An integrated platform of production-learning-research for visualized drag and drop of telematics data fusion analysis

Abstract

The application belongs to the technical field of vehicle networking data analysis system, and is specifically an industry-university-research integration platform supporting visual drag for vehicle networking data fusion analysis, comprising a node management module, a data flow construction module, a program running connection module and a developer community module.The integration platform can integrate academic research, teaching platform and industrial application together, realize industry-university-research integration, solve the problems that business personnel cannot customize algorithm behavior and cannot reuse existing components when using the data analysis system, solve the problem that researchers cannot practically land the developed algorithm into the business system, and solve the problems that the running environment does not match and the algorithm logic cannot be intuitively displayed for teachers and students.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of Internet of Vehicles data analysis system, and more particularly relates to an industry-university-research integration platform supporting visualized drag for Internet of Vehicles data fusion analysis. BACKGROUND

[0002] Analyzing and researching the data of Internet of Vehicles is a problem that must be considered by enterprises engaged in the Internet of Vehicles industry. Advanced and effective data analysis methods can fully play the important role of data in promoting management, participating in decision-making, and research landing. In scientific research, researchers need to preprocess data sets and use them for further data analysis to verify the effectiveness of methods or models. In the business department, the business system needs to analyze and process the related data generated by Internet of Vehicles in real time to realize the functions of monitoring and alarming of abnormal data. In enterprise management, enterprise personnel need to analyze market data from multiple dimensions to obtain market research reports and use them for enterprise product decision-making.

[0003] In the current Internet of Vehicles data analysis system, the common problem is that the use process of the system is biased towards computational thinking rather than business thinking, and the complexity of business analysis by business personnel in the use process has not been reduced. At the same time, the traditional data analysis process focuses on analyzing structured data, making it difficult for business personnel to effectively use artificial intelligence algorithm technology to empower products. The existing Internet of Vehicles data fusion analysis platform in the market has problems such as difficulty in calling operators, difficulty in reusing components, and not supporting custom training models. SUMMARY

[0004] To achieve the above purpose, the technical solution adopted by the present application is to provide an industry-university-research integration platform supporting visualized drag for Internet of Vehicles data fusion analysis, comprising a node management module, a data flow construction module, a program running connection module and a developer community module.

[0005] The node management module is used to create new algorithm nodes or edit existing algorithm nodes, and manage all nodes.

[0006] The data flow construction module is used to support instantiating different types of nodes in the form of visualized drag into a project, and establishing a one-way connection relationship between these instance nodes to form a data analysis flow project.

[0007] The program running connection module is used to connect the code program encapsulated in the node with the code interpreter of the background service to realize the running of the node program and the sharing of the memory of the background service, and realize the integration of debugging, development and deployment of the platform.

[0008] The developer community module is used to support individual developers to publish encapsulated nodes to the community, and these nodes are downloaded and reused by others.

[0009] Optionally, the node management module can create a new algorithm node in the node library. A platform user can encapsulate a piece of code into a node according to the specified form requirement, and define the data types and formats of the input data and output data of the node, and additionally define the configuration items of the node, so that the platform user can modify the parameter values in the configuration items according to the actual project scenario, and then instantiate the node to convert the input data into output data through the operation of the code and output the output data.

[0010] Optionally, the node management module can edit any existing algorithm node, modify and edit the algorithm node encapsulated by others according to specific requirements, and convert it to the node library of the current project and use it.

[0011] Optionally, the data flow construction module supports instantiating different node types in the node library into the current project, and modifying the configuration items of a node to make it an instantiated node. Then a one-way connection relationship can be established between two instantiated nodes, which defines the mapping between part of the output data of the previous node and part of the input data of the next node, realizes the data flow between the instantiated nodes, and establishes a connection relationship between multiple nodes. Finally, a data analysis flow project that can complete the entire data analysis task is formed.

[0012] Optionally, the program running connection module connects the code encapsulated in the instantiated node with the code interpreter of the backend service through the message queue when the project is running, so that the running of the node program shares the memory with the background service. When the node runs, the specific value of each variable can be monitored by the background service, and the running state of each node in the project can be viewed and debugged in the background service, realizing the integration of debugging, development and deployment of the platform.

