Intelligent interactive terminal acoustic automation test method, device and equipment

Abstract

The present application relates to the technical field of automation testing, solves the problem that it is difficult to realize automatic acquisition and accurate determination of intelligent acoustic interaction terminal test results under consistent and traceable standardized test sound field in the prior art, and provides an intelligent interaction terminal acoustic automation testing method, device and equipment, comprising: constructing a target test sound field corresponding to a to-be-tested acoustic interaction terminal, and triggering the to-be-tested acoustic interaction terminal to execute a corresponding acoustic interaction process; collecting running state data in the acoustic interaction process through a state data acquisition unit, and collecting audio feedback data output by the to-be-tested acoustic interaction terminal through a sound field acquisition unit; analyzing and processing the running state data and the audio feedback data in combination with test evaluation index data, and automatically generating corresponding test results. The present application realizes calibration and traceability of acoustic parameters and automatic determination of test results, and improves the consistency of different batch test conditions and the accuracy of test results.

Description

Technical Field

[0001] This invention relates to the field of automated testing technology, and in particular to an automated acoustic testing method, apparatus, and equipment for intelligent interactive terminals. Background Technology

[0002] With the rapid development of artificial intelligence, large-scale model inference, and edge acoustic computing technologies, acoustic interaction has become an important interaction method for intelligent automotive cockpits, smart homes, mobile smart terminals, and intelligent customer service systems. Various intelligent acoustic interaction terminals not only need to possess basic capabilities such as acoustic wake-up, acoustic recognition, semantic understanding, and command execution, but also need to support functions such as continuous dialogue, multi-turn interaction, noise reduction in complex environments, and multi-scenario adaptation. Therefore, during product development, mass production verification, and quality testing, extensive testing of the acoustic interaction performance of intelligent acoustic interaction terminals is necessary to verify the product's functional reliability and interaction effects under different test scenarios.

[0003] In existing technologies, testing of intelligent acoustic interactive terminals typically employs manual or semi-automated methods. This involves testers playing test instructions according to pre-set test cases and judging the results by manually listening, observing the device's response, or reviewing log information. While some testing systems can simulate the test environment using hardware such as human mouths or noise generators, the testing process still relies heavily on manual analysis and judgment of recognition, execution, and response results. On one hand, manual testing is inefficient and prone to omissions, misjudgments, and statistical errors in large-scale testing scenarios. On the other hand, differences in the vocalization methods of different testers, testing environments, and testing procedures lead to insufficient consistency in testing conditions, resulting in a lack of objectivity and repeatability in the test results. Furthermore, existing testing systems often lack the ability to collaboratively analyze terminal operating status data, audio feedback data, and test evaluation indicators, making it difficult to achieve automated evaluation and unified judgment of test results.

[0004] Chinese patent CN121789644A discloses an acoustic testing method, system, device, and medium. It obtains acoustic test instructions, inputs the acoustic test instructions into the vehicle-mounted acoustic interaction model to obtain acoustic response data, and obtains the corresponding expected response text and acoustic recognition text. It further uses a multi-level matching method of text matching, semantic matching, and intent matching to determine the test results. This solution can also generate standardized test acoustics through acoustic synthesis and superimpose traffic noise or environmental noise to simulate different sound scenarios. While the aforementioned patents can achieve standardized generation of test acoustics and automatic matching and judgment of test results to a certain extent, they mainly focus on matching and analyzing acoustic recognition text with expected response text. They have not yet established a closed-loop calibration and metrological traceability mechanism based on standard sound pressure data for artificial mouths, noise simulators, and sound field acquisition devices, making it difficult to achieve consistent control over acoustic sound pressure, background noise sound pressure, and signal-to-noise ratio. At the same time, they have not fully integrated the operating status data of the terminal under test, audio feedback data, and multiple test evaluation indicators for comprehensive evaluation. Therefore, it is still difficult to guarantee the consistency of test sound fields, the traceability of test data, and the comprehensiveness and accuracy of test results under different batches, different devices, and different environmental conditions.

[0005] Therefore, how to construct a consistent and traceable standardized test sound field, and how to automatically acquire test data from intelligent acoustic interaction terminals and accurately determine test results under the standardized test sound field, are technical problems that urgently need to be solved. Summary of the Invention

[0006] In view of this, embodiments of the present invention provide an automated acoustic testing method, apparatus, and equipment for intelligent interactive terminals, in order to solve the problem in the prior art that it is difficult to automatically acquire and accurately determine the test results of intelligent acoustic interactive terminals under a consistent, traceable, and standardized test sound field.

[0007] In a first aspect, embodiments of the present invention provide an automated acoustic testing method for intelligent interactive terminals, applied to an automated acoustic testing system for intelligent interactive terminals. The system includes a testing platform, an acoustic signal output unit, a sound field acquisition unit, an acoustic calibration unit, a status data acquisition unit, and an acoustic interactive terminal under test. The testing platform is communicatively connected to the acoustic signal output unit, the sound field acquisition unit, the acoustic calibration unit, and the status data acquisition unit, respectively. The status data acquisition unit is communicatively connected to the acoustic interactive terminal under test. The method includes: Obtain the test configuration data corresponding to the acoustic interactive terminal to be tested, wherein the test configuration data includes test corpus data, environmental noise data, and test evaluation index data; Based on the test configuration data, the test scenario is configured through the acoustic calibration unit to determine the target test sound field parameters corresponding to the acoustic interaction terminal to be tested. Based on the target test sound field parameters, the sound signal output unit is controlled to output the test audio signal corresponding to the test corpus data and the environmental noise signal corresponding to the environmental noise data, so as to construct the target test sound field corresponding to the acoustic interaction terminal under test and trigger the acoustic interaction terminal under test to execute the corresponding acoustic interaction process. The status data acquisition unit collects the operating status data of the acoustic interaction terminal under test during the acoustic interaction process, and the sound field acquisition unit collects the audio feedback data output by the acoustic interaction terminal under test. Based on the operational status data, the audio feedback data, and the test evaluation index data, the test results are analyzed to obtain the corresponding test results.

[0008] Preferably, the step of configuring the test scene based on the test configuration data through the acoustic calibration unit to determine the target test sound field parameters corresponding to the acoustic interaction terminal under test includes: Obtain the target test sound pressure level, target background noise, and target signal-to-noise ratio parameters corresponding to the test configuration data; The control sound signal output unit outputs a preset calibration audio signal and a preset calibration noise signal; The actual test sound pressure data and actual background noise data in the test area are collected through the acoustic calibration unit; Calculate the actual signal-to-noise ratio parameter based on the actual test sound pressure data and actual background noise data; Calculate the first deviation between the actual test sound pressure data and the target test sound pressure parameter, the second deviation between the actual background noise data and the target background noise parameter, and the third deviation between the actual signal-to-noise ratio parameter and the target signal-to-noise ratio parameter, respectively. The output gain parameter of the acoustic signal output unit is adjusted according to the first deviation value, the second deviation value and the third deviation value. When the first deviation value, the second deviation value and the third deviation value all meet the preset threshold condition, the target test acoustic field parameter is determined.

