Intelligent cabin AI voice interaction test method, computer equipment and storage medium

Abstract

The invention discloses an intelligent cabin AI voice interaction test method, computer equipment and a storage medium. The test method comprises the following steps: establishing an intelligent test database based on function analysis of the intelligent cabin AI voice interaction system; the intelligent test database stores an interaction test case, the interaction test case comprises a plurality of generalization corpus texts, and the generalization corpus texts are endowed with difficulty coefficient values; during testing, in response to an externally input test difficulty coefficient instruction, the intelligent test database intelligently calculates and screens a generalization corpus text selected according with the test difficulty coefficient instruction based on a preset difficulty coefficient value of the generalization corpus text, replaces the generalization corpus text with voice output, and performs an interaction test; and calling the AI intelligent model to judge an interaction test result, and giving a score to the intelligent cabin AI voice interaction system. According to the invention, the intelligent test of the intelligent cabin AI voice interaction system is realized, the test intelligence and the test efficiency are greatly improved, and the consistency and objectivity of a test evaluation system can be ensured.

Description

Technical Field

[0001] This invention relates to the field of vehicle testing technology, and in particular to an AI voice interaction testing method for intelligent cockpits, a computer device, and a storage medium. Background Technology

[0002] With the rapid development of intelligent automotive cockpits, human-vehicle interaction has entered a new stage. In particular, AI voice interaction systems for intelligent cockpits have developed rapidly, gradually becoming an important carrier of cockpit user experience and a core battleground for product competition. AI voice interaction in intelligent cockpits is a necessary means to create advanced human-computer interaction methods. A related problem is how to conduct effectiveness testing on AI voice interaction systems to evaluate their performance and provide users with more objective test evaluations. This is a problem that must be solved.

[0003] Currently, the testing methods for AI voice interaction in smart cockpits are mainly divided into subjective testing and objective testing. Subjective testing includes the most basic and internationally recognized MOS scoring method, which evaluates the product through the most intuitive user experience and terminal usage effects, bypassing the problem of translating consumer language into engineering language. Objective testing avoids the influence of environmental and human factors on experimental results through standardized and automated testing.

[0004] However, for intelligent cockpit AI voice interaction systems, current subjective and objective testing techniques cannot comprehensively and realistically reflect the product performance and user experience of intelligent cockpit AI voice interaction systems in order to effectively test and evaluate their performance. Therefore, developing a testing method for intelligent cockpit AI voice interaction systems is of great significance. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings and defects of the prior art and to provide an AI voice interaction testing method, computer equipment, and storage medium for intelligent cockpits.

[0006] One objective of this invention is to provide a testing method for AI voice interaction in an intelligent cockpit, comprising the following steps:

[0007] Based on the analysis of the functions of the intelligent cockpit AI voice interaction system, an intelligent test database is established. The intelligent test database stores multiple different interactive test cases. Each interactive test case is used to test an interactive scenario of the intelligent cockpit AI voice interaction system. Each interactive test case contains an interactive test scenario. Each interactive test scenario includes multiple interactive tasks. Each interactive task includes multiple interactive test tasks. Each interactive test task contains multiple source corpus texts. The multiple source corpus texts are generalized to form multiple generalized corpus texts for testing. Each generalized corpus text is assigned a difficulty coefficient value.

[0008] When testing the AI ​​voice interaction system of the intelligent cockpit, in response to the test difficulty coefficient command input by the external input, the intelligent test database intelligently calculates and filters out the generalized corpus text that meets the test difficulty coefficient command based on the preset difficulty coefficient value of the generalized corpus text, replaces it with voice output, and conducts interactive testing on the AI ​​voice interaction system of the intelligent cockpit.

[0009] The system collects interaction test results, and the intelligent test database calls the AI ​​intelligent model to judge the interaction test results. Based on the judgment results, it gives a score to the intelligent cockpit AI voice interaction system.

[0010] When the intelligent test database calls the AI ​​intelligent model to judge the interaction test results, it judges whether the feedback results of the intelligent cockpit AI voice interaction system are accurate based on the positive and negative correlation of the interaction results.

[0011] The system will be rated based on the interaction, including an overall rating for the AI ​​voice interaction system and / or a scenario rating for each interaction scenario.

