Automobile universal part automatic test system based on structured analysis and semantic analysis

Abstract

The invention discloses an automatic test system for automobile universal parts based on structured analysis and semantic analysis, and relates to the technical field of automatic test systems for automobile universal parts. The automatic test system comprises an automatic test system, and the receiving end of the automatic test system is in signal connection with the transmitting end of a case import unit; the receiving end of the automatic test system is in signal connection with the transmitting end of the natural language processing unit. By arranging a natural language processing unit, an automatic conversion module, a semantic analysis engine module, a case training model module, a manual calibration module and a test database, the natural language test case of the instrument panel is analyzed and compiled by using the semantic analysis engine module in combination with the case training model module to generate an automatic test case; the test process is simplified, the test efficiency is effectively improved, the repeated test work is reduced, the personal error is reduced, the test accuracy and the use case coverage are improved, the verification test time of the electronic control unit is shortened, the repeated work is reduced, and the test efficiency is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic testing systems for automobile universal parts, and in particular to an automatic testing system for automobile universal parts based on structured parsing and semantic analysis. Background Art

[0002] With the evolution of automotive electronic and electrical architecture and the rapid development of intelligent and networked technologies, the functional complexity of automotive instrument panels, as core components of human-computer interaction, has increased significantly. Modern instrument panels not only need to display basic information such as vehicle speed, RPM, and fuel level, but also integrate data from multiple functional modules such as ADAS status, navigation, multimedia, and vehicle diagnostics. They also need to support personalized interfaces and OTA (Over-the-Air) remote upgrades, placing higher demands on the development and testing of instrument panels.

[0003] Traditional instrument panel functional testing methods mainly rely on manual operation of hardware testing equipment, which has the following limitations: due to the excessive number of integrated functions in the instrument panel, the labor cost of functional testing is high, and the thousands of test cases are complex and numerous, which cannot meet the needs of rapid iteration; the adaptation cost is high, and the interface protocols and driver layers of different hardware platforms are significantly different. Test scripts and environments need to be customized for each platform, with a long development cycle, low reusability, and insufficient test coverage. Manual testing is difficult to cover complex scenarios (such as multi-tasking concurrency and extreme working condition simulation), and cross-platform compatibility cannot be efficiently verified; agile development needs do not match, the iteration speed of smart cars is accelerated, and traditional testing processes are difficult to support continuous integration and continuous testing. Summary of the Invention

[0004] The present invention provides an automatic testing system for automobile universal parts based on structured parsing and semantic analysis to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: An automatic test system for universal automobile parts based on structured parsing and semantic analysis includes an automatic test system, wherein a receiving end of the automatic test system is signal-connected to a transmitting end of a use case import unit, a transmitting end of the automatic test system is signal-connected to a transmitting end of a natural language processing unit, a transmitting end of the automatic test system is signal-connected to a receiving end of a software test unit, a transmitting end of the use case import unit is signal-connected to a receiving end of a natural language processing unit, and a transmitting end of the automatic test system is signal-connected to a receiving end of a open display unit.

[0006] A further improvement of the technical solution of the present invention is that: the natural language processing unit includes an automatic conversion module, a manual calibration module and a test database, the automatic conversion module includes a semantic analysis engine module and a use case training model module, the natural language processing unit receiving end is signal-connected with the automatic conversion module transmitting end, the manual calibration module transmitting end and the test database transmitting end, and the automatic conversion module uses the semantic analysis engine module to process natural language processing technology to automatically convert the test requirements described in text into executable test logic.

[0007] A further improvement of the technical solution of the present invention is that: the semantic analysis engine module includes a structured parsing module, a self-use test case historical data set, a rule engine module and an artificial use case schematic module; the self-use test case historical data set, the rule engine module and the artificial use case schematic module perform data training; the manual calibration module supports engineers to correct the test steps generated after data training; the corrected data will be fed back to the natural language processing unit for continued training, and the data will be imported into the test database to store the calibrated use case conversion results for the natural language processing unit to continuously learn and optimize data; the structured parsing module includes a structured label module and a structured parameter module, which are used to summarize natural language in the form of labels + parameters.