[0009] Preferably, the step of adjusting the output gain parameter of the acoustic signal output unit according to the first deviation value, the second deviation value, and the third deviation value, and determining the target test sound field parameter when the first deviation value, the second deviation value, and the third deviation value all meet the preset threshold condition, includes: Establish a test sound pressure error model between the target test sound pressure parameter and the actual test sound pressure data, and a noise sound pressure error model between the target background noise parameter and the actual background noise data; Based on the test sound pressure error model, calculate the first gain correction parameter corresponding to the test audio output channel; Based on the noise sound pressure error model, calculate the second gain correction parameter corresponding to the noise output channel; Using the first gain correction parameter, adjust the output gain parameter corresponding to the test audio signal, and using the second gain correction parameter, adjust the output gain parameter corresponding to the ambient noise signal. The new actual test sound pressure data and new actual background noise data are re-acquired and adjusted. The process returns to the step of calculating the actual signal-to-noise ratio parameter based on the actual test sound pressure data and actual background noise data, until the first deviation value, the second deviation value and the third deviation value all meet the preset threshold conditions, and then the target test sound field parameter is determined.

[0010] Preferably, after adjusting the output gain parameter corresponding to the acoustic signal output unit according to the first deviation value, the second deviation value, and the third deviation value, and determining the target test sound field parameter when the first deviation value, the second deviation value, and the third deviation value all meet the preset threshold condition, the method further includes: Obtain the target output gain parameter of the acoustic signal output unit when the preset threshold condition is met; A standard sound source output profile is generated based on the target output gain parameters, and a mapping relationship is established between the standard sound source output profile and the target test sound field parameters. The standard sound source output file is invoked, and the sound signal output unit is controlled to output a reference calibration audio signal with a sound pressure level corresponding to the target test sound pressure parameter. The reference calibration audio signal is acquired by the acoustic calibration unit to obtain reference sound pressure data; By comparing the reference sound pressure data with the target test sound pressure parameter, when the deviation between the reference sound pressure data and the target test sound pressure parameter meets the preset calibration threshold condition, the current sound field is determined as the reference sound field.

[0011] Preferably, after comparing the reference sound pressure data with the target test sound pressure parameter, and determining the current sound field as the reference sound field when the deviation between the reference sound pressure data and the target test sound pressure parameter meets a preset calibration threshold condition, the method further includes: In the reference sound field, the sound field acquisition unit is controlled to acquire the reference calibration audio signal to obtain the corresponding sampling electrical signal; Based on the sound pressure conversion function of the sound field acquisition unit, the sampled electrical signal is converted into sampled sound pressure data; Calculate the sound pressure deviation between the sampled sound pressure data and the reference sound pressure data, wherein the reference sound pressure data is the standard reference sound pressure data measured by the acoustic calibration unit; The sensitivity correction coefficient of the corresponding channel of the sound field acquisition unit is calculated based on the sound pressure deviation value; Establish a mapping relationship between the sensitivity correction coefficient and the acquisition channel of the sound field acquisition unit, and store the sensitivity correction coefficient in the parameter library of the test platform.

[0012] Preferably, before controlling the sound signal output unit to output the test audio signal corresponding to the test corpus data and the environmental noise signal corresponding to the environmental noise data based on the target test sound field parameters, so as to construct the target test sound field corresponding to the acoustic interaction terminal under test, and triggering the acoustic interaction terminal under test to execute the corresponding acoustic interaction process, the method further includes: Acquire the device clock information corresponding to the acoustic signal output unit, the acoustic field acquisition unit, and the status data acquisition unit; Based on the global time reference corresponding to the device clock information, a unified clock synchronization relationship is established between the sound signal output unit, the sound field acquisition unit, and the status data acquisition unit. Based on the timing control parameters corresponding to the unified clock synchronization relationship, test the execution timing of audio signal output, environmental noise signal output, audio feedback data acquisition, and running status data acquisition, and generate corresponding timing control results. Based on the timing control results, each hardware unit is controlled to execute corresponding automated test tasks.

[0013] Preferably, the step of acquiring the operating status data of the acoustic interaction terminal under test through the status data acquisition unit during the acoustic interaction process, and acquiring the audio feedback data output by the acoustic interaction terminal under test through the sound field acquisition unit includes: Identify the physical communication interface type of the acoustic interactive terminal under test, and load a communication protocol stack and message parsing rules that match the physical communication interface type. The physical communication interface type includes CAN bus interface, RS232 / RS485 serial port interface and Ethernet interface. A low-level data connection is established between the status data acquisition unit and the acoustic interaction terminal under test based on the communication protocol stack. Listen to and capture the original interactive messages generated by the acoustic interactive terminal under test during the response to test commands; The original interactive message is decoded according to the message parsing rules to extract running status data including wake-up flag, acoustic recognition text, instruction execution status code and timestamp; Simultaneously, the sound field acquisition unit acquires the audio feedback data output by the acoustic interactive terminal under test.

[0014] Preferably, the step of analyzing the test results based on the operating status data, the audio feedback data, and the test evaluation index data to obtain the corresponding test results includes: Acquire the running status data, the audio feedback data, the test corpus data, and the test evaluation index data; Based on the test corpus data and the test evaluation index data, determine the corresponding expected recognition results, expected execution results, and evaluation rules; The identification result data in the running status data is matched and analyzed with the expected identification result to obtain the identification judgment result; The execution result data in the running status data is matched and analyzed with the expected execution result to obtain the execution determination result; Calculate the response latency parameters based on the timestamp data and the playback time data corresponding to the test audio signal, and generate the response latency evaluation results; A wake-up evaluation result is generated based on the wake-up status data in the operation status data; Based on the audio feedback data, audio quality evaluation parameters are extracted, and audio quality evaluation results are generated. A comprehensive evaluation calculation is performed based on the recognition judgment result, the execution judgment result, the response latency evaluation result, the wake-up evaluation result, and the audio quality evaluation result, and the corresponding test result is output.

[0015] Secondly, embodiments of the present invention provide an automated acoustic testing device for intelligent interactive terminals, applied to an automated acoustic testing system for intelligent interactive terminals. The system includes a testing platform, an acoustic signal output unit, a sound field acquisition unit, an acoustic calibration unit, a status data acquisition unit, and an acoustic interactive terminal under test. The testing platform is communicatively connected to the acoustic signal output unit, the sound field acquisition unit, the acoustic calibration unit, and the status data acquisition unit, respectively. The status data acquisition unit is communicatively connected to the acoustic interactive terminal under test. The device includes: The data acquisition module is used to acquire the test configuration data corresponding to the acoustic interactive terminal to be tested, wherein the test configuration data includes test corpus data, environmental noise data and test evaluation index data; The scene configuration module is used to configure the test scene based on the test configuration data through the acoustic calibration unit, and determine the target test sound field parameters corresponding to the acoustic interaction terminal to be tested. The interactive triggering module is used to control the sound signal output unit to output the test audio signal corresponding to the test corpus data and the environmental noise signal corresponding to the environmental noise data based on the target test sound field parameters, so as to construct the target test sound field corresponding to the acoustic interactive terminal under test, and trigger the acoustic interactive terminal under test to execute the corresponding acoustic interaction process. The data acquisition module is used to acquire the operating status data of the acoustic interaction terminal under test during the acoustic interaction process through the status data acquisition unit, and to acquire the audio feedback data output by the acoustic interaction terminal under test through the sound field acquisition unit. The test result analysis module is used to analyze the test results based on the running status data, the audio feedback data, and the test evaluation index data to obtain the corresponding test results.

[0016] Thirdly, embodiments of the present invention provide a testing device, including: at least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the method of the first aspect described above.