[0012] The generalized corpus text is generated by a pre-set corpus generalization system. The source corpus text is processed with colloquial expressions from different regions. Key information is extracted from the colloquial expressions and converted into keywords. The keywords are then generalized according to pre-set rules. The keywords are obtained offline or online by the corpus generalization system by connecting to a keyword database, or obtained online from relevant websites / platforms. The keyword database includes place name database, scenic spot name database, celebrity information database, cuisine name database, historical allusion database, music database, and encyclopedic knowledge database.

[0013] When testing the AI ​​voice interaction system for the intelligent cockpit, the intelligent test database intelligently selects generalized language text from the test cases according to the test difficulty coefficient instruction, using the following method:

[0014] Each test covers all interactive test scenarios, all interactive tasks under each interactive test scenario, and all interactive test tasks under each interactive task. The number of source corpus texts under each interactive test task remains consistent, and the number of corpus interaction rounds is consistent. The difficulty coefficient value of the generalized corpus text is calculated to ensure that the total difficulty coefficient of the selected generalized corpus text matches the test difficulty coefficient.

[0015] The AI ​​intelligent model is multiple. Based on the attributes of the interaction test results, the intelligent test database calls different AI intelligent models to judge the interaction test results, and judges the interaction test results by matching the AI ​​intelligent model.

[0016] For the same interaction test result, the intelligent test database calls at least two different AI intelligent models to judge the same interaction test result, cross-validate, and finally give the positive and negative correlation evaluation result.

[0017] The intelligent test database is based on the preset difficulty coefficient value of generalized corpus text. It intelligently calculates and filters out generalized corpus text that meets the test difficulty coefficient instruction. When replaced with voice output, it outputs dialect voices or standard Mandarin voices corresponding to different regions to conduct interactive tests on the intelligent cockpit AI voice interaction system.

[0018] A second objective of this invention is to provide a computer device, including a processor and a memory, wherein the memory stores a computer program for execution by the processor, and the computer program, when executed by the processor, implements the intelligent cockpit AI voice interaction testing method.

[0019] A third objective of this invention is to provide a storage medium storing a computer program, which, when executed by a processor, implements the intelligent cockpit AI voice interaction testing method.

[0020] The intelligent cockpit AI voice interaction testing method of the present invention, through an intelligent test database, responds to externally input test difficulty coefficient instructions, and intelligently calculates and filters generalized text that meets the test difficulty coefficient instructions based on preset generalized text difficulty values, replaces it with voice output, performs interactive testing on the intelligent cockpit AI voice interaction system, calls an AI intelligent model to judge the interactive test results, and gives a score to the intelligent cockpit AI voice interaction system based on the interaction situation. It can realize the intelligentness testing of the intelligent cockpit AI voice interaction system, greatly improve the intelligence and efficiency of the test, and ensure the consistency and objectivity of the test evaluation system. Attached Figure Description

[0021] Figure 1This is a flowchart of the intelligent cockpit AI voice interaction testing method of the present invention. Detailed Implementation

[0022] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0023] See Figure 1 As shown in the exemplary embodiment of this application, the intelligent cockpit AI voice interaction testing method includes the following steps:

[0024] S1. Based on the analysis of the functions of the intelligent cockpit AI voice interaction system, an intelligent test database is established. The intelligent test database stores multiple different interactive test cases. Each interactive test case is used to test an interactive scenario of the intelligent cockpit AI voice interaction system. Each interactive test case contains an interactive test scenario. Each interactive test scenario includes multiple interactive tasks. Each interactive task includes multiple interactive test tasks. Each interactive test task contains multiple source corpus texts. The multiple source corpus texts are generalized to form multiple generalized corpus texts for testing. Each generalized corpus text is assigned a difficulty coefficient value.

[0025] S2. When testing the intelligent cockpit AI voice interaction system, in response to the external input test difficulty coefficient instruction, the intelligent test database intelligently calculates and filters out the generalized corpus text that meets the test difficulty coefficient instruction based on the preset generalized corpus text difficulty coefficient value, replaces it with voice output, and conducts interactive testing on the intelligent cockpit AI voice interaction system.

[0026] S3. Collect interaction test results. The intelligent test database calls the AI ​​intelligent model to judge the interaction test results and gives a score to the intelligent cockpit AI voice interaction system based on the judgment results.

