CN114443784A

CN114443784A - Local dynamic map implementation method based on high-precision map

Info

Publication number: CN114443784A
Application number: CN202011205222.9A
Authority: CN
Inventors: 王平
Original assignee: Shanghai Zhucheng Information Technology Co ltd
Current assignee: Shanghai Zhucheng Information Technology Co ltd
Priority date: 2020-11-02
Filing date: 2020-11-02
Publication date: 2022-05-06

Abstract

A local dynamic map implementation method based on a high-precision map comprises the steps of firstly determining a region to be processed of the high-precision electronic map, extracting map information and constructing a static layer of the local dynamic map; then, considering different characteristics and complex relations of intelligent traffic participants, respectively constructing entity frames and a knowledge map world model through an LCM (liquid Crystal Module), and storing and updating local dynamic map information by using a map database; when the density of the traffic participants is high, LDM interaction among different traffic participants in a small range is realized based on the 5GV2X and C-V2X technologies, and LDM construction, storage, updating and local interaction are realized. The invention has the characteristics of high precision and reasonable storage, can fully utilize the multi-node property of the local LDM information of each traffic participant, and realizes LDM construction, storage, updating and local interaction.

Description

Local dynamic map implementation method based on high-precision map

Technical Field

The invention relates to a technology in the field of intelligent navigation, in particular to a local dynamic map creating, storing and updating interaction method based on a centimeter-level map.

Background

The Local Dynamic Map (LDM) technology combines a static digital Map, i.e., a Geographic Information System (GIS) Map, with various traffic Dynamic information to form a comprehensive environment-aware description, which becomes a key technology for integrating static, temporary and Dynamic information in a geographic environment, and has a main function of maintaining the latest state of acquiring traffic conditions and providing specific LDM information to specific participating objects in a limited area. At present, no completely mature LDM construction, storage and updating scheme exists, most LDMs are realized based on a common low-precision electronic map and a traditional relational database, accurate real-time information required by intelligent transportation (such as automatic driving) cannot be provided, a key sensing node needs to be selected for data uploading in consideration of the problem that a target is repeatedly sensed in an intelligent transportation scene, and compared with a mode of uploading all data, the new selection mode selects as few intelligent networking automobile nodes as possible, so that redundancy is reduced.

Disclosure of Invention

Aiming at the defects in the prior art, the invention provides a local dynamic map implementation method based on a high-precision map, which has the characteristics of high precision and reasonable storage, and can fully utilize the multi-node property of local LDM information of each traffic participant to realize LDM construction, storage, update and local interaction.

The invention is realized by the following technical scheme:

the invention relates to a local dynamic map implementation method based on a high-precision map, which comprises the steps of firstly determining a region to be processed of the high-precision electronic map, extracting map information and constructing a static layer of the local dynamic map; then, considering different characteristics and complex relations of intelligent transportation participants, respectively constructing an entity frame and a knowledge map world model through LCM (Lightweight Communications and Marshall), and storing and updating local dynamic map information by using a map database; when the density of the traffic participants is high, LDM interaction among different traffic participants in a small range is realized based on the 5G V2X and C-V2X technologies, and LDM construction, storage, updating and local interaction are realized.

The invention relates to a system for realizing the method, which comprises the following steps: the device comprises a static layer information extraction unit, an entity model construction storage unit and a node selection and information interaction unit, wherein: the static layer information extraction unit is connected with the entity model construction storage unit and provides high-precision geographic information for the LDM, the entity model construction storage unit realizes real-time storage and processing of information, and the node selection and information interaction unit is an LDM construction basic module and is respectively connected with the static layer information extraction unit and the entity model construction storage unit to cooperate to improve construction efficiency and real-time performance of the LDM.

Technical effects

The invention integrally solves the technical problems that the prior art is low in positioning accuracy, can only realize lane-level information (such as roads and intersections), cannot refine the geographic environment and cannot carry out LDM construction and information storage in a complex traffic environment, such as a road section with high traffic flow density.