[0017] In summary, the beneficial effects of the present invention are as follows: The automated acoustic testing method, apparatus, and equipment for intelligent interactive terminals provided in this invention acquire test configuration data and construct a target test sound field corresponding to the acoustic interactive terminal under test based on the test configuration data. Test audio signals and environmental noise signals are output in a unified and controllable test environment to trigger the acoustic interactive terminal under test to execute the corresponding acoustic interaction process. Simultaneously, the system collects operational status data during the acoustic interaction process through a status data acquisition unit and collects audio feedback data output by the terminal through a sound field acquisition unit. Furthermore, the test results are analyzed in conjunction with test evaluation index data, thereby achieving automated execution of the intelligent acoustic interactive terminal testing process and automated determination of test results. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments of the present invention will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, and these are all within the protection scope of the present invention.

[0019] Figure 1 This is a schematic diagram of the structure of the intelligent interactive terminal acoustic automated testing system in Embodiment 1 of the present invention; Figure 2 This is a flowchart illustrating the automated acoustic testing method for intelligent interactive terminals in Embodiment 1 of the present invention. Figure 3 This is a structural block diagram of the automated acoustic testing device for intelligent interactive terminals in Embodiment 2 of the present invention; Figure 4 This is a schematic diagram of the test equipment in Embodiment 3 of the present invention. Detailed Implementation

[0020] The features and exemplary embodiments of various aspects of the present invention will now be described in detail. To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only configured to explain the present invention and are not configured to limit the present invention. For those skilled in the art, the present invention can be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the present invention by illustrating examples of the invention.

[0021] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0022] It should be noted that all actions involving the acquisition of signals, information, or data in this invention are carried out in compliance with the relevant data protection laws and regulations of the locality and with authorization from the owner of the relevant device.

[0023] Example 1 This invention provides an automated acoustic testing method for intelligent interactive terminals, applied to an automated acoustic testing system for intelligent interactive terminals. The system includes a test platform, an acoustic signal output unit, a sound field acquisition unit, an acoustic calibration unit, a status data acquisition unit, and an acoustic interactive terminal to be tested. The test platform is communicatively connected to the acoustic signal output unit, the sound field acquisition unit, the acoustic calibration unit, and the status data acquisition unit, respectively. The status data acquisition unit is communicatively connected to the acoustic interactive terminal to be tested.

[0024] Specifically, this automated testing method operates within a testing system uniformly scheduled by a testing platform. The testing platform can be understood as the central control and data processing core of the entire system. It is not merely used to issue simple start / stop commands, but also undertakes functions such as test task configuration, hardware coordination control, acoustic calibration invocation, communication data parsing, result judgment, and report generation. The acoustic signal output unit corresponds to the audio signal acquisition, processing, and output control module, artificial mouth, and noise simulator, among other sound source-side hardware, as specified in the disclosure document. It is used to output standardized test acoustics according to the corpus, sound pressure level, noise type, and signal-to-noise ratio requirements issued by the testing platform. The system includes a sound field acquisition unit and a high-precision microphone array for acquiring the audio response of the external acoustics of the terminal under test or in the test sound field. The acoustic calibration unit, corresponding to a sound level meter and its associated closed-loop calibration algorithm, is used to calibrate the sound pressure, sensitivity, and sound field consistency of the artificial mouth, noise simulator, and microphone array before or during formal testing. The status data acquisition unit, corresponding to a multi-standard data acquisition card, can establish a communication link with the acoustic interactive terminal under test via CAN bus, serial port, or Ethernet to obtain real-time operational status data such as wake-up results, command execution status, interruption feedback, and interaction delay. In actual setup, the test platform establishes communication connections with the sound signal output unit, sound field acquisition unit, acoustic calibration unit, and status data acquisition unit, enabling the test platform to uniformly control acoustic playback, noise superposition, sound field acquisition, acoustic calibration, and status feedback under the same clock reference. The status data acquisition unit is then connected to the acoustic interactive terminal under test, allowing the terminal under test to automatically transmit its internal interaction messages and execution feedback back to the test platform after receiving standard test commands from the artificial mouth and completing acoustic interaction. Through the above connections, the system combines sound source output, sound field detection, acoustic calibration, and terminal status acquisition into a closed-loop test link. This avoids the differences in accent, speech rate, emotion, and environment caused by traditional human voices, and can automatically judge the test results through underlying communication messages. This mechanism reduces the errors caused by manual listening, screen viewing, and manual statistics, and improves the consistency, repeatability, and automation of intelligent acoustic interaction terminals in different product scenarios such as in-vehicle cockpits, smart homes, smart customer service, and mobile terminals.

[0025] Please see Figure 1 , Figure 1 This illustration shows the overall structure of the intelligent interactive terminal acoustic automated testing system and the connection relationships between its components in an embodiment of the present invention. The intelligent acoustic interactive testing platform includes a basic settings module, a hardware management and collaborative control module, a sound source calibration and consistency control module, a corpus management module, a test case design and indicator setting module, a task management and execution monitoring module, a data acquisition and evaluation module, and a result display and report management module. Each module can be implemented by one or more processors in the testing platform executing corresponding program instructions.

[0026] The audio signal acquisition, processing and output control module is connected to the intelligent acoustic interaction test platform, the artificial mouth, the noise simulator and the high-precision microphone array respectively. It is used to receive the sound output control commands issued by the test platform, send the audio signals corresponding to the test data and the audio signals corresponding to the environmental noise to the artificial mouth and the noise simulator respectively, and receive the audio signals acquired by the high-precision microphone array.

[0027] An artificial mouth and a noise simulator are positioned facing the object under test. The artificial mouth outputs test audio signals to the object, while the noise simulator outputs ambient noise signals to the test area where the object is located. This ensures that the test audio signals and ambient noise signals together form a target test sound field corresponding to the target test sound field parameters. A high-precision microphone array is positioned within the test area to collect acoustic feedback from the object and sound field signals from the test area. The collected results are then transmitted to the intelligent acoustic interactive testing platform via an audio signal acquisition, processing, and output control module.

[0028] The sound level meter is set at preset sound field measurement points to measure the actual acoustic sound pressure output by the artificial mouth and the actual background noise sound pressure output by the noise simulator during sound source calibration. It also provides standard reference sound pressure data during the sound field acquisition unit calibration process, assisting the sound source calibration and consistency control module in adjusting the sound source output gain and correcting the sensitivity of the high-precision microphone array. The test object communicates with a multi-standard data acquisition card. Based on the physical communication interface and communication protocol corresponding to the test object, the multi-standard data acquisition card acquires the original interactive messages generated by the test object during its response to the test audio signal, and transmits the parsed wake-up state, acoustic recognition results, command execution state, and timestamp, among other operational status data, to the intelligent acoustic interaction test platform.

[0029] During testing, the hardware management and collaborative control module controls the artificial mouth, noise simulator, high-precision microphone array, and multi-standard data acquisition card to operate collaboratively according to a unified time reference. The data acquisition and evaluation module analyzes and compares the operating status data and audio feedback data of the tested object with the test evaluation indicators of the corresponding test cases to obtain the judgment result of a single test case. After the task is completed, the result display and report management module summarizes the judgment results and raw test data of each test case to generate the corresponding test results and test reports.