[0008] A further improvement of the technical solution of the present invention is that the software testing unit transmitting end is signal-connected to the open display unit receiving end, and the open display unit is used to open the automated test UI operation interface to the user, and is used to display the test progress, hardware status, and signal waveform key information in real time.

[0009] A further improvement of the technical solution of the present invention is that: the transmitting end of the automatic test system is signal-connected to the receiving end of the communication interaction module, the transmitting end of the communication interaction module is signal-connected to the receiving end of the software testing unit, the software testing unit includes a relay module, a programmable power supply control module, a host computer module, an industrial computer module, a message module, a report module and a test engineering module, the relay module is used to switch the circuit to realize the automatic construction of the test circuit, the programmable power supply control module is used to control the voltage output of each channel of the power supply, the host computer module is used to control the switching of the test circuit, the industrial computer module test platform PC end, runs the test software, communicates with each instrument, is used to record and analyze bus messages, and provide test work.

[0010] A further improvement of the technical solution of the present invention is that: the software testing unit is used to drive the underlying function of the standardized use case generated by the use case conversion model to generate a test project, the message module is used to capture and analyze bus messages, and the report module is used to save test results and output test reports.

[0011] A further improvement of the technical solution of the present invention is that: the transmitting end of the industrial control computer module is signal-connected to the receiving end of the communication interaction module, the transmitting end of the communication interaction module is signal-connected to the receiving end of the host computer module, and the communication interaction module is used to realize data interaction between the underlying test software and the host computer.

[0012] A further improvement of the technical solution of the present invention is that the open display unit includes a test report module and a data comparison module. The test report module is used to display the test progress, hardware status, and key information of the signal waveform in real time for engineers to debug and observe the test data. The data comparison module is used to provide a test data comparison view and automatically mark abnormal detection results in red. The structured parsing module specifically analyzes the following steps: Step 1: Written in Python, Python analyzes the semantics of artificial use cases and generates CAPL test scripts. A hybrid architecture of rule-based agents and lightweight NLP models is employed. The rule-based agent approach directly leverages existing keyword mapping rules ("SysBPM = 0 / 1 / 2 / 3 / 4" → parameterized testing) to ensure the accuracy of core logic. The lightweight NLP model is used to parse entities and actions in unstructured text, identify entities for which rules are not clearly defined, and build a rule-based AI agent based on the specified rule engine module. Step 2: Automatically parse and compile natural language test cases into executable scripts, abstracting the natural language into the form of labels + parameters; Step 3: Python identifies and parses the use case labels. Based on the label rules and the vector tool chain Canoe and VTestStudio, it is classified as a Capl function module function (1-n). The test case content of the test dashboard is: Send signal: ipkconfig=1 (parameter). The semantic analysis engine module (8) parses this sentence as [send_signal] ipkconfig=1 (label + parameter). It continues to identify the labels and parameters processed by the semantic analysis, which are reflected in the mapping script as send_signal (ipkconfig, 1). Step 4: Finally automate the test script content and generate the corresponding project signal mapping table.

[0013] Due to the adoption of the above technical solution, the present invention has the following technical advancements compared to the prior art: The present invention provides an automatic test system for general automobile parts based on structured parsing and semantic analysis. By setting a natural language processing unit, an automatic conversion module, a semantic analysis engine module, a use case training model module, a manual calibration module and a test database, the semantic analysis engine module is used in combination with the use case training model module to parse the natural language test cases of the instrument panel, compile and generate automated test cases, simplify the test process, effectively improve the test efficiency, reduce repetitive test work, reduce human errors, improve test accuracy and use case coverage, and shorten the verification test time of the electronic control unit. The natural language processing unit uses the semantic analysis engine module and the use case training model module to convert text into The described test requirements are automatically converted into executable test logic, with a built-in manual calibration interface to support engineers in correcting the test steps generated after data training. The corrected data will be fed back to the natural language processing unit for further training, and its own test database will be generated to store the calibrated use case conversion results for the model to continuously learn and optimize data. Natural language test cases are automatically parsed and compiled into executable scripts, which greatly reduces the time for manual testing, reduces repetitive work, and significantly improves testing efficiency. It can support agile development and continuous integration requirements, that is, you only need to upload the use case, and the test case-driven test automation workflow can be achieved in the CI / CD continuous integration test system.

