception subsystem, massive traffic operation data are uploaded to the system, a service-oriented data flow design is not provided, and under the existing TOCC technical architecture, the system is difficult to quickly respond according to the current traffic operation state to carry out efficient intelligent decision control.

The Cyber-Physical Systems (CPS) is used as an extension and extension of a control system and an embedded system, fully utilizes a multi-scale and multi-probability Physical model and a Physical element comprehensive perception technology by integrating advanced perception, communication, calculation, control and other information technologies and automatic control technologies, constructs an information space capable of reflecting the state of the Physical space in real time, and timely interacts and mutually maps Physical entities, operating environments, information association and other elements of the Physical space and the information space through an associated feedback mechanism, thereby realizing on-demand response, scheme rehearsal and dynamic optimization of system resource configuration in the information space. The CPS information space construction is based on the accuracy of state sensing and the real-time performance of calculation and analysis, and can meet the requirements of accurate real-time information service and intelligent outgoing service.

Different from a Digital Twin technology (Digital Twin), the CPS describes physical entities and states thereof in a system in a time-space domain and a logic domain, abstracts the physical entities and the states into an information model in a virtual space, and also adds an intelligent decision control function. On the basis of mutual mapping of a physical space and an information space, a data-driven intelligent control scheme is previewed in the information space, the control effect of the scheme is evaluated, and iterative optimization is carried out. Unlike the TOCC technical architecture, the CPS greatly accelerates the feedback optimization and adjustment process of the control scheme, and effectively avoids the risk of poor effect of the initial control scheme.

At present, due to the defects of the TOCC technical architecture and the requirement of intelligent outgoing service, a technical architecture of the intelligent highway traffic system suitable for intelligent highway construction, supporting vehicle-road coordination and automatic driving technology landing is required to be provided. Although CPS is a relatively new research field, CPS has made a prominent contribution in the process of the development of informatization and industrialization since the concept description of CPS by the national science foundation in 2006. With the rapid development of technologies such as the Internet of things, big data and artificial intelligence, conditions are created for the land application of the CPS in a new generation of highway intelligent traffic system. The invention provides a technical framework of an intelligent highway information physical system, which is suitable for application scenes of vehicle-road cooperation and automatic driving and based on automatic data flow, and can be used for guiding the construction of the intelligent highway system.

Disclosure of Invention

The invention solves the problem that the existing intelligent expressway system with the TOCC technology architecture cannot meet the intelligent travel service requirement.

In order to solve the above problems, the present invention provides an intelligent road information physical system based on data self-flow, comprising: an information space layer and a connection channel layer connected to each other; the connection channel layer is also used for connecting with the physical space layer; the physical space layer comprises physical entities in the intelligent expressway network and interactive relations among the physical entities; the connection channel layer comprises communication network equipment and a CPS unit node controller, and the communication network equipment is in communication connection with a physical entity in the intelligent expressway network; the information space layer comprises a CPS intelligent control bus, a data information layer, an interactive mapping layer, a historical state layer, a background calculation layer, a scheme preview layer and a knowledge accumulation layer; the CPS unit node controller is in communication connection with the CPS intelligent control bus; the CPS unit node controller has the authority of administrating all controllable physical entities in the administration range, is in communication connection with the information space layer, and is used for acquiring the traffic data of the physical entities, and performing real-time state sensing and computational analysis according to the traffic data to acquire fusion sensing information and mining information based on the real-time traffic data; the CPS intelligent control bus is used for connecting the CPS unit node controllers and the data information layer, classifying and summarizing the data uploaded by the CPS unit node controllers, and distributing the data to the data information layer for storage; receiving the optimal executable scheme output by the scheme preview layer scheme, compiling and issuing the optimal executable scheme to the CPS unit node controller, and directly transmitting data with high real-time requirement in data uploaded by the CPS unit node controller to the interactive mapping layer; the interactive mapping layer comprises a plurality of virtual physical entities, and loads the data of the data information layer or the data uploaded by the CPS intelligent control bus to complete the updating of the virtual physical entities; the historical state layer is used for loading historical data stored in the data information layer or historical data transmitted by the CPS intelligent control bus so as to assist the background calculation layer in carrying out intelligent decision analysis; the background computing layer is used for carrying out real-time intelligent decision analysis according to the interactive mapping layer and the data loaded by the historical state layer so as to obtain at least one feasible scheme; the scheme preview layer is used for previewing the feasible scheme according to a prediction model and determining an optimal executable scheme according to a previewing implementation effect, and sending the optimal executable scheme to the CPS intelligent control bus so that the CPS intelligent control bus sends the optimal executable scheme to the CPS unit node controller for execution; the knowledge accumulation layer is connected with the scheme pre-modeling layer and the physical space layer and is used for storing the optimal executable scheme, receiving a feedback effect of a physical entity of the physical space layer for executing the optimal executable scheme, determining an iterative optimization scheme of the prediction model according to the feedback effect and storing prediction deviation analysis experience so as to optimize the scheme pre-modeling layer according to the iterative optimization scheme.