[0013] Optionally, the steps of the industry-university-research integration platform for developing a data analysis project include:

[0014] Step (1): encapsulating a data analysis logic algorithm into a node;

[0015] Step (2): developing a node for reading in Internet of Vehicles data;

[0016] Step (3): connecting algorithm nodes to complete the development of a data analysis flow;

[0017] Step (4): debugging and running the data analysis flow.

[0018] Optionally, in step (1), in the industry-university-research integration platform of the present application, the running logic of one analysis stage of data analysis is implemented as a code segment, and the code is encapsulated into a node, and the data field type and format of the input data and output data of the node are customized, and the node instance will perform calculation on the received input data through the program logic defined by the code block in the running process, and the calculation result will be transmitted out through the customized output; at the same time, the variable part inside the program logic of the code block is configured;

[0019] The self-written program logic is encapsulated into a reusable node, and is reused in the development of other data analysis flows by modifying the configuration items, defining the input and output, and even developing secondary; the specific operation steps of encapsulating a node are as follows:

[0020] First step: determine a name for the node to identify the node;

[0021] Second step: write a description text for the node to facilitate others to understand what the node does and the applicable scenario;

[0022] Third step: define the type and form of the input data and output data of the algorithm node according to the requirements;

[0023] Fourth step: determine the configuration items required by the algorithm node;

[0024] Fifth step: write the program running logic of the node in the form of code to realize the conversion of the input data of the node into output data according to certain algorithm logic.

[0025] Optionally, in step (2), the industry-university-research integration platform connects the code in the node with the code interpreter of the background service through a message queue, realizes the sharing of the running program of the node in the platform and the memory of the background service, so that the algorithm of the platform can use the running environment and data calculation capability of the background service; developing a node also uses the running environment of the background service in the form of code segment to establish data connection with the Internet of Vehicles through reading files or reading databases, and imports the data required for vehicle networking data analysis into the platform.

[0026] Optionally, in step (3), the algorithm node connection mode is as follows:

[0027] First step: find the algorithm node needed by the project to be developed through the way of new node type or reuse of existing node type;

[0028] Second step: the required node type is dragged into the current developed project in the form of visual drag, and the node configuration item is filled in, so as to complete the instantiation of the node, and make the node become an instance node of the current project;

[0029] Third step: the nodes with logical relationship are established in one-way connection relationship, and the mapping of output and input is defined in the connection relationship, so that the output data of the previous node is mapped into the input data of the next node, thereby realizing the connection of the two nodes.

[0030] Fourth step: repeat the third step until the connection between all nodes with logical relationship is completed.

[0031] Optionally, in step (4), the node code is connected with the code interpreter of the background service through the message queue, the data analysis platform can realize memory sharing with the background service, so that both are in the same running environment, so that the development and debugging environment of the data analysis process is the same as the deployment environment; with the help of memory sharing, the values of various variables in the code in the algorithm node can be monitored by the background service, and the code is debugged in the background.

[0032] The application provides a production-teaching-research integrated platform supporting visual drag for Internet of Vehicles data fusion analysis, which can integrate academic research, teaching platform and industrial application together to realize production-teaching-research integration. The platform can break through the whole link of data analysis and ensure efficient landing of data application. In the platform, a user can abstract and encapsulate data operation logic, which is convenient for others to reuse. Finally, a plurality of instantiated algorithm nodes are obtained by reusing existing algorithms and fine-tuning logic, and then the instantiated nodes are connected to establish a data flow connection relationship, so as to realize a data analysis process meeting the business demand. Meanwhile, the platform visualizes the data analysis process, so as to realize intuitive and easy-to-use effect for data analysis personnel and business personnel, thereby improving the efficiency of data analysis of the related personnel, and the application prospect is very wide. For business personnel, the problem that it is difficult to customize algorithm behavior and reuse existing components when using the data analysis system is solved. For scientific researchers, the problem that the algorithm developed by the researchers is difficult to be actually landed in the business system is solved. For teachers and students, the problem that the running environment is not matched and the algorithm logic is difficult to be intuitively displayed is solved. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0034] Figure 1 A data analysis flow example diagram developed by a production-teaching-research integration platform using the present application.