[0014] The present invention provides an automatic test system for automobile universal parts based on structured parsing and semantic analysis. By setting a software test unit, a relay module, a programmable power supply control module, a host computer module, an industrial computer module, a message module, a report module and a test engineering module, the software test unit mainly includes a relay switching circuit, a programmable power supply control voltage output, and a host computer control switching of the test circuit. The system hardware completes the bottom-level bench design of the automatic test system, and the software level needs to complete communication interaction, script bottom-level function design, use case analysis and hardware bench testing. The standardized use case generated by the use case conversion model drives the bottom-level function to generate a test project.

[0015] The present invention provides an automatic testing system for universal automotive parts based on structured parsing and semantic analysis. By setting an open display unit, a test report module and a data comparison module, an automated test UI operation interface is provided to users, and key information such as test progress, hardware status and signal waveform are displayed in real time for engineers to debug and observe test data. A test data comparison view is provided, abnormal detection results are automatically marked in red, and a visual UI interface is provided to display test progress, hardware status and signal waveform in real time, supporting engineers to quickly debug and observe data, lowering the operating threshold, eliminating the need for engineers to send individual messages and then observe the message signals, and significantly shortening the verification cycle of the electronic control unit. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1This is a connection diagram of the automatic test system of the present invention; Figure 2 This is a connection diagram of the natural language processing unit of the present invention; Figure 3 This is a connection diagram of the semantic analysis engine module of the present invention; Figure 4 This is a schematic diagram of the connection of the software testing engineering unit of the present invention; Figure 5 This is a schematic diagram of the connection of the open display unit of the present invention; Figure 6 This is a schematic diagram of the connection of the structured parsing module of the present invention.

[0017] In the figure: 1. Automatic test system; 2. Use case import unit; 3. Natural language processing unit; 4. Software testing unit; 5. Communication interaction module; 6. Open display unit; 7. Automatic conversion module; 8. Semantic analysis engine module; 9. Use case training model module; 10. Manual calibration module; 11. Test database; 12. Self-use test case historical data set; 13. Rule engine module; 14. Manual use case schematic module; 15. Relay module; 16. Programmable power supply control module; 17. Host computer module; 19. Industrial computer module; 20. Message module; 21. Report module; 22. Test engineering module; 23. Test report module; 24. Data comparison module, 25 Structured parsing module, 26 Structured labeling module, 27 Structured parameter module. DETAILED DESCRIPTION