Further, the CPS unit node controller is configured to perform calculation and analysis according to the traffic data, generate a preliminary execution scheme, and send the preliminary execution scheme to the physical entity.

Further, the data information layer is used for storing the classification data uploaded by the CPS unit node controller in a classification mode;

the classification data comprises traffic information data, interactive relation data among physical entities and mapping rule data among the physical entities and the virtual physical entities.

Further, the interactive mapping layer is used for loading the real-time data stored in the data information layer or the real-time data uploaded by the CPS intelligent control bus, and exemplarily, the interactive mapping layer can display the regional road network macro basic graph and the fitting function information thereof, the main road network macro basic graph and the fitting function information thereof, the road section related information, the basic road network information and the traffic operation state information, the traffic event information and the facility equipment information loaded thereon.

Further, the background calculation layer is used for calculating methods or models related to system control strategy design, decision basis analysis and the like, such as a traffic accident recognition model for example; the traffic accident recognition model is used for carrying out traffic accident recognition and analysis according to traffic flow parameter change trend prediction based on particle calculation and traffic accident recognition based on Bayesian inference.

Further, the scheme preview layer is configured to preview each of the feasible schemes according to a prediction model to obtain a preview implementation effect on the basis of the interactive mapping layer in the current state, and determine an optimal executable scheme according to the preview implementation effect determined by a preset index system.

Further, the scheme preview layer is also used for evaluating the current system state, predicting the vehicle running state and predicting the traffic flow state.

Further, the knowledge accumulation layer is used for performing reason analysis when the control effect of the scheme preview layer for executing the preview scheme does not meet a preset requirement, accumulating traffic state prediction deviation analysis empirical data, and performing iterative optimization on the preview scheme until the control effect meets the preset requirement;

the knowledge accumulation layer is also used for storing traffic incident management and control experience data.

Further, the interaction relationship between the physical entities is an association rule existing between the physical entities; and a uniform compiling rule of the interaction relation between the physical entities is preset in the physical space layer.

The intelligent highway information physical system based on the data self-flow determines the functional requirements of a physical space, a connecting channel and an information space in the information physical system, the system operation logic based on the data self-flow and an intelligent control decision and optimization mechanism based on scheme rehearsal and knowledge accumulation, can provide accurate, timely and personalized travel auxiliary information for travelers, supports the landing application of technologies such as vehicle-road cooperation and automatic driving, improves the operation efficiency of a highway network, and can effectively promote the standardized work of infrastructure in the process of establishing a landing scheme of the information physical system in the traffic field and promoting the development of the vehicle-road cooperation automatic driving technology.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.

FIG. 1 is a schematic diagram illustrating a design concept of an intelligent highway information physics system according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of a technical architecture of an intelligent highway information physics system according to an embodiment of the present invention;

FIG. 3 is a functional block diagram of a connection channel according to an embodiment of the present invention;

fig. 4 is a functional architecture diagram of an information space in an embodiment of the invention.

Detailed Description

In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

The embodiment of the invention fully utilizes the technologies of Internet of things, big data, artificial intelligence, automatic control and the like, provides a technical framework of an intelligent highway information physical system based on data self-flowing, and can make up for the defects of the traditional highway traffic operation monitoring and dispatching center. The physical system for highway information designed and developed according to the technical architecture has the functions of real-time sensing of traffic operation information, information world synchronous mapping, scheme rehearsal and intelligent decision making, provides accurate, timely and personalized travel auxiliary information for travelers, supports landing application of technologies such as vehicle-road cooperation and automatic driving, and improves the operation efficiency of a highway network.

The intelligent highway information physical system provided by the embodiment of the invention is a system capable of realizing autonomous data flow, intelligent road network operation, autonomous monitoring control and iterative optimization, and can be used as a design scheme of a novel highway operation management platform. CPS is an extension of embedded systems, and the overall system design is oriented to service requirements for both the traveler and the system management.

On the basis of a basic architecture framework of a traffic information physical system, interaction with a physical world is arranged outside a hierarchical structure, so that the information system and the physical system are deeply integrated, multi-dimensional cooperation among all levels is realized according to specific services, the system architecture is optimized, the optimal data circulation process is realized, a service scheme is obtained more efficiently, and the service scheme is fed back to a physical object.

Referring to a schematic diagram of a design concept of the intelligent road information physical system shown in fig. 1, an information space capable of mapping a physical space in real time is constructed by combining a basic transportation network and distributed traffic information acquisition equipment to form a sensing layer. On the basis of the perception layer, a communication layer, a calculation layer and a control layer are constructed and superposed in an information space for the contents to be realized by the service layer, so that the information space is ensured to have the functions of real-time analysis and scientific decision. The physical world carries out comprehensive perception, feedback perception and strategy execution, and the information world carries out receiving perception data and strategy making.

The embodiment of the invention provides an intelligent highway information physical system based on data self-flow, which comprises an information space layer and a connecting channel layer which are connected with each other. The connection channel layer is used to connect the physical space layer and the information space layer.