[0035] Figure 2 An overall architecture diagram of the platform of the present application;

[0036] Figure 3 An external behavior framework diagram of a node of the present application. DETAILED DESCRIPTION

[0037] In order to make the technical problems, technical solutions and beneficial effects of the present application clearer, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0038] Now a production-teaching-research integration platform supporting visual drag-and-drop vehicle networking data fusion analysis provided by an embodiment of the present application will be described. As shown in Figure 1 The production-teaching-research integration platform supporting visual drag-and-drop vehicle networking data fusion analysis encapsulates the algorithm logic of data analysis into reusable algorithm nodes, and interfaces the vehicle networking data sources into the system in the form of reusable nodes, and allows users to customize algorithm logic according to requirements and encapsulate multiple algorithm nodes, then creates a one-way connection relationship between the nodes after instantiating the nodes, forms a data analysis flow composed of multiple nodes, and finally obtains an analysis process that can intuitively show the data flow direction, and realizes rapid deployment and running in the cloud platform.

[0039] The production-teaching-research integration platform is composed of the following four core modules, and the architecture relationship thereof is shown in Figure 2 , which shows the relationship between different modules.

[0040] Node management module: can create new algorithm nodes or edit existing algorithm nodes, and manage all nodes. Each node needs to define its input data form and output data form, and then the node can calculate the input data according to the code logic in the node to obtain the output data. The running logic of a specific node can be realized by customizing the code that meets the formal requirements, which solves the problem of enabling data analysis process for artificial intelligence.

[0041] Data flow construction module: support to instantiate different types of nodes in the form of visual drag to a project, so that a node encapsulated algorithm logic can be conveniently called, solve the algorithm logic in the previous data analysis platform is difficult to call the problem. At the same time support and establish a one-way connection between the nodes, form a data analysis flow project, and can directly see the connection between the nodes and the dependent relationship.

[0042] Program running connection module: connect the code program encapsulated in the node with the code interpreter of the background service, so that the running of the node program shares the memory with the background service, realizing the integration of platform debugging, development and deployment.

[0043] Developer community module: support individual developers to publish the encapsulated nodes to the community, which can be downloaded and reused by others, which solves the problem that the components of the previous data analysis platform are difficult to reuse.

[0044] Through the above four core modules, the steps of using the production-teaching-research integrated platform of the application for data analysis project development include the following four steps:

[0045] Step (1): encapsulating data analysis logic algorithm as a node;

[0046] Step (2): develop a node for reading in Internet of Vehicles data;

[0047] Step (3): connect algorithm nodes to complete the development of data analysis flow;

[0048] Step (4): debug and run the data analysis flow.

[0049] Among them, in step (1), in the production-teaching-research integrated platform of the application, the running logic of one analysis stage of data analysis is realized as a code segment (such as a segment of Python code), and the code is encapsulated into a kind of node, and the data field type and format of the input data and the output data of the node are customized. Therefore, the node instance can perform calculation on the received input data through the pre-defined program logic of the code block during the running process, and the calculation result is transmitted outward through the customized output. At the same time, in order to realize the configuration of the variable part (such as the connection address of a certain database) inside the program logic of the code block, the encapsulated node can explicitly declare a plurality of configuration items, and when the node is reused, the user needs to fill in the configuration value for each configuration item in a separate operation interface after instantiating the node, so as to customize the value of some parameter variable of the code block during actual running according to the actual environment or actual needs. Therefore, the external behavior of a node is as shown in the following table. Figure 3 Program running connection module: connect the code program encapsulated in the node with the code interpreter of the background service, so that the running of the node program shares the memory with the background service, realizing the integration of platform debugging, development and deployment.

[0043] Developer community module: support individual developers to publish the encapsulated nodes to the community, which can be downloaded and reused by others, which solves the problem that the components of the previous data analysis platform are difficult to reuse.

[0044] Through the above four core modules, the steps of using the production-teaching-research integrated platform of the application for data analysis project development include the following four steps:

[0045] Step (1): encapsulating data analysis logic algorithm as a node;

[0046] Step (2): develop a node for reading in Internet of Vehicles data;

[0047] Step (3): connect algorithm nodes to complete the development of data analysis flow;

[0048] Step (4): debug and run the data analysis flow.