[0018] The present invention is described in further detail below in conjunction with the embodiments: like Figure 1-6 As shown, the present invention provides an automatic test system for automobile universal parts based on structured parsing and semantic analysis, including an automatic test system 1, a receiving end of the automatic test system 1 is signal-connected to a transmitting end of a use case import unit 2, a receiving end of the automatic test system 1 is signal-connected to a transmitting end of a natural language processing unit 3, a transmitting end of the automatic test system 1 is signal-connected to a receiving end of a software test unit 4, a transmitting end of the use case import unit 2 is signal-connected to a receiving end of the natural language processing unit 3, and a transmitting end of the automatic test system 1 is signal-connected to a receiving end of a open display unit 6; The natural language processing unit 3 includes an automatic conversion module 7, a manual calibration module 10 and a test database 11. The automatic conversion module 7 includes a semantic analysis engine module 8 and a use case training model module 9. The semantic analysis engine module 8 includes a structured parsing module 25, a self-use test case history data set 12, a rule engine module 13 and a manual use case schematic module 14. The self-use test case history data set 12, the rule engine module 13 and the manual use case schematic module 14 perform data training. The manual calibration module 10 supports engineers to correct the test steps generated after data training. The corrected data will be fed back to the natural language processing unit 3 for further training, and the data will be imported into the test database 11 to store the calibrated use case conversion results for the natural language processing unit 3 to continuously learn and optimize data. The structured parsing module 25 includes a structured label module 26 and a structured parameter module 27, which are used to summarize natural language into the form of labels + parameters; The transmitting end of the software testing unit 4 is connected to the receiving end of the open display unit 6. The open display unit 6 is used to open the automated test UI operation interface to the user and display the test progress, hardware status, signal waveform and other key information in real time. The receiving end of the natural language processing unit 3 is connected to the transmitting end of the automatic conversion module 7, the transmitting end of the manual calibration module 10 and the transmitting end of the test database 11. The automatic conversion module 7 uses the semantic analysis engine module 8 to process the natural language processing technology, and automatically converts the test requirements described in text, such as "simulating abnormal vehicle speed signal", into executable test logic. By setting the natural language processing unit 3, the automatic conversion module 7, the semantic analysis engine module 8, the use case training model module 9, the manual calibration module 10 and the test database 11, the semantic analysis engine module 8 is used in combination with the use case training model module 9 to parse the natural language test cases of the instrument panel, compile and generate automated test cases, simplify the test process, effectively improve the test efficiency, reduce repetitive testing work, reduce human errors, improve test accuracy and use case coverage, and shorten the electronic control unit. To reduce the verification test time, the natural language processing unit 3 uses the semantic analysis engine module 8 and the use case training model module 9 to automatically convert text-described test requirements such as "simulating abnormal vehicle speed signals" into executable test logic. It has a built-in manual calibration interface to support engineers in correcting the test steps generated after data training. The corrected data will be fed back to the natural language processing unit 3 for continued training, and its own test database will be generated to store the calibrated use case conversion results for the model to continuously learn and optimize data. It automatically parses and compiles natural language test cases into executable scripts, greatly reducing the time for manual testing, reducing repetitive work, and significantly improving test efficiency. It can support agile development and continuous integration requirements, that is, only need to upload the use case, and the test case-driven test automation workflow can be achieved in the CI / CD continuous integration test system; The transmitting end of the automatic test system 1 is connected to the receiving end of the communication interaction module 5 for signal connection, and the transmitting end of the communication interaction module 5 is connected to the receiving end of the software test unit 4 for signal connection. The software test unit 4 includes a relay module 15, a programmable power supply control module 16, a host computer module 17, an industrial computer module 19, a message module 20, a report module 21 and a test engineering module 22. The relay module 15 is used to switch the circuit to realize the automatic construction of the test circuit. The programmable power supply control module 16 is used to control the voltage output of each channel of the power supply. The host computer module 17 is used to control the switching of the test circuit. The industrial computer module 19 is the PC end of the test platform, runs the test software, communicates with each instrument, is used to record and analyze bus messages, and provides test work; The transmitting end of the industrial control computer module 19 is connected to the receiving end of the communication interaction module 5, and the transmitting end of the communication interaction module 5 is connected to the receiving end of the host computer module 17. The communication interaction module 5 is used to realize data interaction between the bottom test software and the host computer; The software testing unit 4 is used to drive the underlying functions of the standardized use cases generated by the use case conversion model to generate a test project. The message module 20 is used to capture and analyze bus messages. The report module 21 is used to save test results and output a test report. By setting the software testing unit 4, the relay module 15, the programmable power supply control module 16, the host computer module 17, the industrial computer module 19, the message module 20, the report module 21 and the test project module 22, the software testing unit 4 mainly includes the relay switching circuit, the programmable power supply control voltage output, and the host computer control switching of the test circuit. The system hardware completes the underlying bench design of the automatic test system 1. The software level needs to complete communication interaction, script underlying function design, use case analysis and hardware bench testing. The standardized use case generated by the use case conversion model is used to drive the underlying function to generate a test project. The open display unit 6 includes a test report module 23 and a data comparison module 24. The test report module 23 is used to display the test progress, hardware status, and key information of the signal waveform in real time for engineers to debug and observe the test data. The data comparison module 24 is used to provide a test data comparison view and automatically mark the abnormal detection results in red. By setting the open display unit 6, the test report module 23 and the data comparison module 24, an automated test UI operation interface is provided to the user, and the test progress, hardware status, and key information of the signal waveform are displayed in real time for engineers to debug and observe the test data. A test data comparison view is provided, and the abnormal detection results are automatically marked in red. A visual UI interface is provided to display the test progress, hardware status, and signal waveform in real time, supporting engineers to quickly debug and observe data, lowering the operation threshold, and eliminating the need for engineers to send individual messages separately and then observe the message signals, thereby significantly shortening the verification cycle of the electronic control unit. The structured analysis module 25 specifically analyzes the following steps: Step 1: Written in Python, Python analyzes the semantics of artificial use cases and generates CAPL test scripts. A hybrid architecture of rule-based agents and lightweight NLP models is employed. The rule-based agent approach directly leverages existing keyword mapping rules ("SysBPM = 0 / 1 / 2 / 3 / 4" → parameterized testing) to ensure the accuracy of core logic. The lightweight NLP model is used to parse entities and actions in unstructured text, identify entities for which rules are not clearly defined, and build a rule-based AI agent based on the specified rule engine module13. Step 2: Automatically parse and compile natural language test cases into executable scripts, abstracting the natural language into the form of labels + parameters; Step 3: Python identifies and parses the use case labels. Based on the label rules and the Vector toolchain Canoe and VTestStudio, it is classified as a CAPL functional module function (1-n). The test case content of the test dashboard is: Send signal: ipkconfig=1 (parameter). The semantic analysis engine module 8 parses this sentence as [send_signal]ipkconfig=1 (label + parameter). It continues to identify the labels and parameters processed by the semantic analysis, and reflects it in the mapping script as send_signal(ipkconfig, 1). Step 4: Finally automate the test script content and generate the corresponding project signal mapping table.