Specifically, the physical space layer comprises physical entities in the intelligent expressway network and interactive relations among the physical entities; the connection channel layer comprises communication network equipment and a CPS unit node controller; the information space layer comprises a CPS intelligent control bus, a data information layer, an interactive mapping layer, a historical state layer, a background calculation layer, a scheme preview layer and a knowledge accumulation layer.

The communication network equipment of the connection channel layer is in communication connection with a physical entity in the intelligent expressway network of the physical space layer; the CPS unit node controller of the connection channel layer is in communication connection with the CPS intelligent control bus of the information space layer. Based on the structure, the self-flow of data among a physical space, a connecting channel and an information space can be realized. The specific functions of each component module are as follows:

the CPS unit node controller has the authority to control all controllable physical entities in the scope of jurisdiction and is in communication connection with the information space layer, can acquire the traffic data of the physical entities, and performs real-time state sensing and calculation analysis according to the traffic data to acquire fusion sensing information and mining information based on the real-time traffic data;

the CPS intelligent control bus can be connected with the CPS unit node controller and the data information layer, classifies and summarizes the data uploaded by the CPS unit node controller, and distributes the data to the data information layer for storage; receiving an optimal executable scheme output by the scheme preview layer scheme, compiling and issuing the optimal executable scheme to the CPS unit node controller, and directly transmitting data with high real-time requirement in data uploaded by the CPS unit node controller to the interactive mapping layer;

the interactive mapping layer comprises a plurality of virtual physical entities and can load data of the data information layer or data uploaded by the CPS intelligent control bus to complete the updating of the virtual physical entities;

the historical state layer can load historical data stored in the data information layer or historical data transmitted by a CPS intelligent control bus so as to assist the background computing layer to carry out intelligent decision analysis;

the background computing layer can perform real-time intelligent decision analysis according to the data loaded by the interactive mapping layer and the historical state layer so as to obtain at least one feasible scheme;

and the scheme preview layer can determine an optimal executable scheme according to the prediction model preview feasible scheme and the preview implementation effect, and sends the optimal executable scheme to the CPS intelligent control bus, so that the CPS intelligent control bus sends the optimal executable scheme to the CPS unit node controller for execution.

And the knowledge accumulation layer is connected with the scheme preview layer and the physical space layer and is used for storing the optimal executable scheme, receiving the feedback effect of the optimal executable scheme executed by the physical entity of the physical space layer, determining the iterative optimization scheme of the prediction model according to the feedback effect and storing the prediction deviation analysis experience so as to optimize the scheme preview layer according to the iterative optimization scheme.

Fig. 2 is a schematic diagram of a technical architecture of an intelligent road information physical system according to an embodiment of the present invention. The following describes each part of the intelligent road information physical system provided by the embodiment of the invention in detail with reference to fig. 2.

(1) Physical space

The physical space in the intelligent highway information physical system refers to all physical entities in the intelligent highway network and the interactive relationship among the physical entities.

The physical entities of the physical space refer to things which exist objectively and can be distinguished from each other, so that for a complex system such as an intelligent road, the physical entities relate to a wide range of physical objects, including: an infrastructure road transportation network (e.g., each lane, ramp, service area, Toll station, electromechanical facility along the line, marker line, etc.), roadside terminal devices (e.g., video monitor, radar detector, weather detector, vehicle detector, traffic event detector, variable message sign, and other intelligent roadside devices with communication function and control function, etc.) operating vehicles (e.g., general road use vehicles, emergency rescue vehicles, maintenance vehicles, special vehicles such as police cars and ambulances, etc., autonomous vehicles), and On-board devices (e.g., On-board navigator, Electronic Toll Collection-On board Unit (ETC-OBU), intelligent mobile terminal), drivers, highway operation related personnel, ETC.

The interactive relationship between the physical entities refers to an association rule existing between the physical entities, and a uniform compiling rule of the interactive relationship between the physical entities is preset in the physical space layer. Such as the connection and adjacent relation between lanes, the charging rules of the toll station for different vehicles, the interaction relation between running vehicles, the related information of the traffic flow composed of a plurality of vehicles, and the like. Because the number of physical entities is large, the interaction relationship between the physical entities is complex and is intersected with each other, and the data belongs to a large amount of implicit data contained in a physical space.

The intelligent road CPS needs to compile physical entities and interaction relations among the physical entities uniformly in a physical space. Taking a conventional road use vehicle as an example, in the system, a large number of such physical objects exist, which can be uniformly defined as an entity set a. In entity set A, any ith vehicle can be denoted as A_i. For vehicle A_iIt itself has several attributes, such as license plate number, vehicle type, vehicle color, vehicle-mounted device, current driving lane, current driving speed, etc. An in-vehicle device such as the ith vehicle may be denoted as A_i,obuIt is noted that the attributes of an entity itself may exist in another set of entities. Generally, the on-board unit (OBU) belongs to a vehicle information acquisition entity device and is described in a physical spaceWhen the vehicle information acquisition system is used, the OBU (on-board unit) is recorded as an entity set. For vehicle A_iAnd vehicle A_jThere is an interactive relationship, and such relationship includes a relative positional relationship, whether there is a lateral disturbance, whether it is a following running state, or not. For example vehicle A_iAnd vehicle A_jIn the presence of lateral interference, then A_i∩A_j(hor) ═ 1, and the interaction forces between the two can be respectively described as

And

the compiling in the physical space is a detailed description of the physical world, and is also a key point for realizing the real-time mapping of the physical world in the information space.