[0049] Among them, in step (1), in the production-teaching-research integrated platform of the application, the running logic of one analysis stage of data analysis is realized as a code segment (such as a segment of Python code), and the code is encapsulated into a kind of node, and the data field type and format of the input data and the output data of the node are customized. Therefore, the node instance can perform calculation on the received input data through the pre-defined program logic of the code block during the running process, and the calculation result is transmitted outward through the customized output. At the same time, in order to realize the configuration of the variable part (such as the connection address of a certain database) inside the program logic of the code block, the encapsulated node can explicitly declare a plurality of configuration items, and when the node is reused, the user needs to fill in the configuration value for each configuration item in a separate operation interface after instantiating the node, so as to customize the value of some parameter variable of the code block during actual running according to the actual environment or actual needs. Therefore, the external behavior of a node is as shown in the following table. Figure 3

[0050] Researchers can flexibly encapsulate the program logic they write into reusable nodes, and reuse them in the development of other data analysis flows by modifying the configuration items, defining the input and output, and even secondary development. The specific steps for encapsulating a node are as follows:

[0051] Step 1: Determine a name for the node to identify the node;

[0052] Step 2: Write a description text for the node to facilitate others to understand what the node does and the applicable scenarios;

[0053] Step 3: Define the types and forms of input data and output data of the algorithm node according to the requirements;

[0054] Step 4: Determine the configuration items required by the algorithm node (i.e. the field values that need to be determined according to the actual environment of the user);

[0055] Step 5: Write the program running logic of the node in the form of code to implement the conversion of the input data of the node into output data according to certain algorithm logic.

[0056] Through the above steps, a node that can run a certain algorithm logic can be obtained, and the data flow for data analysis of the Internet of Vehicles can be regarded as the algorithm logic calculation of the original data through multiple nodes. These logic calculations achieve the final calculation results required from the original data. In most cases, the development of data analysis process can be reused by others, and the completed data analysis logic can also be reused in other systems (such as SVM algorithm for sample classification in multiple systems). However, there is currently a lack of a platform that can visually reuse these developed program code fragments to complete a data analysis process. Therefore, the code fragments that can be reused in the data analysis process are encapsulated into a node, and the node is instantiated in the current analysis process in the form of visual dragging in the data analysis process that needs to reuse the code fragments, and combined with other nodes to complete the development of the entire data analysis process, so that these nodes can be reused to solve the problem of difficult reuse of code fragments in the past.

[0057] In step (2), the industry-university-research integration platform connects the code in the node with the code interpreter of the background service through a message queue, realizes the sharing of the running program of the node in the platform and the memory of the background service, so that the algorithm of the platform can use the running environment and data calculation capacity of the background service. A node is also developed in the form of code fragments to use the running environment of the background service to establish a data connection with the Internet of Vehicles through the form of reading files or reading databases, and to import the data required for data analysis of the Internet of Vehicles into the platform. This process can encapsulate an algorithm node specially used for reading a certain type of data according to the encapsulation method described in step (1), and then in a specific data analysis project, the platform user only needs to select the required data source node, instantiate this node into the current data analysis stream through the drag-and-drop method in the visual interface of the platform, and then modify a few configuration items of the node to realize the import of the Internet of Vehicles data.

[0058] In step (3), a complete data analysis stream is composed of multiple data analysis stages, and in the platform, multiple algorithm nodes are needed to combine to complete the analysis of the original data. After the Internet of Vehicles data is accessed according to step (2), further data analysis process is still needed for the source data, so the next step is for the data analysis personnel to create new algorithm nodes that are exclusive to the current system according to the method described in step (1), or to reuse existing algorithm nodes to develop the data analysis stream. The developer obtains multiple nodes encapsulating code fragments through new creation or reuse according to business requirements, and these nodes constitute a node library. The next step is to instantiate the nodes in the node library into the current project and establish connection relationships for the instance nodes. Because each node has its own input and output, establishing connection relationships for these nodes is to connect these individual instance nodes through the creation of a one-way connection relationship. For example, if A node and B node need to be connected, only a connection relationship from A to B needs to be created, and it is specified that which part of the output of the A node is mapped to which part of the input of the B node. In this way, when deployed and run, after the code of the A node is run, the output data will flow into the input of the B node according to the mapping specified in the connection relationship, and the B node will continue to process it. The data analysis process obtained through this node connection method has the characteristics of being intuitive and flexible, and business personnel can see the data flow and the macro process of data analysis from the connection relationship of the algorithm nodes, and can reuse existing nodes and integrate them into the current data analysis project through the drag-and-drop method. After the project development is completed and enters the maintenance stage, the data analysis personnel or scientific researchers can change the implementation method of the code inside the node flexibly without affecting the stable operation of the entire project without damaging the exposed input and output.