Compared with the existing dynamic map construction mode based on a common electronic map, a central system LDM as a center and a traditional relational database, the method introduces a high-precision electronic map and map database storage based on a world model, and simultaneously increases an LDM interaction scheme within a certain range. By constructing, storing and updating the interactive LDM, the LDM accuracy can be greatly improved, the time delay of communication and data processing can be reduced by the graph database storage and the small-range LDM interaction, dynamic and real-time perception and updating can be realized by combining the LCM information interaction technology, the system efficiency is improved, and resources are saved.

Drawings

FIG. 1 is a flow chart of the present invention;

FIGS. 2a and 2b are the overall architecture diagrams of the LDM of the present invention;

FIG. 3 is a LDM data hierarchy diagram according to the present invention;

FIG. 4a is a flow chart of LDM high-precision map data processing, and FIG. 4b is a process of LDM high-precision map data extraction and storage according to the present invention;

FIG. 5a is a flow chart of LDM data access process, FIG. 5b is a schematic diagram of data storage entities and relationships, and FIG. 5c is a diagram of world model entity relationships based on a knowledge-graph;

fig. 6a is an LDM interaction flow chart between traffic participants in a road segment with a large traffic density according to the present invention, and fig. 6b is an interaction schematic diagram.

Detailed Description

The intelligent traffic participants in the embodiment are universal, for example, various automobiles equipped with various sensors, roadside devices, various traffic infrastructures, network monitoring centers and even pedestrian handheld devices can use the method to construct, store and update the interactive LDM under the wireless network link. The equipment terminal sends data acquired by various sensors to an upper data processor, such as roadside equipment, through mobile communication or various wireless communication equipment interaction information, the data center is used for constructing LDM, simultaneously subscribes real-time LDM information through LCM, and shares LDM information through a short-distance communication mode, such as 5G V2X, within a certain range.

As shown in fig. 2b, the present embodiment relates to a method for constructing, storing, updating, and locally interacting an LDM based on a high-precision map and graph database storage technology, which includes the following steps:

1) determining a to-be-processed area of a high-precision electronic map, extracting map information and constructing a static layer of a local dynamic map, which specifically comprises the following steps:

1.1) selecting an SHP (Shappefile) file of a local area high-precision map, and obtaining data information of each entity object of the map from the SHP file by using a Post GIS;

as shown in fig. 4a, the high-precision map data processing flow includes two directions, the first direction is information extraction and storage, and the second direction is pure map release for various display terminals;

1.2) after the SHP file is obtained, taking the node as an entity object and the edge as a relation between the entity objects, storing static map entity data into a map database, and synchronizing the SHP file to a map server for rendering and issuing at the same time, as shown in FIG. 5 b;

the physical objects include, but are not limited to, lane lines.

The node attributes include, but are not limited to, color, position, shape.

The edges include but are not limited to membership, containment.

The release is used for calling traffic participants, such as traffic monitoring center display and vehicle-mounted display.

As shown in fig. 4B, the high-precision map information is extracted and stored in a schematic diagram of a database, and in the storage stage, an entity is constructed by using LCM-gen, wherein an SHP file is a high-precision map file at a certain intersection, a map information extraction tool is Postgre SQL, an electronic map publishing server is GeoServer, a storage database is Neo4j, the entity construction is realized by using LCM programming, nodes a, B and C are three map data entities, a is a lane marking arrow, B is a lane obstacle, and C is a road intersection central point.

2) As shown in fig. 5, in consideration of different characteristics and complex relationships of intelligent transportation participants, entity frames and a knowledge map world model are respectively constructed through LCM programming, and local dynamic map information is stored and updated by using a map database, which specifically includes:

2.1) monitoring multi-process data in real time through an LCM notification/subscription function, constructing a frame structure of a corresponding entity according to monitored information of each traffic participant object, and storing the acquired information as an intermediate state.

The LCM is a collection of libraries and tools for message passing and data marshalling, and can be obtained by those skilled in the art through the following methods: http:// lcm-proj.

The multi-process data refers to: and uploading the communication process data by the traffic participants.