[0019] Structured parsing test case: 1. Natural language test content (original test case): KL15 ON The voltage is normal and the CCP_CFCAN_FrP00 (0x169) message node signals are all valid. Default configuration 0xC111 insUnitsPrm=0; 2. Content processed by the semantic analysis model: [def_msg] BMSChrgStsIO;IMCUVehElecRngV;IMCUVehElecRng [send_msg_with_data] CCP_CFCAN_FrP00,FF,FF,FF,FF,FF,FF,FF,FF,20 [set_case_header] 101-080-010-001, Endurance display function test and KL15 ON, check whether the endurance display function is normal when KL15 is ON [set_io] KL30=12V;KL15=12V [change_config] insUnitsPrm=0; 3. Post-mapping script content (final script execution test content): def_msg(BMSChrgStsIO,IMCUVehElecRngV,IMCUVehElecRng) send_msg_with_data(CCP_CFCAN_FrP00,FF,FF,FF,FF,FF,FF,FF,FF,20) set_case_header("101-080-010-001, Endurance display function test and KL15 ON, check whether the endurance display function is normal when KL15 is ON") set_io(12,12) change_config(insUnitsPrm,0).

[0020] The above generally describes the present invention in detail. However, it is obvious to those skilled in the art that modifications or improvements may be made thereto. Therefore, modifications or improvements that do not depart from the spirit of the present invention are within the scope of protection of the present invention.

Claims

1. An automatic testing system for general automotive parts based on structured parsing and semantic analysis, comprising an automatic testing system (1), characterized in that: The receiving end of the automatic test system (1) is connected to the transmitting end of the use case import unit (2) by signal, the receiving end of the automatic test system (1) is connected to the transmitting end of the natural language processing unit (3) by signal, the transmitting end of the automatic test system (1) is connected to the receiving end of the software test unit (4) by signal, the transmitting end of the use case import unit (2) is connected to the receiving end of the natural language processing unit (3) by signal, and the transmitting end of the automatic test system (1) is connected to the receiving end of the open display unit (6) by signal.

2. The automatic testing system for universal automotive parts based on structured parsing and semantic analysis according to claim 1 is characterized in that: The natural language processing unit (3) includes an automatic conversion module (7), a manual calibration module (10) and a test database (11); the automatic conversion module (7) includes a semantic analysis engine module (8) and a use case training model module (9); the natural language processing unit (3) receiving end is signal-connected to the automatic conversion module (7) transmitting end, the manual calibration module (10) transmitting end and the test database (11) transmitting end; the automatic conversion module (7) utilizes the semantic analysis engine module (8) to process natural language processing technology and automatically converts the text-described test requirements into executable test logic.