[0030] Please see Figure 2 The method includes: Obtain the test configuration data corresponding to the acoustic interactive terminal to be tested, wherein the test configuration data includes test corpus data, environmental noise data, and test evaluation index data; Specifically, before starting the test, the testing platform first obtains test configuration data matching the product type, test standards, and test tasks of the acoustic interactive terminal under test. The test corpus data is mainly used to determine the acoustic content to be played to the acoustic interactive terminal, which may include wake-up words, control commands, interactive recognition statements, and multi-turn dialogue statements; environmental noise data is used to determine the type and intensity of background noise to be superimposed during the test; and test evaluation index data is used to limit the judgment conditions for each test case, such as whether wake-up was successful, whether the recognized content is consistent with expectations, whether the command was executed, and whether the response latency exceeded limits. In actual use, testers can select the standard test template for the corresponding product in the task management module of the testing platform. The system automatically binds the test corpus, environmental noise file, and evaluation index set in the corpus. They can also configure the delay time, number of loops, or expected feedback results according to non-standard test requirements. In this way, the test task has already formed complete input conditions and judgment rules before execution, and the subsequent test process does not require manual selection of acoustic elements or temporary judgment results, which helps reduce statistical errors caused by differences in tester operation.

[0031] Based on the test configuration data, the test scenario is configured through the acoustic calibration unit to determine the target test sound field parameters corresponding to the acoustic interaction terminal to be tested. Specifically, after acquiring the test configuration data, the test platform extracts the target test sound pressure parameters, target background noise parameters, and target signal-to-noise ratio parameters corresponding to the acoustic interaction terminal under test from the test configuration data, and calls the acoustic calibration unit to measure the actual acoustic parameters in the test area. The test platform controls the sound signal output unit to output a preset calibration audio signal and a preset calibration noise signal, respectively, and acquires the actual test sound pressure data and actual background noise data at the standard measurement point through the acoustic calibration unit, and calculates the actual signal-to-noise ratio parameter based on the two. Subsequently, the test platform compares the actual test sound pressure data, actual background noise data, and actual signal-to-noise ratio parameter with the corresponding target parameters, adjusts the output level, power amplifier gain, or channel gain of the artificial mouth and noise simulator according to the deviations, and re-acquires the adjusted actual acoustic parameters until all deviations meet the corresponding preset threshold conditions, thereby determining the target test sound field parameters corresponding to the acoustic interaction terminal under test. For test scenarios such as vehicle cockpits, smart homes, smart customer service, or mobile terminals where sound field calibration has been completed, the test platform can also call up pre-saved sound field configuration schemes and use the acoustic calibration unit to verify and correct the current actual acoustic parameters as necessary, so that the test sound field formed subsequently based on the target test sound field parameters maintains high consistency, repeatability, and traceability.

[0032] Based on the target test sound field parameters, the sound signal output unit is controlled to output the test audio signal corresponding to the test corpus data and the environmental noise signal corresponding to the environmental noise data, so as to construct the target test sound field corresponding to the acoustic interaction terminal under test and trigger the acoustic interaction terminal under test to execute the corresponding acoustic interaction process. Specifically, after determining the target test sound field parameters corresponding to the acoustic interaction terminal under test, the test platform, according to the execution order of the current test cases, calls the corresponding test corpus data and environmental noise data from the test configuration data, and determines the output gain, output time, playback duration, and loop count of the test audio signal and environmental noise signal respectively based on the target test sound field parameters. Subsequently, the test platform controls the artificial mouth in the sound signal output unit to output the test audio signal corresponding to the test corpus data, and controls the noise simulator to output the environmental noise signal corresponding to the environmental noise data, so that the actual acoustic sound pressure, actual background noise sound pressure, and actual signal-to-noise ratio in the test area match the target test sound field parameters, thereby constructing the target test sound field corresponding to the acoustic interaction terminal under test. The test audio signal may include wake-up acoustics, control command acoustics, or multi-turn dialogue acoustics, and the environmental noise signal may include indoor noise, urban noise, high-speed driving noise, or background noise corresponding to other test scenarios. After receiving the test audio signal in the target test sound field, the acoustic interaction terminal under test executes at least one of the following acoustic interaction processes sequentially or in parallel: wake-up detection, acoustic recognition, semantic parsing, command execution, and acoustic feedback, to generate corresponding operating status data and audio feedback data. Since both the test audio signal and the ambient noise signal are standardized according to the calibrated target test sound field parameters, the impact of differences in the tester's speaking style, speech rate, volume, accent, distance, and emotional state on the test results can be reduced, allowing different acoustic interaction terminals under test or different batches of tests to be performed under relatively consistent acoustic conditions.

[0033] The status data acquisition unit collects the operating status data of the acoustic interaction terminal under test during the acoustic interaction process, and the sound field acquisition unit collects the audio feedback data output by the acoustic interaction terminal under test. Specifically, during the acoustic interaction process performed by the acoustic interactive terminal under test, the test platform does not only collect a single acoustic result, but simultaneously acquires two types of mutually corroborating data through the status data acquisition unit and the sound field acquisition unit. The status data acquisition unit can connect to the acoustic interactive terminal under test via CAN bus, serial port, or Ethernet, monitoring the interactive messages output by the device under test in real time and extracting operational status data such as wake-up status, recognition results, control feedback commands, execution status, interruption feedback, and interaction latency from the messages. The sound field acquisition unit can be implemented using a high-precision microphone array, used to collect acoustic broadcasts, prompts, or other audio feedback output by the acoustic interactive terminal under test during the interaction process. To ensure that the acquisition results accurately correspond to specific test cases, the test platform can uniformly control the working sequence of the artificial mouth, noise simulator, status data acquisition unit, and microphone array, establishing a temporal correspondence between a certain audio segment, a certain communication message, and a certain test corpus. In this way, when the tested terminal fails to wake up, is misidentified, responds slowly, or broadcasts abnormally, the system can see the actual internal state of the terminal from the communication data and retain the external acoustic feedback from the audio data, which facilitates subsequent automatic judgment and problem tracing.

[0034] Based on the operational status data, the audio feedback data, and the test evaluation index data, the test results are analyzed to obtain the corresponding test results.

[0035] Specifically, after obtaining operational status data and audio feedback data, the testing platform analyzes the test results in conjunction with test evaluation index data. The data acquisition and evaluation module can first parse the raw communication messages uploaded by the status data acquisition unit, extract the actual feedback content corresponding to the current test case, and then compare it with the pre-set expected results. For items involving audio quality evaluation, it can further process the audio feedback data acquired by the sound field acquisition unit to calculate acoustic output quality, MOS score, or other acoustic evaluation results. If the terminal under test completes wake-up, correctly recognizes acoustic content, and executes the corresponding instructions within the specified time, and the audio feedback meets the evaluation index requirements, the testing platform can determine that the test case is passed. If there are situations such as failure to wake up, false wake-up, inconsistent recognition results, unexecuted instructions, response timeout, or abnormal audio feedback, it can be recorded as failure or abnormality. After completing multiple test cases, the testing platform can further statistically analyze results such as wake-up rate, recognition rate, false wake-up frequency, average interaction latency, and test case pass rate, and provide a basis for generating subsequent test reports and raw data tables. Since the judgment process is automatically completed based on communication messages and audio acquisition data, it does not require manual listening, screen viewing, or manual counting, thus improving testing efficiency and reducing errors caused by subjective human judgment.

[0036] Preferably, the step of configuring the test scene based on the test configuration data through the acoustic calibration unit to determine the target test sound field parameters corresponding to the acoustic interaction terminal under test includes: Obtain the target test sound pressure level, target background noise, and target signal-to-noise ratio parameters corresponding to the test configuration data; The control sound signal output unit outputs a preset calibration audio signal and a preset calibration noise signal; The actual test sound pressure data and actual background noise data in the test area are collected through the acoustic calibration unit; Calculate the actual signal-to-noise ratio parameter based on the actual test sound pressure data and actual background noise data; Calculate the first deviation between the actual test sound pressure data and the target test sound pressure parameter, the second deviation between the actual background noise data and the target background noise parameter, and the third deviation between the actual signal-to-noise ratio parameter and the target signal-to-noise ratio parameter, respectively. The output gain parameter of the acoustic signal output unit is adjusted according to the first deviation value, the second deviation value and the third deviation value. When the first deviation value, the second deviation value and the third deviation value all meet the preset threshold condition, the target test acoustic field parameter is determined.