As shown in fig. 1, the physical frame structure includes: the name, attribute and state of the entity object, and the incidence relation and state record list between the entity object and other entities; further, different names, attributes, states, associations, and state records may be included for different types of data entities.

The attributes include but are not limited to the shape, size, dimension, color and other characteristics of the entity;

the state includes but is not limited to the information of the position, speed, acceleration, motion direction and the like of the entity;

2.2) defining and recording the association relation existing among the entities through pointers according to the static map entity data stored in the map database in the step 1.2).

The association relationship includes, but is not limited to, membership, inclusion, distance, etc.

As shown in fig. 5B, which is a schematic diagram of map database storage, A, B, C are 3 node entities, which are respectively a moving vehicle a, a moving vehicle B, and a traffic light C in operation. Each node has its own attribute and state, the edge represents the association between each node, and the association in the actual storage code is represented by a pointer;

2.3) generating and maintaining an event information list of one time dimension aiming at the event information of each entity, and the information of 0 moment is shown in figure 1.

The event information list stores the corresponding state of the entity within a certain time range.

The event information list preferably further records a future state predicted by a prediction algorithm, and the prediction algorithm is implemented by using a target state prediction algorithm which is not limited to a time series.

2.4) constructing a knowledge graph world model through the LCM-gen, storing the entity information into a graph database, and accessing and maintaining the entity information through the LCM-gen, wherein the method specifically comprises the following steps: storing each node, namely the relationship between entities, namely the association between entities in a graph mode by using a pointer, namely, starting from a bottom layer node, constructing a knowledge graph through the relationship between the nodes, and realizing data layering at the same time, wherein the first layer uses high-precision map information, and other layers of information are dynamically acquired from an LCM-IPC (liquid Crystal display-personal computer) as shown in FIG. 3.

Preferably, in the process of building the knowledge-graph world model, when entity information is ambiguous, the entity information is stored as a concept object or a knowledge-graph algorithm is called to deduce the information.

As shown in FIG. 5c, an example is constructed for a world model based on a knowledge-graph, comprising: the data of each layer of LDM and the lower layer of the knowledge graph are the basis of calculation and reasoning of higher layers. In the knowledge graph, the top-level entity is a universal world model object which has a standard object library storing typical objects of various rules summarized according to experience. The sub-entities of the top-level entity include static objects, semi-dynamic objects and dynamic objects, which are all objects identified according to various sensor sensing signals and stored after being acquired by the LCM, may be irregular and atypical, but belong to objects which need to be tracked by paying attention to the current vehicle, and are actually existing entities, such as other vehicles, traffic lights and pedestrians.

The entity information is not clear: when the sensing signal is greatly interfered, no external corresponding entity may exist after processing, the concept object is used for representing at the moment, uncertainty exists when the uncertainty model is used for representing classification, and an algorithm can be triggered to carry out merging, calculation and deduction according to the relation in the known map so as to obtain the attributes of the related entity and the relation. And the attributes in the entities and the relations can be modified according to the identified results, so that dynamic and real-time perception and knowledge updating are realized.

The deduction of the knowledge graph algorithm refers to the following steps: the method comprises the following steps of traditional rule-based reasoning, distributed representation-based reasoning, neural network-based reasoning and a mixed reasoning algorithm based on the method. The modeling algorithm of multi-source information fusion based on DKRL is used in the embodiment.

And 2.5) organizing the association among the nodes according to time through the dynamic and semi-dynamic information acquired by the LCM-IPC from the multi-process data, and modifying the attributes in the entities and the relationships, thereby realizing dynamic and real-time perception and knowledge updating and completing the construction of the LDM.

The dynamic and semi-dynamic information refers to: including but not limited to vehicle information, traffic light information, pedestrian information, traffic conditions, weather, etc.

The time organization refers to that: and corresponding according to the event information list of each time dimension generated and maintained by the event information of each entity in the step 2.3).

2.6) pushing the LDM constructed in the step 2.5) to the traffic participants in a subscription/notification mode through the LCM.