(2) Connecting channel

The connection channel is mainly used as a connection bridge between a physical space and an information space in the CPS and comprises an infrastructure (such as an LTE-V/5G base station, a bridge, a switch, a modem, an optical fiber network and the like) of a communication network and a CPS unit node controller. All facilities within a connection channel are subordinate to the physical space, but because of the functional particularities of their implementation, in the technical architecture they are presented as one module only.

Referring to the functional architecture diagram of the connection channel shown in fig. 3, a CPS unit node controller is shown, each CPS unit node controller being connected to a CPS intelligent control bus and connected to the device layer of the segment N.

In the device layer are shown a millimeter wave radar, a high definition camera, a traffic event detector, a weather detector, a Road Side Unit (RSU) and a plurality of vehicle OBUs. The equipment can collect vehicle type information, license plate number information, driver information, vehicle positioning information, running speed information, acceleration information, vehicle interval running speed information, traffic flow speed information, traffic flow headway distribution information, traffic event information, meteorological environment information and the like. And the CPS unit node controller stores the redundant data and the copied importance data into a temporary storage data information base, performs aggregation analysis on the data and then sends the data to a CPS intelligent control bus. The CPS intelligent control bus can issue and arrange a fusion sensing and data mining algorithm library, and a CPS unit node controller can call the fusion sensing and data mining algorithm library to perform data sensing mining; the fusion perception and data mining algorithm library can also be corrected and updated.

Specifically, the CPS unit node controller is a server which has both calculation and analysis capabilities and a capability of communicating with a CPS intelligent control bus. According to the actual situation, the CPS unit node controller can be arranged on a single road segment or node, or can be arranged on a small regional road network formed by a plurality of road segments and nodes, and has the authority of dominating all controllable physical entities in the range to which the CPS unit node controller belongs.

The CPS unit node controller can perform calculation analysis according to the traffic data, including accident recognition, traffic operation situation perception and the like, generate a preliminary execution scheme and send the preliminary execution scheme to the physical entity; the traffic operation situation awareness can comprise traffic flow state information and a traffic event time-space domain influence range.

According to traffic information collected by all devices of a device layout layer within the range, such as real-time position information, running speed information, traffic incident information, traffic environment information and the like of a vehicle, a CPS unit node controller conducts real-time state sensing and calculation analysis, converts all recessive data contained in a physical space into dominant data and adjusts the dominant data into structured data, and the CPS unit node controller not only contains a fusion sensing result of mass traffic information which can be directly detected by the device layout layer, but also contains information obtained by data mining based on the detected traffic information, such as traffic flow state information, traffic incident time-space domain influence range and the like. For example, as follows, the video image data acquired by the high-definition camera and the microwave detector data are subjected to data fusion in the CPS unit node controller to compensate for the influence of illumination, bad weather and the like on the traffic flow detection precision. Meanwhile, in order to reduce the calculation pressure of the information space and improve the response speed, the CPS unit node controller can also carry out statistical analysis on the node-level to road section-level traffic flow parameters and preliminary calculation on the change trend of the current traffic flow parameters of the governed road sections, and data preparation is made for calculating the traffic accident occurrence probability of a background calculation layer in the subsequent information space.

In addition to the accident identification scene, in the aspect of sensing the traffic operation situation, after the detector collects node-level traffic data, such as section traffic flow, the node-level traffic data needs to be collected into the CPS unit node controllers of the branch pipes to calculate average density of road section average flow, and then the CPS intelligent control bus stores the road section average density into the corresponding database in a classified mode according to road section space topological relation or other related relations. Therefore, complicated calculation is distributed in each CPS unit node controller, calculation resources are fully utilized, a background calculation layer in an information space is convenient to directly call associated data, and a corresponding macro basic diagram is drawn.

The CPS unit node controller senses the real-time state, calculates, analyzes and screens the data, uploads the screened data to a CPS intelligent control bus for aggregation analysis, classifies the data into a corresponding database, and then loads the data to an interactive mapping layer in an information space. And screening the left data and the data to be copied and stored, and temporarily storing the data and the data to be copied and stored in the CPS unit node controller until the data storage time limit is over.

In the connection channel, the CPS unit node controller needs to have the capability of issuing control decision information to all devices with communication functions in the device layer within the range to which the CPS unit node controller belongs. The control decision information is generally sent to the unit node controller by a CPS intelligent control bus in the information space. In addition, the CPS unit nodes have certain calculation and analysis capacity, and when the CPS unit nodes are applied to a low-delay scene, the unit nodes directly generate a preliminary decision scheme and send the preliminary decision scheme to the equipment layer. For example, when a major traffic accident occurs in a road segment and an affected lane needs to be quickly closed, the CPS unit node controller can directly control the closing of the lane and simultaneously issue lane closing information to the equipment layer. Emergency rescue management in the face of an accident, a cooperative control scheme and the like are generated subsequently by an information space. In addition, reliable network connection is required among the CPS unit node controllers in different belonging ranges, and the implementation conditions of real-time sharing and cooperative management and control of information are guaranteed.