[0037] Specifically, in the process of constructing the target test sound field based on the test configuration data, the test platform first reads the target test sound pressure parameter, target background noise parameter, and target signal-to-noise ratio parameter from the current test task. The target test sound pressure parameter is used to limit the sound pressure level that should be reached at the standard measurement point when the artificial mouth outputs test acoustics. The target background noise parameter is used to limit the noise decibel that should be reached when the noise simulator outputs environmental noise. The target signal-to-noise ratio parameter is used to limit the strength difference between the test acoustics and the background noise, so as to ensure that the terminal under test is under acoustic conditions that match the actual use scenario.

[0038] Subsequently, the test platform controls the acoustic signal output unit to output a preset calibration audio signal and a preset calibration noise signal, respectively. The preset calibration audio signal can be an acoustic segment or a sine wave with a fixed amplitude, and the preset calibration noise signal can be pink noise or a background noise signal corresponding to the target scene, used to form a measurable calibration sound field before the formal test. The sound level meter in the acoustic calibration unit is arranged at the standard measurement point position in the test area to collect the actual test sound pressure data when the artificial mouth outputs and the actual background noise data when the noise simulator outputs. The test platform then calculates the actual signal-to-noise ratio parameter under the current sound field based on the two types of measured sound pressure data. Because the output sound pressure of the artificial mouth and noise simulator may deviate from the set value due to speaker unit aging, amplifier gain drift, or changes in ambient temperature and humidity after long-term use, the test platform will further compare the actual test sound pressure data with the target test sound pressure parameter to obtain the first deviation value, compare the actual background noise data with the target background noise parameter to obtain the second deviation value, and compare the actual signal-to-noise ratio parameter with the target signal-to-noise ratio parameter to obtain the third deviation value. When any deviation value does not meet the preset threshold condition, the test platform will adjust the output gain parameter of the corresponding sound signal output unit according to the deviation direction and deviation magnitude. For example, it may increase or decrease the level of the artificial mouth output channel, the amplifier gain, or adjust the gain of the noise simulator output channel, so that the acoustic sound pressure, background noise, and the difference between the two gradually approach the target configuration.

[0039] After one or more feedback adjustments, if the first, second, and third deviation values ​​all fall within the allowable range, it indicates that the acoustic output intensity, noise output intensity, and signal-to-noise ratio within the test area have met the test conditions required for the acoustic interactive terminal under test. At this point, the test platform determines the test sound field parameters as the target test sound field parameters. This closed-loop configuration method based on actual sound level meter data avoids errors caused by estimating sound pressure solely based on software-set levels. It establishes a correspondence between the output parameters of the artificial mouth and noise simulator and the actual sound field results, thereby improving the acoustic consistency and reproducibility of results when testing acoustic interactive terminals of different batches and types.

[0040] Preferably, the step of adjusting the output gain parameter of the acoustic signal output unit according to the first deviation value, the second deviation value, and the third deviation value, and determining the target test sound field parameter when the first deviation value, the second deviation value, and the third deviation value all meet the preset threshold condition, includes: Establish a test sound pressure error model between the target test sound pressure parameter and the actual test sound pressure data, and a noise sound pressure error model between the target background noise parameter and the actual background noise data; Based on the test sound pressure error model, calculate the first gain correction parameter corresponding to the test audio output channel; Based on the noise sound pressure error model, calculate the second gain correction parameter corresponding to the noise output channel; Using the first gain correction parameter, adjust the output gain parameter corresponding to the test audio signal, and using the second gain correction parameter, adjust the output gain parameter corresponding to the ambient noise signal. The new actual test sound pressure data and new actual background noise data are re-acquired and adjusted. The process returns to the step of calculating the actual signal-to-noise ratio parameter based on the actual test sound pressure data and actual background noise data, until the first deviation value, the second deviation value and the third deviation value all meet the preset threshold conditions, and then the target test sound field parameter is determined.

[0041] Specifically, when adjusting the output gain parameter of the acoustic signal output unit, the test platform does not simply determine whether the sound field is qualified based on the sound pressure data measured once. Instead, it matches the target test sound pressure parameter with the actual test sound pressure data collected by the acoustic calibration unit to establish a test sound pressure error model, and matches the target background noise parameter with the actual background noise data to establish a noise sound pressure error model.

[0042] The test sound pressure error model is used to characterize the deviation of the actual output sound pressure of the test audio output channel where the artificial mouth is located from the target sound pressure. The noise sound pressure error model is used to characterize the deviation of the actual output sound pressure of the noise output channel where the noise simulator is located from the target noise sound pressure. Both can include the direction and magnitude of the deviation, enabling the test platform to determine whether the output gain needs to be increased or decreased, and within what range the adjustment range should be controlled.

[0043] Subsequently, the testing platform calculates the first gain correction parameter corresponding to the test audio output channel based on the test sound pressure error model, and calculates the second gain correction parameter corresponding to the noise output channel based on the noise sound pressure error model. The first gain correction parameter is used to correct the output level or amplifier gain when the artificial mouth plays test acoustics, while the second gain correction parameter is used to correct the output level or channel gain when the noise simulator plays ambient noise. Since acoustic sound pressure and background noise jointly determine the actual signal-to-noise ratio (SNR), after adjusting the acoustic output gain and noise output gain respectively, the testing platform re-collects new actual test sound pressure data and new actual background noise data through the acoustic calibration unit, recalculates the adjusted actual SNR parameter, and continues to compare it with the target test sound pressure parameter, target background noise parameter, and target SNR parameter. If the new first deviation value, second deviation value, or third deviation value still does not fall within the preset threshold range, the testing platform continues to generate new gain correction parameters based on the updated error model and performs the next round of adjustments until the acoustic sound pressure, background noise sound pressure, and SNR all meet the requirements.

[0044] By using this channel-specific error modeling and cyclic feedback correction method, the artificial mouth and the noise simulator can each form stable output parameters, while ensuring that the sound field relationship after their superposition meets the test configuration requirements. This avoids the problem of deviation in the actual signal-to-noise ratio caused by calibrating only a single sound source, thus enabling the final determined target test sound field parameters to have good acoustic consistency, reproducibility, and metrological traceability.

[0045] Preferably, after adjusting the output gain parameter corresponding to the acoustic signal output unit according to the first deviation value, the second deviation value, and the third deviation value, and determining the target test sound field parameter when the first deviation value, the second deviation value, and the third deviation value all meet the preset threshold condition, the method further includes: Obtain the target output gain parameter of the acoustic signal output unit when the preset threshold condition is met; A standard sound source output profile is generated based on the target output gain parameters, and a mapping relationship is established between the standard sound source output profile and the target test sound field parameters. The standard sound source output file is invoked, and the sound signal output unit is controlled to output a reference calibration audio signal with a sound pressure level corresponding to the target test sound pressure parameter. The reference calibration audio signal is acquired by the acoustic calibration unit to obtain reference sound pressure data; By comparing the reference sound pressure data with the target test sound pressure parameter, when the deviation between the reference sound pressure data and the target test sound pressure parameter meets the preset calibration threshold condition, the current sound field is determined as the reference sound field.