[0027] In this application, the interactive test scenarios can be determined based on the needs of testing, interaction, and user research. These scenarios are formed through methods such as scenario classification and deduplication. For example, interactive test scenarios are divided into six categories, including travel assistant, car management, entertainment recommendation, encyclopedia query, fun chat, and image / text generation. The travel assistant test scenario includes interactive tasks such as destination query, route filtering, waypoint setting, traffic condition query, travel planning, and business trip booking. The destination query interactive task includes tasks such as querying destinations near a specified location, querying historical destinations, querying event venues, querying destinations for specific needs, querying destinations under specific conditions, querying nearby destinations, querying destinations by rating, querying destinations by price, querying destinations with fuzzy semantics, querying destinations with limited distance range, querying destinations with combined conditions, and multi-round destination recommendations. Similarly, the car management test scenario includes interactive tasks such as vehicle travel checks, vehicle operation guidance, proactive vehicle services, important schedule reminders, and intelligent telephone dialing. The vehicle travel check interactive task includes checking door status, tire pressure status, and remaining range. Test tasks include inspection and maintenance reminders, and interactive tasks such as music recommendations, audiobook recommendations, video recommendations, radio recommendations, and game recommendations. Music recommendation interactive tasks include volume settings, turning music on / off, searching by artist, searching by year, searching by music charts, searching by music genre, searching by listener type, searching by user habits, combined conditional music search, random music switching, and switching songs based on context. Encyclopedia query interactive tasks include natural sciences, humanities and social sciences, engineering technology, and general knowledge. Natural science interactive tasks include biological knowledge, geographical knowledge, and health knowledge. Chat companion interactive tasks include child car companionship, adult emotional communication, news hot topic Q&A, and future weather queries. Child car companionship interactive tasks include early childhood education, storytelling, and emotional soothing. Image and text generation interactive tasks include custom image generation and custom text generation. Custom image generation interactive tasks include wallpaper generation, multi-round wallpaper generation, and image generation.

[0028] Each interactive test scenario includes multiple source texts for testing under each test task. These source texts are generalized to form generalized texts that constitute the specific test texts. Each test text corresponds to a different test task to complete the testing of the task. For example, the travel assistant test scenario includes test tasks such as destination query, route filtering, waypoint setting, traffic condition query, travel planning, and travel booking. The interactive test task of querying destinations near a specified location under the destination query category includes source texts such as "[Recommended [Hotels] within [West Lake Scenic Area]]". After generalization, this source text can be formatted as "Quickly help me find a hotel near West Lake Scenic Area? I'm a bit tired. I want to find a place to stay inside West Lake Scenic Area. Do you have any recommendations? What hotels are there in West Lake Scenic Area? Please list them for me." For example, in the interactive test task of querying the venue of an event under the destination query interactive task, the source corpus text could be the query "Where will a certain celebrity's concert be held next month?". After generalizing this source corpus text, it becomes something like, "Quickly help me find out where a certain celebrity's concert will be held next month? I want to know the specific location of a certain celebrity's concert next month. Where is a certain celebrity's concert scheduled for next month? Please check." The number of generalized corpus texts can be many, such as hundreds, and can be pre-set to a limit, generated by the generalization corpus system according to preset generalization rules.

[0029] The above is merely an illustrative example and is not limited to it. There is no limit to the specific classification and the number of test texts. The intelligent test database can be infinitely expanded based on the obtained keywords according to the corresponding rules.

[0030] In this embodiment of the application, when testing the intelligent cockpit AI voice interaction system, the intelligent test database can output voice through the test robot connected to it, including outputting different regional accents or dialects to test the intelligent cockpit AI voice interaction system. The test robot collects the interaction feedback results of the intelligent cockpit AI voice interaction system test and feeds them back to the intelligent test database along with the corresponding test data. The intelligent test database then calls the AI ​​intelligent model to judge the interaction test results.

[0031] In an exemplary embodiment of this application, when the intelligent test database calls the AI ​​intelligent model to judge the interaction test results, it judges whether the feedback results of the intelligent cockpit AI voice interaction system are accurate based on the positive or negative correlation of the main or key information contained in the interaction results. For example, when booking a business trip, after outputting a request to query a flight to Shanghai next Monday morning, after obtaining the flight information results from the intelligent cockpit AI voice interaction system, it extracts the main or key information, such as the departure and destination, flight information, and departure time. Combined with the information in the test corpus text, it calls an AI intelligent model to verify the positive or negative correlation of the flight information results and judges whether they are related. If they are positively correlated, the result is considered correct; if they are negatively correlated, the result is considered incorrect. Based on the preset rules of the relationship between correlation and scoring, the test results of the intelligent cockpit AI voice interaction system are scored.