(3) Information space

The information space in the intelligent highway information physical system is a virtual space which is created by fully utilizing an advanced cloud computing mode and an automatic control technology and virtually operating a plurality of digital models (virtual physical entities) equivalent to the physical entities on the basis of the operating state data of all the physical entities in the system. The information space has the functions of synchronous mapping with the physical space, scheme rehearsal, intelligent decision and the like. The information space of the intelligent highway information physical system in the embodiment of the invention has a technical framework of '1 bus +1 information layer +5 layers', and the automatic data flow mode of the intelligent highway information physical system is realized based on data flow in the information space and the correlation feedback between the information space and the physical space, so that a closed loop enabling system of traffic data comprehensive perception, real-time calculation analysis, intelligent optimization decision, accurate execution, effect feedback, iterative optimization and knowledge accumulation is formed, the 'capability' of realizing resource optimization in a certain range is given to all entities of the physical space in the system, and the overall resource allocation efficiency of the system is improved. Fig. 4 shows a functional architectural diagram of an information space.

The '1 bus' of the information space refers to a CPS intelligent control bus, the CPS unit node controller and a data information layer are connected, uploading data of the CPS unit node controller are classified and summarized, and the data are distributed to the data information layer to be stored. The CPS intelligent control bus can classify and summarize mass and multi-source traffic data and system operation data uploaded by the CPS unit node controller and distribute and store the data to the data information layer. Meanwhile, when the scheme preview layer calculates the optimal scheme capable of being issued and executed, the CPS intelligent control bus can quickly receive the scheme and compile the scheme into a command which can be directly issued by the CPS unit node controller and is convenient for the equipment layer to execute accurately.

The '1 information layer' refers to a data information layer and can store classified data uploaded by the CPS unit node controller in a classified manner; the classification data comprises traffic information data, interactive relation data among physical entities, mapping rule data among the physical entities and virtual physical entities and the like.

The data information layer can store and call various data information uploaded in the physical space, and the data information comprises conventional traffic information data, such as traffic flow, average traffic flow speed, vehicle positioning information, vehicle running speed information, traffic event information and the like, and also comprises interactive relation data between physical entities in the physical space and mapping rule data between the physical entities and the virtual physical entities. The mapping rule data mainly stores parameter correction experience of a digital model for mapping the physical entity in real time, and the digital model can be iteratively optimized by using the continuously accumulated physical entity running data of the system, so that the mapping effect of the interactive mapping layer is more accurate.

The data information layer is responsible for storing various information and mainly comprises an interactive relation database, a traffic flow information database, a traffic accident information database, an infrastructure database, a vehicle information database and the like. The interactive relation database stores interactive relations among physical entities in a physical space, such as the belongingrelations between roads and road sections, the belongingrelations between road side terminal equipment and road sections, traffic flow parameters, the belongingrelations between road sections and the like; the traffic accident information base stores traffic accident related information, such as occurrence area names (pile number block), accident numbers, accident occurrence time, occurrence places, accident categories (accident disturbance within the elastic range of the system and accident disturbance beyond the elastic range), accident influence recovery prediction time, traffic control schemes and the like; the infrastructure database is mainly divided into basic transportation road network information and equipment information, wherein the basic transportation road network information mainly comprises road names, road numbers, road lengths, road section names, road section numbers, road section lengths and the like, and the equipment information comprises equipment names, models, layout positions and the like; the vehicle information database records vehicle license plate information, vehicle type information, driver information, vehicle geometric information, speed information, positioning information and the like.

The '5 layers' are an interactive mapping layer, a historical state layer, a background calculation layer, a scheme preview layer and a knowledge accumulation layer respectively. The structure and the function of each layer are introduced as follows:

the interactive mapping layer comprises a plurality of virtual physical entities and can load data of the data information layer or data uploaded by the CPS intelligent control bus to complete the updating of the virtual physical entities. Specifically, the interactive mapping layer may load real-time data stored in the data information layer or real-time data transmitted by the CPS intelligent control bus. Illustratively, a regional road network macroscopic basic graph and fitting function information thereof, a road network main road macroscopic basic graph and fitting function information thereof, basic road network information and traffic running state information loaded thereon, traffic event information, facility equipment information and the like are displayed.

The interactive mapping layer is composed of a large number of virtual physical entities, and data of the data information layer can be loaded to finish updating of the virtual physical entities. For some virtual physical entities, when the service requirements of specific scenes are met, the requirement on the time delay standard is high, and data can be directly called in through a CPS intelligent control bus to complete real-time updating. Illustratively, the interactive mapping layer can display the regional road network macro basic graph and fitting function information (function expressions, extreme point coordinates and the like) thereof, the main road network macro basic graph and fitting function information (function expressions, extreme point coordinates and the like) thereof and road section related information (affiliated roads, road section numbers, current traffic operation states and the like), the display information is updated regularly, and designated road or road section information can be displayed according to user requirements, so that the interactive mapping layer can be used for sensing, evaluating and predicting traffic operation situations and providing quantitative decision bases for making traffic control measures.