[0046] Specifically, after the acoustic sound pressure, background noise, and signal-to-noise ratio are all adjusted to meet the preset threshold conditions, the test platform can further read the target output gain parameters of the acoustic signal output unit. These target output gain parameters may include the output level of the test audio output channel corresponding to the artificial mouth, the power amplifier gain, the channel gain of the noise output channel corresponding to the noise simulator, and other control parameters related to the current sound field configuration. Since these parameters are obtained through actual measurement feedback from the acoustic calibration unit and after multiple rounds of adjustment, they can reflect the actual output conditions required to form the target test sound field within the current test area.

[0047] The testing platform then generates a standard sound source output profile based on the target output gain parameters and establishes a mapping relationship between this profile and the target test sound field. This means that the software-side output control parameters are correspondingly saved with the acoustic sound pressure, background noise, and signal-to-noise ratio in the actual sound field. Subsequent testing of similar acoustic interactive terminals or use of the same test scenario can directly load this profile without needing to search for gain parameters from the initial state each time. To further confirm the reliability of this profile, the testing platform calls the standard sound source output profile to control the sound signal output unit to output a reference calibration audio signal with a preset sound pressure level. This reference calibration audio signal can be understood as a fixed test signal used to verify the standard sound source state of the artificial mouth; its sound pressure level is already determined by the target output gain parameters in the current profile. The sound level meter in the acoustic calibration unit then collects this reference calibration audio signal again at the standard measurement point to obtain the reference sound pressure data.

[0048] The testing platform compares the reference sound pressure data with the target test sound pressure parameters. If the deviation between the two meets the preset calibration threshold conditions, it means that the currently saved standard sound source output file can stably reproduce the target acoustic sound pressure. The output sound pressure of the artificial mouth has met the conditions to be used as the reference for subsequent testing or reverse calibration of the microphone array. At this time, the current sound field can be determined as the reference sound field.

[0049] This process not only completes the sound field configuration before a test, but also solidifies the qualified output parameters into a callable and traceable sound source file, enabling the test platform to quickly restore consistent acoustic conditions when testing different batches or switching between different product categories. It also provides a parameter basis for the dual-layer closed-loop calibration described in the disclosure document, which uses the sound level meter as the true value to calibrate the artificial mouth and noise simulator, and then uses the calibrated artificial mouth as a standard sound source to calibrate the microphone array.

[0050] Preferably, after comparing the reference sound pressure data with the target test sound pressure parameter, and determining the current sound field as the reference sound field when the deviation between the reference sound pressure data and the target test sound pressure parameter meets a preset calibration threshold condition, the method further includes: In the reference sound field, the sound field acquisition unit is controlled to acquire the reference calibration audio signal to obtain the corresponding sampling electrical signal; Based on the sound pressure conversion function of the sound field acquisition unit, the sampled electrical signal is converted into sampled sound pressure data; Calculate the sound pressure deviation between the sampled sound pressure data and the reference sound pressure data, wherein the reference sound pressure data is the standard reference sound pressure data measured by the acoustic calibration unit; The sensitivity correction coefficient of the corresponding channel of the sound field acquisition unit is calculated based on the sound pressure deviation value; Establish a mapping relationship between the sensitivity correction coefficient and the acquisition channel of the sound field acquisition unit, and store the sensitivity correction coefficient in the parameter library of the test platform.

[0051] Specifically, after the current sound field has been confirmed to be able to stably output the target sound pressure, the test platform can further use this reference sound field to calibrate the sensitivity of the sound field acquisition unit. The sound field acquisition unit here is typically a high-precision microphone array. Its function is not to generate sound, but to convert the sound pressure changes within the test area into sampled electrical signals that can be processed by the test platform. Due to potential differences in component batches, long-term aging, temperature and humidity drift, or gain differences in the front-end acquisition circuitry between different microphone channels, even under the same standard sound field, the sound pressure data calculated by different acquisition channels may deviate from the actual sound pressure.

[0052] Therefore, in the reference sound field, the test platform controls the sound signal output unit to continue outputting the aforementioned reference calibration audio signal, and controls the sound field acquisition unit to acquire the reference calibration audio signal to obtain the sampled electrical signal output by the corresponding channel. Subsequently, the test platform converts the sampled electrical signal into sampled sound pressure data according to the sound pressure conversion function of the sound field acquisition unit. The sound pressure conversion function can be understood as the calculation rule for the microphone channel to convert voltage amplitude, digital sampled value, or sound pressure level conversion relationship into sound pressure data. Since the reference sound pressure data is the standard reference sound pressure data measured by the sound level meter in the acoustic calibration unit under the same reference sound field, its reliability is higher than the sampling results of the uncalibrated microphone array. Therefore, the test platform uses the reference sound pressure data as a reference value to calculate the sound pressure deviation value between the sampled sound pressure data and the reference sound pressure data, and obtains the sensitivity correction coefficient of the corresponding acquisition channel accordingly. This sensitivity correction factor is used to compensate for measurement errors caused by the sound field acquisition unit's sensitivity being too high or too low during actual sound pickup. For example, when the sampled sound pressure calculated by a certain microphone channel is lower than the reference sound pressure measured by the sound level meter, the test platform can improve the subsequent sound pressure conversion result of that channel through the correction factor; when the sampled sound pressure is higher than the reference sound pressure, reverse compensation is performed through the correction factor.

[0053] After completing the calculations, the test platform establishes a mapping relationship between the sensitivity correction coefficients and the corresponding sound field acquisition unit acquisition channels, and stores this information in the parameter library. This allows the system to automatically load the corresponding correction coefficients according to the channels during subsequent formal testing, correcting the audio feedback data acquired by the microphone array. Through this process, the sound pickup results of the sound field acquisition unit no longer rely solely on the factory default sensitivity, but are corrected through reference data from the benchmark sound field and sound level meter. This helps reduce the impact of microphone array discrete errors, aging drift, and environmental changes on the audio evaluation results, resulting in higher accuracy and traceability when subsequently judging MOS scores, acoustic feedback quality, or sound field consistency based on audio feedback data.

[0054] Preferably, before controlling the sound signal output unit to output the test audio signal corresponding to the test corpus data and the environmental noise signal corresponding to the environmental noise data based on the target test sound field parameters, so as to construct the target test sound field corresponding to the acoustic interaction terminal under test, and triggering the acoustic interaction terminal under test to execute the corresponding acoustic interaction process, the method further includes: Acquire the device clock information corresponding to the acoustic signal output unit, the acoustic field acquisition unit, and the status data acquisition unit; Based on the global time reference corresponding to the device clock information, a unified clock synchronization relationship is established between the sound signal output unit, the sound field acquisition unit, and the status data acquisition unit. Based on the timing control parameters corresponding to the unified clock synchronization relationship, test the execution timing of audio signal output, environmental noise signal output, audio feedback data acquisition, and running status data acquisition, and generate corresponding timing control results. Based on the timing control results, each hardware unit is controlled to execute corresponding automated test tasks.

[0055] Specifically, before officially outputting the test audio signal and ambient noise signal, the test platform first acquires the device clock information corresponding to the sound signal output unit, sound field acquisition unit, and status data acquisition unit to confirm the current time reference used by the hardware such as the artificial mouth, noise simulator, high-precision microphone array, and multi-standard data acquisition card. Since different hardware units may have independent control clocks or sampling clocks, directly starting them separately can easily lead to problems such as acoustic playback occurring before audio acquisition begins, inconsistencies between noise superposition time and acoustic time, or inaccurate matching between running status messages and corresponding speech data. Therefore, the test platform needs to establish a unified clock synchronization relationship based on the clock information of each device and generate a global time reference on this basis. The global time reference can be understood as a time reference followed by the entire automated test task. Subsequent arrangements regarding when the acoustic output occurs, when ambient noise is superimposed, when the sound field acquisition unit starts recording audio, and when the status data acquisition unit starts listening to the interactive messages of the terminal under test are all based on this global time reference. The testing platform further generates timing control parameters based on the global time base. These timing control parameters may include the acoustic playback start time, noise playback advance or delay, audio acquisition start time, status data monitoring window, duration of a single test case, acquisition end time, and interval between different test cases.