3) As shown in fig. 6, for traffic flow density K ═ N/L, where: n is the number of vehicles, L is the length of a road section, and when K is larger than a set threshold value, LDM interaction among different traffic participants in a small range is realized based on the technologies of 5G V2X and C-V2X, so that dynamic data redundancy is reduced, and information updating efficiency and real-time performance are improved; otherwise, each intelligent transportation participant is set to directly upload/subscribe the data, and the method specifically comprises the following steps:

as shown in fig. 6a, it is a flow chart of LDM interaction among traffic participants for a road segment with a large traffic flow density; as shown in fig. 6b, it is an LDM interaction diagram, wherein A, B, C, D are all traffic participant nodes (cars); the central nodes selected in the figure are node B and node F.

3.1) based on the communication technology of 5G V2X and C-V2X, selecting a central node from traffic flow according to the geographical position of a traffic participant, and constructing a small-range communication network, specifically: firstly, each vehicle judges whether the vehicle is a traffic head vehicle or not through the position and the course sensed by the vehicle-mounted equipment, the vehicle head vehicle in fig. 6B is the vehicle A, the vehicle head vehicle sends the position information and the sensing range information of the vehicle head vehicle to the sensed vehicle rear vehicle, the vehicle rear vehicle judges whether the distance between the vehicle head vehicle and the sensed vehicle rear vehicle is smaller than R, and the vehicle with the farthest distance is selected from the vehicle A and the vehicle rear vehicle as the first selected vehicle, because the vehicle B is farthest in the range of the vehicle R of the vehicle A, the first key vehicle is located in the communication range of the vehicle head vehicle, if the vehicle B is selected as the first selected vehicle in fig. 6B, namely, the central node uploads data, whether the vehicle is provided with V2X equipment or not and works normally or not is judged, whether the central node is available or not is judged, if the vehicle is not available, the next distant vehicle is selected as the central node, and the like, and if the vehicle B is not available, the vehicle A needs to be selected as the central node. And so on, selecting the subsequent node F as the second key node.

The communication network is as follows: radius R is the circular range of on-board unit (OBU) communication distance or roadside unit (RSU) communication distance, with each participant in the communication network being a node; taking the node B as an example, A, C, E nodes are included in the communication coverage area at this time;

3.2) the central node in the communication range uploads the sensing information and subscribes the LDM information, and other nodes in the communication range of the central node upload the sensing information and subscribes the LDM information generated in the step 1) and the step 2) to the central node, thereby realizing real-time and rapid LDM transmission sharing and updating.

Taking node B as an example, the central node directly communicates with the data center bidirectionally through the road side device, the wireless base station and the network, accesses the LDM resources of the data center or takes the coverage of the node B as an example, A, C, E node directly performs bidirectional LDM interaction with the node B, thereby realizing real-time LDM update.

The LDM transmission comprises: ascending, namely sensing information transmitted to the central node by other nodes, and transmitting the sensing information of the central node and other nodes to the LDM by the central node; and downlink, namely the LDM information transmitted by the LDM to other nodes of the central node and subscribed by the central node, and the LDM information transmitted by the central node to other nodes of other nodes and subscribed by the other nodes.

The updating means that: and the LDM updates the data in the graph database according to the data uploaded by the central node.

3.3) when a new traffic participant enters the communication network, adding the traffic participant into the communication network and updating the LDM shared in the network and uploading the LDM to the road side unit and the data center.

Taking the coverage range of the node B as an example, judging whether a new node is added into the communication range of the node, wherein the node D in the graph is a node to be newly added, adding the node into a node network, establishing communication with a central node, updating the LDM and synchronizing to a data center;

the updating of the shared LDM in the network refers to: the entities in the LDM and their association information with other entities are modified/updated.

The road side unit includes, but is not limited to, RSU, AP, LTE-V, camera, etc.

The data center is characterized in that: the method comprises the steps of generating a cloud server for maintaining and issuing the LDM.

3.4) when the existing traffic participant leaves the communication network, deleting the traffic participant from the communication network, updating the LDM shared in the network and uploading the LDM to the roadside unit and the control center.

Taking the coverage of the node B as an example, judging whether a node leaves the communication range of the node, wherein the node E in the graph is a node to be newly added, moving the node out of the node network, disconnecting the communication of the central node, updating the LDM and synchronizing to the data center.