And the historical state layer can load historical data stored in the data information layer or historical data transmitted by the CPS intelligent control bus so as to assist the background calculation layer to carry out intelligent decision analysis. The historical state layer is constructed by historical operating data loaded by the data information layer, can be understood as an interactive mapping layer of the historical state, and is mainly used for assisting the background computing layer to perform intelligent decision analysis. The history state layer is an embedded structure, and interaction mapping layers of a plurality of history time periods can be embedded in the history state layer according to actual needs. Illustratively, the historical state map layer can load accident historical states and restore accident beginning and end according to the historical data of the traffic flow parameters provided by the data information layer, thereby providing a rechecking function, facilitating the confirmation of accident responsibility and the accumulation of accident experience.

And the background computing layer can perform real-time intelligent computing analysis according to the interactive mapping layer and the data loaded by the historical state layer so as to obtain at least one feasible scheme. The background calculation layer is provided with a plurality of calculation analysis models and model correction schemes. Illustratively, the background calculation layer is used for calculating methods or models related to system control strategy design, decision basis analysis and the like, such as a traffic accident recognition model for example; the traffic accident recognition model is used for carrying out traffic accident recognition and analysis according to traffic flow parameter change trend prediction based on particle calculation and traffic accident recognition based on Bayesian inference.

Because the whole CPS technical architecture is characterized in that data automatically flows, and a calculation analysis model in a background calculation layer is corrected in real time, such as a vehicle running characteristic analysis model, different traffic flow compositions and different vehicle running characteristics appear in different time periods, each parameter in the real-time correction model is particularly important for an intelligent decision analysis result.

For example, the vehicle dynamics model of the background calculation map layer can be used for predicting the vehicle running state, and the judgment and the rehearsal of the traffic flow state can be realized by combining the traffic flow dynamic simulation model. The core model of the background calculation layer is a traffic accident recognition model, and the model consists of two parts, namely traffic flow parameter change trend prediction based on particle calculation and a traffic accident recognition method based on Bayesian inference. The former can read out the traffic flow parameter change trend before the accident happens according to the historical traffic accident information, give out the prior probability of the accident occurrence by combining the current traffic flow parameter change trend, and deduces the traffic flow parameter trend. Then, on the basis of prior probability, the traffic accident identification method based on Bayesian inference continuously performs probability correction according to traffic flow parameter data called by a data information layer or real-time data directly transmitted by a CPS intelligent control bus, so that the real-time performance of computational analysis is ensured.

The traffic accident recognition method based on Bayesian inference comprises the following steps: based on the accident occurrence prior probability obtained by analyzing historical accident data, the probability value is corrected in real time according to the change trend of traffic flow parameters, the model is input into the length of a detection area of a detector (infrastructure information database), the length of a vehicle body (vehicle information database from a video image or a vehicle-mounted OBU), the running speed of a vehicle (vehicle information database from a millimeter wave radar), the average speed of a section (traffic flow information database from a CPS unit node controller), the time distance of a vehicle head (traffic flow information database from a millimeter wave or a laser radar), and the occurrence probability of a certain type of accident in a road section and a time period to which the data belong is output.

The traffic flow parameter change trend prediction based on particle calculation can read out the traffic flow parameter change trend before the accident happens according to the historical traffic accident information and deduces the traffic flow parameter trend. The model inputs the current time and the historical time, the traffic flow, the traffic density and the traffic average running speed information (a traffic flow information database) of the current road section, the upstream road section and the downstream road section, and outputs a change interval with reasonable traffic flow parameters. The highway accident recognition technology is mainly used for calculating the prior probability required by Bayesian inference and providing the traffic flow parameter change trend, and is convenient for the Bayesian inference to correct the accident probability.

The scheme preview layer is a specific structure in a CPS technical framework, is oriented to a complex application scene, is not unique in the result of intelligent decision analysis, and is always difficult to evaluate the effect of a plurality of feasible schemes. In the scheme preview layer in this embodiment, an optimal executable scheme may be determined according to a prediction model preview feasible scheme and according to a preview implementation effect, and the optimal executable scheme is sent to the CPS intelligent control bus, so that the CPS intelligent control bus sends the optimal executable scheme to the CPS unit node controller for execution.

On the basis of the interactive mapping layer in the current state, the scheme preview layer can preview each feasible scheme according to the prediction model to obtain a preview implementation effect, and the preview implementation effect is judged according to a preset index system to determine the best executable scheme. Further, the scheme preview layer can also evaluate the current system state, predict the vehicle running state and predict the traffic flow state.

The scheme preview layer essentially previews implementation effects of a plurality of feasible schemes through a prediction model on the basis of an interactive mapping layer in the current state, and places the implementation effects into an index system of a selection scheme, so that an optimal executable scheme is decided.