[0032] In an exemplary embodiment of this application, a score is given to the intelligent cockpit AI voice interaction system based on the interaction situation, including an overall score for the intelligent cockpit AI voice interaction system and / or a scenario score for each interaction test scenario. The intelligent test database in this application can evaluate the intelligent cockpit AI voice interaction system based on the interaction response results and intelligent verification using an AI intelligent model. A score is given based on the positive and negative correlation evaluation results of the interaction results. Then, based on the score, the overall test score of the intelligent cockpit AI voice interaction system can be obtained, or the test score of the intelligent cockpit AI voice interaction system for different scenarios can be obtained. This allows for an objective evaluation of the overall and partial performance of the intelligent cockpit AI voice interaction system, facilitating targeted improvements or performance enhancements by users.

[0033] In this embodiment, the intelligent test database can also be connected to a configured test response information acquisition module. The response information acquisition module collects response information from the intelligent cockpit AI voice interaction system. This module can be a functional module of a test robot or a configured intelligent acquisition robot connected to the intelligent test database. The response information includes voice response information and screen output response information. The response information acquisition module includes an image acquisition module and a sound acquisition module. The collected response information is used by the AI ​​intelligent model to determine the positive or negative correlation of the response of the intelligent cockpit AI voice interaction system, thereby realizing the evaluation of the response of the intelligent cockpit AI voice interaction system.

[0034] In an exemplary embodiment of this application, the generalized corpus text can be generated by a preset corpus generalization system. This system processes the collected source corpus text with colloquial expressions from different regions, extracts key information from the colloquial expressions, converts them into keywords, and then generalizes them according to preset rules based on the keywords. The corpus generalization system can be a standalone system connected to the intelligent test database. It generalizes the source corpus text to form generalized corpus text, and then inputs the formed generalized corpus text into the intelligent test database for storage. Alternatively, the corpus generalization system can be used as a sub-unit of the intelligent test database.

[0035] When performing corpus text generalization, the corpus generalization system analyzes the expressions of the input source corpus text, extracts key information, and then forms keywords based on the key information. Based on these keywords, it generates various different expressions, thus creating multiple generalized corpus texts. For example, if the source corpus text is "Go to that [coffee shop] I went to last time," the system analyzes the colloquial expression to extract the key information (coffee shop), then forms the keyword "coffee shop," and finally generates multiple generalized corpus texts, such as "I really want to drink coffee, let's navigate to that coffee shop I went to last time.", "I'm a little tired, take me to that coffee shop I went to last time to rest," "Navigate to that coffee shop I went to last time," "Find that coffee shop I went to last time, then navigate there," etc. This greatly enriches the corpus of generalized corpus texts, thereby improving the intelligence of the test.

[0036] In an exemplary embodiment of this application, the keywords can also be obtained offline or online by the corpus generalization system connecting to the keyword database, or obtained online from relevant websites / platforms via the network. Then, based on the obtained keywords, keyword replacement operations are performed. Based on the generalization rules, the source corpus text is further generalized to form a generalized corpus text, which can be automatically updated to form multiple generalized corpus texts, expanding the generalized corpus text database. The keyword database includes place name information database, scenic spot name database, celebrity information database, cuisine name database, historical allusion database, encyclopedia knowledge database, image information database, music collection database, music melody database, etc., which can obtain the required keyword database or collection. It can be added or modified later according to the usage of the corpus database, and is not limited to this.

[0037] In an exemplary embodiment of this application, when testing the intelligent cockpit AI voice interaction system, the intelligent test database extracts generalized text that meets the test difficulty coefficient based on the externally input test difficulty coefficient, and tests the intelligent cockpit AI voice interaction system through voice output.

[0038] In an exemplary embodiment of this application, when testing the AI ​​voice interaction system for an intelligent cockpit, the preset scenario difficulty coefficient is used to intelligently filter generalized text corpus from test cases according to the test difficulty coefficient instruction, and the selection is performed using the following method:

[0039] Each test covers all interactive test scenarios, all interactive tasks under each interactive test scenario, and all interactive test tasks under each interactive task. The number of source texts under each interactive test task remains consistent, and the number of corpus interaction rounds is consistent. Under the above screening conditions, the test difficulty coefficient is calculated based on the difficulty coefficient value of the selected generalized corpus texts, so that the total difficulty coefficient of the selected generalized corpus texts matches the test difficulty coefficient.