Preferably, whether the node density is lower than a dismissal threshold value or not is periodically judged, the network is dismissed when the condition is met, the nodes directly and bidirectionally communicate with the data center through the road side equipment, the wireless base station and the network, and the LDM resources of the data center are accessed.

As shown in fig. 2, the data center obtains information of traffic participants from outside through the LCM and performs access operation to the graph database, and the external device also subscribes to the LDM information through the LCM.

Through specific practical experiments, a simulation platform is MATLAB + PRESCAN + Neo4j, MATLAB + PRESCAN + Mysql is adopted in a comparison scheme, and the simulation process comprises the following steps:

A. scene parameter setting

Table 1 scene parameter setting table

Parameter(s)	Numerical value
		Intelligent networking automobile number	86
Number of general vehicles	14
		Number of pedestrians	6
Number of lanes	8*4
		Width of lane	3.5m
Number of vehicle-mounted V2X devices	86
		Number of radar units on vehicle	86
V2X radar sensing radius R	20m
		Number of cameras on road side	4
Number of traffic signal lamps	4*2
		Threshold value of traffic flow density	40 pieces/km

B. Map with a plurality of maps

And (3) high-precision map shapefile files of certain intersection (precision: error is less than 10 cm). The comparison scheme adopts the OpenStreetMap file of the ordinary electronic map of the intersection.

C. Test results

TABLE 2 test results

In summary, it can be seen that the significant improvements of the present invention over the prior art include:

1. providing more accurate underlying static map information. The high-precision map is used as static map information, the precision can reach centimeter level, various element data in the high-precision map are more detailed, and the position and shape information is more precise, so that the method is favorable for providing lane-level navigation information.

2. The method adapts to the characteristics of various data in intelligent traffic and the complex relationship among traffic participants, abandons the traditional relational database, adopts a more practical database to store LDM, and is easy to expand and maintain. By introducing a world model, entity frames are constructed, the relationship between data and data exists in nodes and edges, and the concept of a base table field is not provided. If the relationship among data is complex, the data exists in a plurality of tables and intermediate tables exist, the traditional database needs to inquire some data and obtain the data only through various join table operations, the sql statement is complex, maintenance is not facilitated, and meanwhile performance is not high. The graph database can realize the query modification (entity attribute, state and pointer modification) function through simple sentences, and meanwhile, the execution speed is high, thereby being beneficial to ensuring the real-time property of the LDM.

3. The association between the nodes is organized according to time, so that the data redundancy is greatly reduced while the change of the data in the time dimension can be described in real time: as the additional overhead ratio in the test result shows, the data redundancy is directly reduced by constructing in a node selection mode.

4. The method has expandability. Data obtained by the LCM can trigger various algorithms to carry out combination, calculation and deduction according to the relation in the knowledge graph, and attributes of related entities and relations are obtained. Therefore, the problem that the sensor entity data acquired by the LCM is inaccurate is solved.

5. The real-time performance and the effectiveness of the LDM information are ensured by fully utilizing the information interaction among the traffic participants. By constructing a local communication network on a road section with a complex traffic condition, LDM sharing in a small range is realized, and participants do not need to request LDM from an intelligent center, so that the pressure of the LDM in the center is reduced, the efficiency is improved, the risk of high concurrency delay is reduced, and resources are saved.

The foregoing embodiments may be modified in many different ways by those skilled in the art without departing from the spirit and scope of the invention, which is defined by the appended claims and all changes that come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

Claims

1. A local dynamic map implementation method based on a high-precision map is characterized by comprising the steps of firstly determining a region to be processed of the high-precision electronic map, extracting map information and constructing a static layer of the local dynamic map; then, considering different characteristics and complex relations of intelligent traffic participants, respectively constructing entity frames and a knowledge map world model through an LCM (liquid Crystal Module), and storing and updating local dynamic map information by using a map database; when the density of the traffic participants is high, LDM interaction among different traffic participants in a small range is realized based on the technologies of 5G V2X and C-V2X, and LDM construction, storage, updating and local interaction are realized;

the step of constructing the static layer of the local dynamic map comprises the following steps: selecting an SHP file of a high-precision map of a local area, and obtaining data information of each entity object of the map from the SHP file by using a Post GIS; and then, the LCM-gen is used for constructing an entity in the form of the relationship between the nodes serving as the entity objects and the edges serving as the entity objects, and static map entity data is stored into a map database.