3. The automatic testing system for universal automobile parts based on structured parsing and semantic analysis according to claim 2 is characterized in that: The semantic analysis engine module (8) includes a structured parsing module (25), a self-use test case history data set (12), a rule engine module (13) and an artificial use case schematic module (14). The self-use test case history data set (12), the rule engine module (13) and the artificial use case schematic module (14) perform data training. The artificial calibration module (10) supports engineers to correct the test steps generated after data training. The corrected data will be fed back to the natural language processing unit (3) for continued training, and the data will be imported into the test database (11) to store the calibrated use case conversion results for the natural language processing unit (3) to continuously learn and optimize data. The structured parsing module (25) includes a structured label module (26) and a structured parameter module (27) for summarizing natural language into the form of label + parameter.

4. The automatic testing system for universal automotive parts based on structured parsing and semantic analysis according to claim 1 is characterized in that: The software test unit (4) transmitting end is signal-connected to the open display unit (6) receiving end. The open display unit (6) is used to open an automated test UI operation interface to the user, and is used to display key information such as test progress, hardware status, and signal waveform in real time.

5. The automatic testing system for universal automobile parts based on structured parsing and semantic analysis according to claim 1 is characterized in that: The transmitting end of the automatic test system (1) is connected to the receiving end of the communication interaction module (5) by signal connection, and the transmitting end of the communication interaction module (5) is connected to the receiving end of the software test unit (4) by signal connection. The software test unit (4) includes a relay module (15), a programmable power supply control module (16), a host computer module (17), an industrial computer module (19), a message module (20), a report module (21) and a test engineering module (22). The relay module (15) is used to switch the circuit to realize the automatic construction of the test circuit. The programmable power supply control module (16) is used to control the voltage output of each channel of the power supply. The host computer module (17) is used to control the switching of the test circuit. The industrial computer module (19) is a test platform PC end, runs the test software, communicates with each instrument, and is used to record and analyze bus messages to provide test work.

6. The automatic testing system for universal automobile parts based on structured parsing and semantic analysis according to claim 5 is characterized in that: The software testing unit (4) is used to drive the underlying function with the standardized use case generated by the use case conversion model to generate a test project. The message module (20) is used to capture and analyze bus messages. The report module (21) is used to save the test results and output a test report.

7. The automatic testing system for universal automobile parts based on structured parsing and semantic analysis according to claim 5 is characterized in that: The transmitting end of the industrial control computer module (19) is connected to the receiving end of the communication interaction module (5) by signal connection, and the transmitting end of the communication interaction module (5) is connected to the receiving end of the host computer module (17) by signal connection. The communication interaction module (5) is used to realize data interaction between the bottom test software and the host computer.

8. The automatic testing system for universal automobile parts based on structured parsing and semantic analysis according to claim 1 is characterized in that: The open display unit (6) includes a test report module (23) and a data comparison module (24). The test report module (23) is used to display the test progress, hardware status, and key information of the signal waveform in real time for engineers to debug and observe the test data. The data comparison module (24) is used to provide a test data comparison view and automatically mark abnormal detection results in red.

9. The automatic testing system for universal automobile parts based on structured parsing and semantic analysis according to claim 3 is characterized in that: The structured analysis module (25) specifically analyzes the following steps: Step 1: Use Python programming language to write, Python analyzes the semantics of artificial use cases, generates capl test scripts, and adopts a hybrid architecture of rule-based agent + lightweight NLP model. The rule-based agent directly uses the existing keyword mapping rules ("SysBPM=0 / 1 / 2 / 3 / 4" → parameterized test) to ensure the accuracy of the core logic. The lightweight NLP model is used to parse entities and actions in unstructured text and identify entities without clear rules. Based on the specified rule engine module (13), a rule-based agent (AI Agent) is constructed. Step 2: Automatically parse and compile natural language test cases into executable scripts, abstracting the natural language into the form of labels + parameters; Step 3: Python identifies and parses the use case labels. Based on the label rules and the vector tool chain Canoe and VTestStudio, it is classified as a Capl function module function (1-n). The test case content of the test dashboard is: Send signal: ipkconfig=1 (parameter). The semantic analysis engine module (8) parses this sentence as [send_signal] ipkconfig=1 (label + parameter). It continues to identify the labels and parameters processed by the semantic analysis, which are reflected in the mapping script as send_signal (ipkconfig, 1). Step 4: Finally automate the test script content and generate the corresponding project signal mapping table.

Images