And (4) carrying out decision and preview on the accident management and control scheme in the scheme preview layer, if the preview effect is not good, analyzing reasons by the knowledge accumulation layer, accumulating the traffic state prediction deviation analysis experience, and carrying out iterative optimization on the original scheme until the control effect of the preview scheme meets the control effect requirement. At the moment, the management and control scheme issues an instruction to the equipment layout layer through the CPS intelligent control bus-CPS unit node controller, and the roadside terminal equipment completes the accurate execution of the scheme. Meanwhile, the roadside terminal equipment detects the scheme control effect index in real time and feeds back the scheme control effect index to a knowledge accumulation layer of an information space, the knowledge accumulation layer analyzes whether the control effect meets the requirement or not, if the control effect does not meet the expected control effect, the knowledge accumulation layer analyzes the reason, the scheme effect feedback experience is accumulated, the scheme preview layer adjusts the management and control scheme, and the issuing, executing, feeding back, evaluating and other operations are completed again; and if the expected control effect is achieved, continuously controlling until the accident influence is reduced to an acceptable degree, recording the accident information to a traffic incident information database of the information data layer, and accumulating the traffic accident control experience to the knowledge accumulation layer.

And the knowledge accumulation layer can perform reason analysis under the condition that the control effect of the scheme preview layer for executing the preview scheme does not meet the preset requirement so as to accumulate traffic state prediction deviation analysis empirical data and perform iterative optimization on the preview scheme until the control effect meets the preset requirement. And the knowledge accumulation layer can also store traffic incident control experience data.

The knowledge accumulation layer is a key structure for automatic data flow, stores all system control schemes executed in the past in the knowledge accumulation layer, has the function of directly connecting with a physical space equipment layer, and obtains the feedback effect of the executed control schemes. According to the scheme feedback effect, the knowledge accumulation layer can analyze the deficiency of the current CPS system in scheme preview, and further provides the iterative optimization direction of the prediction model in the scheme preview layer. Through the knowledge accumulation layer, the CPS system can be made to be more intelligent along with the increase of the running time.

Taking traffic accident management and control as an example, after the accident recognition model of the background calculation layer calculates that a certain road section has a traffic accident which needs to be managed and controlled, the calculated accident information is uploaded to the scheme preview layer, and the scheme preview layer formulates a management and control scheme. For example, in the aspects of 'a toll station closes 2 entrance ramps for 10 minutes', 'vehicles in an accident area bypass through xx routes', and emergency rescue vehicles, the rescue strategies including the types, the number, the rescue paths and the like of the rescue vehicles are determined. After the management and control scheme is formulated, performing accident management and control scheme rehearsal on the scheme rehearsal layer to obtain an accident management and control effect parameter estimated value, wherein the accident area traffic volume is reduced to 70% of the original volume after 10 minutes, the rescue vehicle estimated arrival time is 5 minutes, and the accident estimated processing time is 8 minutes. According to the index system, the accident is considered to be effectively processed, the congestion caused by the accident is effectively relieved, and the management and control scheme is feasible.

If the control scheme is judged to be feasible after previewing, the control scheme can be issued to the CPS unit node controller which governs the accident area through the CPS intelligent control bus, and the CPS unit node controller controls the related facility equipment to complete the implementation of the scheme.

After the scheme is implemented, the roadside device in the physical space acquires traffic data in real time, and directly feeds back information capable of reflecting a real control effect to the knowledge accumulation layer of the information space, if the number of toll station entrance ramps is 2, after the management and control is carried out for 10 minutes, the traffic volume is reduced to 65% of the accident time, the arrival time of the rescue vehicle is 4 minutes, and the processing time is 10 minutes, if the real value of the accident management and control effect parameter meets an accident management and control index system, the formulated scheme is considered to be feasible and effective, and the management and control scheme is recorded to the scheme preview layer to serve as a management and control experience for reference.

On the contrary, if the accident control effect fed back to the knowledge accumulation layer by the roadside device is extremely poor or basically invalid and does not meet related indexes, such as that '2 toll station entrance ramps are closed, after the management and control are carried out for 10 minutes, the traffic volume of an accident area is 90% of the accident time' or 'the actual arrival time of a rescue vehicle is 15 minutes', the knowledge accumulation layer analyzes the reason of failure management and control, collects the historical management and control measure information of the previous similar events, retrains the scheme decision model, adjusts related parameters, feeds back the trained new scheme decision model to the scheme preview layer, makes a new management and control scheme, and re-issues the scheme and records the management and control information and the accident information; if the control effect is not much different from the index requirement, only accumulating the control experience and optimizing a decision model, so that the optimization of a subsequent control scheme is facilitated, a new scheme is not issued any more in the control, and control information and accident information are recorded; if the control effect is identical with the preview effect and meets the standard requirement, the model is not optimized, and only the control information and the accident information are recorded to the data information layer.