[0056] Subsequently, the testing platform uses these timing control parameters to orchestrate the output of test audio signals, environmental noise signals, audio feedback data acquisition, and operational status data acquisition, generating corresponding timing control results. Based on these results, it controls each hardware unit to execute automated testing tasks. Through this processing, the test acoustics played by the artificial mouth, the background noise output by the noise simulator, the audio feedback data acquired by the high-precision microphone array, and the operational status data obtained by the multi-standard data acquisition card can establish a corresponding relationship on the same timeline. This allows the testing platform to accurately determine the causal relationship between a communication message, a segment of audio feedback, and a piece of test data, avoiding misjudgments of interaction delays, incorrect statistical identification results, or the inability to trace abnormal data due to hardware startup asynchrony. This improves the repeatability of the automated testing process and the reliability of the result analysis.

[0057] Preferably, the step of acquiring the operating status data of the acoustic interaction terminal under test through the status data acquisition unit during the acoustic interaction process, and acquiring the audio feedback data output by the acoustic interaction terminal under test through the sound field acquisition unit includes: Identify the physical communication interface type of the acoustic interactive terminal under test, and load a communication protocol stack and message parsing rules that match the physical communication interface type. The physical communication interface type includes CAN bus interface, RS232 / RS485 serial port interface and Ethernet interface. A low-level data connection is established between the status data acquisition unit and the acoustic interaction terminal under test based on the communication protocol stack. Listen to and capture the original interactive messages generated by the acoustic interactive terminal under test during the response to test commands; The original interactive message is decoded according to the message parsing rules to extract running status data including wake-up flag, acoustic recognition text, instruction execution status code and timestamp; Simultaneously, the sound field acquisition unit acquires the audio feedback data output by the acoustic interactive terminal under test.

[0058] Specifically, when collecting operational status data of the acoustic interactive terminal under test, the test platform first identifies the actual physical communication interface type provided by the terminal and loads the matching communication protocol stack and message parsing rules accordingly. The physical communication interface type refers to the hardware interface used by the acoustic interactive terminal under test when interacting with external test equipment. Vehicle-mounted acoustic cockpits typically use a CAN bus interface, while smart home control systems, smart customer service devices, or other terminals may use RS232, RS485 serial ports, or Ethernet interfaces depending on the product design. The communication protocol stack ensures that the test platform can complete handshake, data reception, and link maintenance according to the communication rules of the corresponding interface, while the message parsing rules explain the meaning of different fields in the original interactive messages.

[0059] After the protocol loading is completed, the status data acquisition unit establishes a low-level data connection with the acoustic interaction terminal under test based on the corresponding communication protocol stack, and continuously listens to the raw interaction messages generated by the terminal under test during the acoustic test. The raw interaction messages here are not the interface results observed manually, but rather the low-level operational data generated by the terminal under test when receiving test commands, performing wake-up recognition, performing semantic processing, and outputting control feedback, which can more directly reflect the internal interaction state of the terminal.

[0060] The testing platform decodes the original interactive messages according to the loaded message parsing rules, extracting operational status data such as wake-up flags, acoustic recognition text, command execution status codes, and timestamps. The wake-up flags indicate whether the terminal has been successfully woken up; the acoustic recognition text characterizes the acoustic content actually recognized by the terminal; the command execution status code determines whether the corresponding control command was executed successfully; and the timestamp calculates the interaction delay between acoustic input and terminal feedback. Simultaneously, the sound field acquisition unit synchronously acquires the audio feedback data output by the acoustic interaction terminal under test. This allows the testing platform to obtain both the terminal's internal state from the communication messages and acoustic feedback such as external acoustics or prompts from the audio acquisition results. Through this parallel acquisition method of communication and acoustic links, the system is compatible with products using different hardware protocols, such as automotive, smart home, smart customer service, and mobile terminals. By using underlying messages and audio data as the basis for automatic judgment, it reduces errors caused by manual listening, screen viewing, and manual recording, improving the objectivity and traceability of test results.

[0061] Preferably, the step of analyzing the test results based on the operating status data, the audio feedback data, and the test evaluation index data to obtain the corresponding test results includes: Acquire the running status data, the audio feedback data, the test corpus data, and the test evaluation index data; Based on the test corpus data and the test evaluation index data, determine the corresponding expected recognition results, expected execution results, and evaluation rules; The identification result data in the running status data is matched and analyzed with the expected identification result to obtain the identification judgment result; The execution result data in the running status data is matched and analyzed with the expected execution result to obtain the execution determination result; Calculate the response latency parameters based on the timestamp data and the playback time data corresponding to the test audio signal, and generate the response latency evaluation results; A wake-up evaluation result is generated based on the wake-up status data in the operation status data; Based on the audio feedback data, audio quality evaluation parameters are extracted, and audio quality evaluation results are generated. A comprehensive evaluation calculation is performed based on the recognition judgment result, the execution judgment result, the response latency evaluation result, the wake-up evaluation result, and the audio quality evaluation result, and the corresponding test result is output.

[0062] Specifically, during test result analysis, the testing platform first acquires the operational status data uploaded by the status data acquisition unit, the audio feedback data collected by the sound field acquisition unit, and the test evaluation index data bound to the current test task, and then associates these three types of data according to the same test case. The test evaluation index data is used to provide the expected interaction results that the test case should achieve. Among them, the expected recognition result can be the text content or semantic instructions that the acoustic interaction terminal under test should recognize after receiving a certain test audio signal; the expected execution result can be the control feedback, functional status change, or execution success flag that the terminal should output; and the evaluation rules are used to define how the results of recognition, execution, wake-up, latency, and audio quality are judged as pass, fail, or abnormal.

[0063] The testing platform matches the recognition results in the operational status data with the expected recognition results to determine whether the tested terminal correctly understands the test acoustic content. It then matches the execution results in the operational status data with the expected execution results to determine whether the terminal completed the corresponding instruction action after recognition. For interaction latency, the testing platform uses timestamp data from the operational status data, combined with the playback time data of the test audio signal, to calculate the response latency parameter from acoustic playback, terminal wake-up, recognition completion, to execution feedback, and then determines whether it exceeds the latency range allowed in the evaluation rules. For wake-up performance, the testing platform generates a wake-up evaluation result based on the wake-up status data in the operational status data to distinguish between normal wake-up, no wake-up, false wake-up, or repeated wake-up.

[0064] At the same time, the testing platform can also process the audio feedback data and extract audio quality evaluation parameters, such as the clarity of the external acoustics of the tested terminal, volume stability, abnormal noise, broadcast integrity or MOS score, etc., to determine whether the terminal has not only completed the interaction, but also completed the feedback with the required acoustic quality.