[0059] The node connection mode is as follows:

[0060] First step: find the algorithm nodes (such as SVM algorithm nodes, MySQL connection nodes, etc.) needed by the project to be developed by creating a new node type or reusing an existing node type;

[0061] Second step: drag the required node type into the current development project in a visual drag-and-drop manner, and fill in the node configuration items, thereby completing the instantiation of these nodes and making these nodes become instance nodes of the current project;

[0062] Third step: establish a one-way connection relationship between nodes with logical relationships, and define the mapping of output and input in this connection relationship, so that the output data of the previous node is mapped to the input data of the next node, thereby realizing the connection between the two nodes;

[0063] Fourth step: repeat the third step until the connection between all nodes with logical relationships is completed.

[0064] In step (4), the node code is connected to the code interpreter of the background service through the message queue, and the data analysis platform can share memory with the background service, so that they are in the same running environment (specifically, the same file system), thereby making the development and debugging environment of the data analysis flow the same as the deployment environment, improving the efficiency of product delivery. With the help of memory sharing, the values of variables in the code in the algorithm node can be monitored by the background service, so that the developer can debug the code in the background. This process reduces the cost of code debugging by keeping the consistency of memory space and running environment.

[0065] When the development of the entire data analysis flow is completed through debugging, the user can run the project. After the project is run, the first node (usually the node developed in step (2) to connect the data source) starts running, and the output data of the node running is mapped to the input of the next node through the node connection relationship, thereby starting the continuous running of the next node. All these node connections form a pipeline that may have a branch structure, and the above running process is repeated until the running of all nodes is completed, and the analysis result of the entire data analysis flow is obtained in the tail node.

[0066] The present application abstracts and encapsulates the code of algorithm logic into reusable nodes, and defines the input, output and configuration items of each algorithm node, and then the developer modifies the configuration items according to the actual environment when reusing a node, and can reuse the node in the construction of other data analysis flows through visual drag-and-drop;

[0067] By establishing a one-way connection relationship for the nodes encapsulating algorithm logic, and defining the mapping of output data and input, the flow of data between nodes is realized, the running process of the entire data analysis can be intuitively displayed, different personnel can be facilitated to be familiar with the business logic, and when the demand changes, the logic of part of the nodes can be quickly changed, thereby realizing agile development, solving the problems that it is difficult to reuse the developed code fragments and difficult to reflect the entire data analysis process in the previous development process of data fusion analysis.

[0068] The data analysis platform is connected with the code interpreter of the background service through the message queue to realize memory sharing of program running, integrates debugging development and product deployment, and improves the efficiency of product delivery.

[0069] The production-teaching-research integrated platform of the present application supports the whole process operation from vehicle networking data collection, data access, data development, data analysis, model establishment, model operation and data application, and abstracts and encapsulates the operation logic of different processes into algorithm nodes, establishes and displays the data analysis flow process in a visual manner on the platform, and supports one-key operation of the data flow, which is convenient for researchers to debug and modify the operation logic, and is convenient for business personnel to customize and reuse the operator nodes. In addition, the existing data sets or existing algorithms in the developer community matched with the platform can be conveniently introduced into the development platform of the user in the form of encapsulation into nodes, so that the business developers can reuse these nodes in multiple data analysis processes. In addition, the platform of the present application can serve scientific researchers in the research and landing of frontier algorithms, and serve teaching personnel in the intuitive and visual display of algorithms. Therefore, the use of the vehicle networking data analysis platform can greatly improve the efficiency of data fusion analysis of developers, business personnel, scientific researchers and teaching personnel.