The embodiment of the invention provides a smart highway information physical system technical framework based on data self-flow by combining a networking technology and a cloud computing mode to meet the requirements and challenges of technologies such as vehicle-road coordination, automatic driving and the like on the conventional highway system, supports a vehicle-road coordination application scene and various smart information issues required by intelligent internet vehicle operation, provides basic technical support for the design and development of a new-generation intelligent traffic control system which is the core of the conventional smart highway, and effectively improves the operation efficiency and service level of the highway network.

Furthermore, the invention determines the functional requirements of a physical space, a connecting channel and an information space in the information physical system, the system operation logic based on automatic data flow and an intelligent control decision and optimization mechanism based on scheme rehearsal and knowledge accumulation, and can effectively promote the standardization work of infrastructure in the process of making a landing scheme of the information physical system in the traffic field and promoting the development of the vehicle-road cooperative automatic driving technology.

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent highway information physical system based on data self-flow is characterized by comprising: an information space layer and a connection channel layer connected to each other; the connection channel layer is also used for connecting with the physical space layer; the physical space layer comprises physical entities in the intelligent expressway network and interactive relations among the physical entities;

the connection channel layer comprises communication network equipment and a CPS unit node controller, and the communication network equipment is in communication connection with a physical entity in the intelligent expressway network; the information space layer comprises a CPS intelligent control bus, a data information layer, an interactive mapping layer, a historical state layer, a background calculation layer, a scheme preview layer and a knowledge accumulation layer; the CPS unit node controller is in communication connection with the CPS intelligent control bus;

the CPS unit node controller has the authority of administrating all controllable physical entities in the administration range, is in communication connection with the information space layer, and is used for acquiring the traffic data of the physical entities, and performing real-time state sensing and computational analysis according to the traffic data to acquire fusion sensing information and mining information based on the real-time traffic data;

the CPS intelligent control bus is used for connecting the CPS unit node controllers and the data information layer, classifying and summarizing the data uploaded by the CPS unit node controllers, and distributing the data to the data information layer for storage; receiving the optimal executable scheme output by the scheme preview layer scheme, compiling and issuing the optimal executable scheme to the CPS unit node controller, and directly transmitting data with high real-time requirement in data uploaded by the CPS unit node controller to the interactive mapping layer;

the interactive mapping layer comprises a plurality of virtual physical entities and is used for loading the data of the data information layer or the data uploaded by the CPS intelligent control bus so as to complete the updating of the virtual physical entities;

the historical state layer is used for loading historical data stored in the data information layer or historical data transmitted by the CPS intelligent control bus so as to assist the background calculation layer in carrying out intelligent decision analysis;

the background computing layer is used for carrying out real-time intelligent decision analysis according to the interactive mapping layer and the data loaded by the historical state layer so as to obtain at least one feasible scheme;

the scheme preview layer is used for previewing the feasible scheme according to a prediction model and determining an optimal executable scheme according to a previewing implementation effect, and sending the optimal executable scheme to the CPS intelligent control bus so that the CPS intelligent control bus sends the optimal executable scheme to the CPS unit node controller for execution;

the knowledge accumulation layer is connected with the scheme pre-modeling layer and the physical space layer and is used for storing the optimal executable scheme, receiving a feedback effect of a physical entity of the physical space layer for executing the optimal executable scheme, determining an iterative optimization scheme of the prediction model according to the feedback effect and storing prediction deviation analysis experience so as to optimize the scheme pre-modeling layer according to the iterative optimization scheme.

2. The system as claimed in claim 1, wherein the CPS unit node controller is configured to perform calculation analysis based on the traffic data and to generate a preliminary execution plan and to issue it to the physical entity.

3. The system according to claim 1, wherein the data information layer is used for storing classification data uploaded by the CPS unit node controller;

4. The system as claimed in claim 1, wherein the interaction mapping layer is configured to load real-time data stored in the data information layer or real-time data uploaded by the CPS intelligent control bus.

5. The system according to claim 1, wherein the background computation layer is used for computation methods or models involved in system control strategy design and decision-making basis analysis, and the models comprise traffic accident recognition models;

the traffic accident recognition model is used for carrying out traffic accident recognition and analysis according to traffic flow parameter change trend prediction based on particle calculation and traffic accident recognition based on Bayesian inference.

6. The system according to claim 1, wherein the scheme preview layer is configured to preview each of the feasible schemes according to a prediction model to obtain a preview implementation effect based on the interactive mapping layer in the current state, and determine an optimal executable scheme according to the preview implementation effect determined by a preset index system.

7. The system of claim 6, wherein the solution preview layer is further configured to evaluate a current system state, a predicted vehicle operating state, and a predicted traffic flow state.

8. The system according to claim 1, wherein the knowledge accumulation layer is configured to perform cause analysis when a control effect of a preview scheme executed by the scheme preview layer does not meet a preset requirement, to accumulate traffic state prediction deviation analysis empirical data, and to perform iterative optimization on the preview scheme until the control effect meets the preset requirement;

9. The system according to any one of claims 1-8, wherein the interaction relationship between the physical entities is an association rule existing between the physical entities;

and a uniform compiling rule of the interaction relation between the physical entities is preset in the physical space layer.