[0040] The number of interaction rounds for each interactive test task can be set according to the needs of each interactive test task. It can be at least one round, or it can be two, three, or more rounds, depending on the testing needs of different interactive test scenarios. In this setting, one interaction round can correspond to one source text, and two interaction rounds can correspond to two source texts. For example, in the case of two rounds, in the interactive test task of setting the volume of music recommendation, the source text is "play song" + "the music is a bit noisy". In the case of one round, in the interactive test task of searching by year in music recommendation, the source text is "search for [classic songs] from [the 1980s]".

[0041] In an exemplary embodiment of this application, there are multiple AI intelligent models. Based on the attributes of the interaction test results, the intelligent test database calls different AI intelligent models to judge the interaction test results. The interaction test results are judged by matching AI intelligent models. That is, when judging whether an interaction test result is successful or not, at least one matching AI intelligent model can be called to make the judgment. When calling, the matching AI intelligent model can be called according to the applicable scenario or scope of the AI ​​intelligent model, or its application field or professional field, and its relevance to the interaction test, so as to ensure the accuracy of the interaction test result judgment.

[0042] In an exemplary embodiment of this application, the AI ​​intelligent model can be embedded within the intelligent test database. After the test information acquisition device feeds back the test results to the intelligent test database, the intelligent test database calls the AI ​​intelligent model to judge the interactive test results, and judges the interactive test results through a matched AI intelligent model. Alternatively, the AI ​​intelligent model can be deployed independently of the intelligent test database, and can be connected through a communication system to achieve the calling and judgment.

[0043] For the same interaction test result, the intelligent test database calls at least two different matching AI intelligent models to judge the same interaction test result, cross-validates, and finally gives a positive or negative correlation evaluation result. If two different AI intelligent models both give a positive correlation evaluation for the same interaction test result, a positive correlation evaluation is given for the interaction response result. If two different AI intelligent models give opposite correlation evaluations for the same interaction test result, a third AI intelligent model is called to judge the same interaction test result. The final judgment result is given based on the judgment result of the third AI intelligent model. If the judgment result of the third AI intelligent model is positive, it is ultimately considered positively correlated; if it is negatively correlated, it is ultimately considered negatively correlated. Through multi-phase AI intelligent model judgment and interaction verification, the accuracy or correlation of the judgment is improved. An odd number of different AI intelligent models can also be used, such as 5 or 7, etc., and the final result is derived based on the statistically counted number of positive and negative correlations, with the one with the most counts being the final result.

[0044] In an exemplary embodiment of this application, the intelligent test database intelligently calculates and filters generalized text that meets the test difficulty coefficient instructions based on a preset difficulty coefficient value. When this text is replaced with voice output, it is connected to a voice conversion system. The voice conversion system converts the generalized text into different accents or dialects, thereby outputting dialectal or standard Mandarin voices corresponding to different regions to conduct interactive tests on the intelligent cockpit AI voice interaction system. By outputting voice as dialects, standard Mandarin, or other voices, the intelligent cockpit AI voice interaction system's ability to recognize and respond to different voices or accents can be tested.

[0045] An exemplary embodiment of this application also provides a computer device, including a processor and a memory, wherein the memory stores a computer program for being invoked and executed by the processor, and the computer program, when executed by the processor, implements the intelligent cockpit AI voice interaction testing method.

[0046] An exemplary embodiment of this application also provides a storage medium storing a computer program, which, when executed by a processor, implements the intelligent cockpit AI voice interaction testing method. The storage medium may be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0047] The exemplary embodiment of this application provides an intelligent cockpit AI voice interaction testing method that, through an intelligent test database, calls up the test corpus text corresponding to the test task in the test scenario, replaces it with voice output, performs interactive testing on the intelligent cockpit AI voice interaction system, and gives a score to the intelligent cockpit AI voice interaction system based on the interaction. This method can realize the intelligentness testing of the intelligent cockpit AI voice interaction system, greatly improves the intelligence and efficiency of the test, and ensures the consistency and objectivity of the test evaluation system.

[0048] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. It will be apparent to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention.

[0049] Therefore, the embodiments should be regarded as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of the equivalents of the claims be included within the invention.