2. The local dynamic map implementation method based on the high-precision map as claimed in claim 1, wherein when the traffic flow density is greater than a set threshold, the LDM interaction among different traffic participants in a small range is implemented based on 5G V2X and C-V2X technologies, specifically: each vehicle judges whether the vehicle is a vehicle flow head vehicle or not through the position and the course sensed by the vehicle-mounted equipment, the head vehicle sends the position information and the sensing range information of the vehicle to the sensed rear vehicle, the rear vehicle judges whether the distance between the rear vehicle and the head vehicle is smaller than the communication distance range or not, the vehicle which is farthest away, namely, the distance between the rear vehicle and the head vehicle is smaller than the communication distance range is selected from the communication distance range and is closest to the communication distance range is used as a first selected vehicle, the first selected vehicle is used as a central node to upload data, the central node in the communication range uploads the sensing information and subscribes LDM information, and other nodes in the communication range of the central node upload the sensing information and subscribes the LDM information to the central node, so that real-time and quick LDM transmission sharing and updating are realized.

3. The local dynamic map implementation method based on the high-precision map as claimed in claim 2, wherein whether the central node is available is determined according to whether the first selected vehicle is equipped with V2X equipment and whether the first selected vehicle works normally, and when the central node is unavailable, the second distant vehicle is further selected as the central node.

4. The method for realizing the local dynamic map based on the high-precision map as claimed in claim 1, wherein the step of respectively constructing the entity frame and the knowledge map world model by the LCM specifically comprises the steps of: monitoring multi-process data in real time through an LCM notification/subscription function, constructing a frame structure of a corresponding entity according to monitored object information of each traffic participant, and storing the acquired information as an intermediate state; defining and recording the association relationship existing between the entities through a pointer according to the static map entity data stored in the map database; generating and maintaining an event information list of a time dimension for the event information of each entity; constructing a knowledge graph world model through the LCM-gen, storing entity information into a graph database, and accessing and maintaining the entity information through the LCM-gen; dynamic and semi-dynamic information acquired from multi-process data by LCM-IPC organizes association among nodes according to time, modifies attributes in entities and relationships, thereby realizing dynamic and real-time perception and knowledge update and completing construction of LDM.

5. The local dynamic map realization method based on high precision map as claimed in claim 4, wherein the constructed LDM is pushed to the traffic participants by LCM in a subscription/notification manner.

6. The method for implementing the local dynamic map based on the high-precision map as claimed in claim 2, wherein the updating means: when a new traffic participant enters a communication network, adding the traffic participant into the communication network, updating the LDM shared in the network and uploading the LDM to a road side unit and a data center; when the existing traffic participant leaves the communication network, the traffic participant is deleted from the communication network, and the LDM shared in the network is updated and uploaded to the roadside unit and the control center.

7. The method as claimed in claim 2, wherein the method for implementing a local dynamic map based on a high-precision map periodically determines whether the node density is lower than a resolution threshold, and when the condition is satisfied, the network is resolved, and the node directly communicates with the data center in a bidirectional manner through the road side device, the wireless base station and the network to access the LDM resource of the data center.

8. A high-precision map-based local dynamic map system for implementing the method of any preceding claim, comprising: the device comprises a static layer information extraction unit, an entity model construction storage unit and a node selection and information interaction unit, wherein: the static layer information extraction unit is connected with the entity model construction storage unit and provides high-precision geographic information for the LDM, the entity model construction storage unit realizes real-time storage and processing of information, and the node selection and information interaction unit is an LDM construction basic module and is respectively connected with the static layer information extraction unit and the entity model construction storage unit to cooperate to improve construction efficiency and real-time performance of the LDM.