[0070] The above only describes the preferred embodiments of the present application and does not limit the present application, and any modification, equivalent replacement and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. An industry-university-research integration platform supporting visual drag-and-drop for data fusion analysis of Internet of Vehicles, characterized in that: The node management module, the data flow construction module, the program running connection module and the developer community module are included. The node management module is used for creating a new algorithm node or editing an existing algorithm node, and managing all nodes. The data flow construction module is used for supporting instantiation of different types of nodes in a visual drag-and-drop form into a project, establishing a one-way connection relationship between the instance nodes, and forming a data analysis flow project. The program running connection module is used for connecting the code program encapsulated in the node with the code interpreter of the background service, realizing running of the node program and sharing of the memory of the background service, and realizing integration of debugging, development and deployment of the platform. The developer community module is used for supporting personal developers to publish the encapsulated nodes to the community, so that other people can download and reuse the nodes. The node management module can create a new algorithm node in the node library, and a platform user can encapsulate a piece of code into a node according to a specified form requirement, and define data types and formats of input data and output data of the node, and additionally define configuration items of the node, so that the platform user can modify parameter values in the configuration items according to actual project scenarios, and then instantiation of the node can convert the input data into output data through operation of the code and output the output data. The node management module can edit any existing algorithm node, modify and edit the algorithm node encapsulated by others according to specific requirements, convert the algorithm node to the node library of the current project, and use the algorithm node. The data flow construction module supports instantiation of different node types in the node library into the current project, and modification of configuration items of a node to make the node become an instantiated node, and then a one-way connection relationship can be established between two instantiated nodes, the relationship defines mapping between part of output data of a previous node and part of input data of a subsequent node, realizes data flow between the instantiated nodes, multiple nodes establish connection relationships with each other, and finally a data analysis flow project that can complete an entire data analysis task is formed. The program running connection module connects the code encapsulated in the instantiated node with the code interpreter of the background service through a message queue when the project is running, so that running of the node program and sharing of the memory of the background service are realized, and when the node is running, specific values of each variable can be monitored by the background service, the running state of each node in the project can be viewed and debugged in the background service, and integration of debugging, development and deployment of the platform is realized.

2. The industry-university-research integration platform supporting visual drag-and-drop for telematics data fusion analysis according to claim 1, wherein: The steps of the industry-university-research integration platform for developing a data analysis project include: Step (1): encapsulating a data analysis logic algorithm as a node; Step (2): developing a node for reading in Internet of Vehicles data; Step (3): connecting algorithm nodes to complete development of a data analysis flow; Step (4): debugging and running the data analysis flow.

3. The industry-university-research integration platform supporting visual drag-and-drop for telematics data fusion analysis according to claim 2, characterized in that: In step (1), in the production-study-research integration platform, the running logic of one analysis stage of data analysis is implemented as a code segment, and the code is encapsulated into a node, and the data field type and format of the input data and output data of the node are customized. The node instance receives the input data and performs calculation through the pre-defined program logic of the code block, and transmits the calculation result to the outside through the customized output. Meanwhile, the variable part inside the program logic of the code block is configured; The self-written program logic is encapsulated into a reusable node, and is reused in the development of other data analysis flows by modifying the configuration items, defining the input and output, and even developing secondary. The specific operation steps of encapsulating a node are as follows: First step: determine a name for the node to identify the node; Second step: write a description text for the node to facilitate others to understand what the node does and the applicable scenario; Third step: define the type and form of the input data and output data of the algorithm node according to the requirements; Fourth step: determine the configuration items required by the algorithm node; Fifth step: write the program running logic of the node in the form of code to realize the conversion of the input data of the node into output data according to certain algorithm logic.

4. The industry-university-research integration platform supporting visual drag-and-drop for telematics data fusion analysis of claim 2, wherein: In step (2), the production-study-research integration platform connects the code in the node with the code interpreter of the background service through a message queue, realizes the sharing of the running program of the node and the memory of the background service, and enables the platform algorithm to use the running environment and data calculation capability of the background service. A node also uses the running environment of the background service in the form of code segment to establish data connection with the Internet of Vehicles by reading files or databases, and imports the data required for vehicle Internet data analysis into the platform.

5. The industry-university-research integration platform supporting visual drag-and-drop for telematics data fusion analysis of claim 2, wherein: In step (3), the algorithm node connection method is as follows: First step: find the algorithm node required by the project to be developed by creating a new node type or reusing an existing node type; Second step: drag the required node type to the current development project in a visual drag-and-drop manner, and fill in the node configuration items, thereby completing the instantiation of these nodes and making these nodes become instance nodes of the current project; Third step: establish a one-way connection relationship between nodes with logical relationship, and define the mapping of output and input in the connection relationship, so that the output data of the previous node is mapped to the input data of the next node, thereby realizing the connection between two nodes; Fourth step: repeat the third step until the connection between all nodes with logical relationship is completed.

6. The industry-university-research integration platform supporting visual drag-and-drop for telematics data fusion analysis of claim 2, wherein: In step (4), the node code is connected with the code interpreter of the background service through the message queue, the data analysis platform can share the memory with the background service, so that they are in the same running environment, thereby making the development and debugging environment of the data analysis process the same as the deployment environment. With the help of memory sharing, the values of variables in the algorithm node code can be monitored by the background service during running, and the code is debugged in the background.

Images