[0050] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A testing method for AI voice interaction in intelligent cockpits, characterized in that, Includes the following steps: Based on the functional analysis of the intelligent cockpit AI voice interaction system, an intelligent test database is established. The intelligent test database stores multiple different interactive test cases. Each interactive test case is used to test an interactive scenario of the intelligent cockpit AI voice interaction system. Each interactive test case contains an interactive test scenario. Each interactive test scenario includes multiple interactive tasks. Each interactive task includes multiple interactive test tasks. Each interactive test task contains multiple source corpus texts. The multiple source corpus texts are generalized to form multiple generalized corpus texts for testing. Each generalized corpus text is assigned a difficulty coefficient value. When testing the AI ​​voice interaction system of the intelligent cockpit, in response to the test difficulty coefficient command input by the external input, the intelligent test database intelligently calculates and filters out the generalized corpus text that meets the test difficulty coefficient command based on the preset difficulty coefficient value of the generalized corpus text, replaces it with voice output, and conducts interactive testing on the AI ​​voice interaction system of the intelligent cockpit. The system collects interaction test results, and the intelligent test database calls the AI ​​intelligent model to judge the interaction test results. Based on the judgment results, it gives a score to the intelligent cockpit AI voice interaction system.

2. The intelligent cockpit AI voice interaction testing method according to claim 1, characterized in that, When the intelligent test database calls the AI ​​intelligent model to judge the interaction test results, it judges whether the feedback results of the intelligent cockpit AI voice interaction system are accurate based on the positive and negative correlation of the interaction results.

3. The intelligent cockpit AI voice interaction testing method according to claim 1, characterized in that, The system is rated based on the interaction, including an overall rating for the intelligent cockpit AI voice interaction system and / or a scenario rating for each interaction scenario.

4. The intelligent cockpit AI voice interaction testing method according to claim 1, characterized in that, The generalized corpus text is generated by a preset corpus generalization system. The source corpus text is processed with colloquial expressions from different regions. Key information is extracted from the colloquial expressions and converted into keywords. The generalization is then performed based on the keywords according to preset rules. The keywords are obtained offline or online by the corpus generalization system by connecting to a keyword database, or obtained online from relevant websites / platforms. The keyword database includes a place name database, a scenic spot name database, a celebrity information database, a cuisine name database, a historical allusion database, a music library, and an encyclopedia knowledge database.

5. The intelligent cockpit AI voice interaction testing method according to claim 1, characterized in that, When testing the AI ​​voice interaction system for the intelligent cockpit, the intelligent test database intelligently selects generalized language text from the test cases according to the test difficulty coefficient instruction, using the following method: Each test covers all interactive test scenarios, all interactive tasks under each interactive test scenario, and all interactive test tasks under each interactive task. The number of source texts under each interactive test task remains the same, and the number of corpus interaction rounds is consistent. Based on the difficulty coefficient values ​​of the generalized corpus texts, the final total difficulty coefficient of the selected generalized corpus texts is calculated to match the test difficulty coefficient.

6. The intelligent cockpit AI voice interaction testing method according to claim 1, characterized in that, The AI ​​intelligent model is multiple. Based on the attributes of the interaction test results, the intelligent test database calls different AI intelligent models to judge the interaction test results, and judges the interaction test results through the matched AI intelligent model.

7. The intelligent cockpit AI voice interaction testing method according to claim 6, characterized in that, For the same interaction test result, the intelligent test database calls at least two different AI intelligent models to judge the same interaction test result, cross-validate, and finally give the positive and negative correlation evaluation result.

8. The intelligent cockpit AI voice interaction testing method according to claim 1, characterized in that, The intelligent test database is based on the preset difficulty coefficient value of generalized corpus text. It intelligently calculates and filters out generalized corpus text that meets the test difficulty coefficient instruction. When replaced with voice output, it outputs dialect voices or standard Mandarin voices corresponding to different regions to conduct interactive tests on the intelligent cockpit AI voice interaction system.

9. A computer device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program for being called and executed by the processor, and the computer program, when executed by the processor, implements the intelligent cockpit AI voice interaction test method as described in any one of claims 1 to 8.

10. A storage medium, characterized in that, The system contains a computer program that, when executed by a processor, implements the intelligent cockpit AI voice interaction testing method as described in any one of claims 1 to 8.

Images