[0065] Finally, the testing platform comprehensively evaluates and calculates the recognition and judgment results, execution judgment results, response latency evaluation results, wake-up evaluation results, and audio quality evaluation results. According to the evaluation rules, it outputs a pass, failure, or exception conclusion for each test case. After multiple test cases are completed, it further generates test results such as wake-up rate, recognition rate, command execution success rate, average response latency, and audio quality statistics. Through this analysis method, test conclusions no longer rely on testers manually listening, viewing screens, or manually compiling statistics. Instead, they are supported by underlying runtime messages, audio acquisition results, and preset evaluation rules, improving the consistency and objectivity of test judgments and facilitating the tracing of the original data basis for each conclusion in the test report.

[0066] Preferably, the step of calculating the response delay parameter based on the timestamp data and the playback time data corresponding to the test audio signal, and generating the response delay evaluation result, includes: Obtain the playback start timestamp and playback end timestamp corresponding to the test audio signal; Extract the node timestamps corresponding to multiple processing nodes when the acoustic interactive terminal under test performs the interactive response process from the running status data, wherein the multiple processing nodes include at least two of the following: wake-up completion node, recognition completion node, instruction issuance node, and instruction execution completion node; The audio feedback data is used to detect valid audio segments and determine the feedback start timestamp corresponding to when the acoustic interaction terminal under test starts outputting audio feedback. Based on a unified time reference, the playback start timestamp, the playback end timestamp, the timestamps of each node, and the feedback start timestamp are time-aligned, and an interactive event sequence is generated in chronological order. Based on the time difference between adjacent events in the interactive event sequence, the stage response delay corresponding to multiple interactive stages is calculated respectively, and the total response delay is calculated based on the time difference between the playback timestamp and the feedback start timestamp. The response latency of each stage and the total response latency are compared with the corresponding preset latency thresholds to determine the target interaction stage with latency anomalies, and the response latency evaluation result is generated based on the target interaction stage and the corresponding stage response latency.

[0067] Specifically, when the test platform sends out the test audio signal, it simultaneously records the playback start timestamp and the playback end timestamp. The former is used to characterize the moment when the audio signal output unit starts to output the test content to the test area, and the latter is used to characterize the moment when the complete playback of the test content ends. Both can be directly generated by the audio playback interface, the sound card driver layer, or the audio signal output control module and written into the timing record of the current test case.

[0068] After receiving the test audio signal, the acoustic interaction terminal under test will sequentially complete the wake-up detection, content recognition, command generation or issuance, command execution, and feedback output. The status data acquisition unit can listen to the status messages generated inside the terminal through the CAN bus, serial port, or Ethernet, and extract the node timestamp when each processing node is completed from the corresponding flag bits, status codes, or operation logs. The node timestamp is the time data used to mark the time when a certain internal processing event occurs. The specific nodes selected can be determined based on the status information that the terminal under test can open and the current test indicators. It is not required that all terminals provide the exact same nodes.

[0069] At the same time, the sound field acquisition unit continuously acquires audio feedback data from the terminal's external speaker. The test platform can set an energy threshold based on the pre-measured ambient noise level and combine it with continuous frame energy, spectrum changes, silence duration or effective content duration to identify the audio frame when the terminal actually starts to emit sound. The acquisition time corresponding to this audio frame is determined as the feedback start timestamp to avoid misjudging ambient noise, background noise, short pulses or device startup noise as the start of feedback.

[0070] Since the audio signal output unit, status data acquisition unit, and sound field acquisition unit may use their own local clocks, the test platform needs to perform clock offset and transmission delay compensation on each timestamp based on a pre-established global time reference. This ensures that playback events, terminal internal processing events, and feedback output events can be mapped to the same timeline, forming an interactive event sequence according to their actual occurrence order. Therefore, by subtracting adjacent event timestamps, the stage response delay consumed in different stages such as wake-up processing, content recognition, command issuance, command execution, and feedback output can be obtained. Based on the test evaluation rules, either the playback start timestamp or playback end timestamp is selected as the timing starting point for the total delay, and the time difference between it and the feedback start timestamp yields the total response delay.

[0071] The testing platform compares the preset latency thresholds corresponding to each processing stage and the overall interaction process. When the actual latency of a certain stage exceeds the corresponding threshold, the stage is marked as the target interaction stage, and information such as the actual latency, allowable threshold, excess amount, and corresponding test cases are recorded to form a response latency evaluation result. This makes the test results no longer limited to judging whether the overall terminal response times out, but can further determine whether the latency is caused by the terminal wake-up, recognition, instruction processing, execution control, or feedback output, which is convenient for locating the performance bottleneck and abnormal cause of the terminal under test.

[0072] Preferably, the step of performing a comprehensive evaluation calculation based on the recognition judgment result, the execution judgment result, the response latency evaluation result, the wake-up evaluation result, and the audio quality evaluation result, and outputting the corresponding test result includes: Based on the identification judgment result, execution judgment result, response latency evaluation result and wake-up evaluation result corresponding to the running status data, a first test judgment result is generated; Based on the audio content matching results and audio quality evaluation results corresponding to the audio feedback data, a second test judgment result is generated; Obtain at least one of the message integrity, message verification status and timestamp continuity corresponding to the running status data, and determine the first data credibility corresponding to the first test judgment result; Obtain at least one of the signal-to-noise ratio, effective audio segment completeness, and audio content recognition confidence corresponding to the audio feedback data, and determine the second data credibility corresponding to the second test judgment result; By comparing the first test judgment result and the second test judgment result, when the first test judgment result is consistent with the second test judgment result, the two are fused according to the first data credibility and the second data credibility to obtain the test result; When the first test result is inconsistent with the second test result, the source of the data anomaly is determined based on the first data credibility and the second data credibility. When the data credibility corresponding to one of the test judgment results does not meet the preset credibility condition, the other test judgment result that meets the preset credibility condition is determined as the test result; When both the first data credibility and the second data credibility meet the preset credibility condition, the current test case is marked as a decision conflict test case, the acoustic interaction terminal under test is controlled to re-execute the decision conflict test case, and the test result is determined based on the re-collected running status data and audio feedback data.

[0073] Specifically, the testing platform can process the operating status data obtained from the internal acoustic interactive terminal under test and the audio feedback data obtained from the external sound field acquisition unit as two independent judgment links. The first test judgment result mainly reflects whether the internal processing of the terminal meets expectations. It can be generated by combining information such as the matching of the recognized text with the expected recognition result, the consistency of the instruction execution status code with the expected execution result, whether the response delay of each processing node exceeds the corresponding threshold, and whether the wake-up flag is valid. The second test judgment result mainly reflects whether the actual external feedback output of the terminal meets the requirements. It can first remove silent segments, extract effective segments, and recognize the content of the audio feedback data. Then, it matches the recognized feedback content with the expected feedback content corresponding to the current test case and determines the result by combining audio quality parameters such as loudness, distortion, clarity, signal-to-noise ratio, or MOS evaluation value.

[0074] To avoid distortion of a judgment link due to packet loss, message anomalies, or environmental noise interference, the test platform also needs to evaluate the reliability of the two types of data separately. Message completeness can be determined based on the ratio between the expected number of messages and the actual number received, whether key fields are missing, and whether the beginning and end of the message are complete. Message verification status can be determined based on checksum, cyclic redundancy check code, or protocol frame verification results. Timestamp continuity is used to determine whether there are abnormal jumps, repetitions, or long periods of absence between adjacent state messages. The platform can calculate the first data credibility based on the above indicators. For audio feedback data, the signal-to-noise ratio can be determined based on the ratio between the effective audio signal power and the background noise power. The audio segment completeness can be determined based on whether the effective segment contains a complete start segment, main content, and end segment. The audio content recognition confidence can be determined based on the probability value, similarity, or confidence score output by the content recognition model, thus obtaining the second data credibility. When the two judgment results are consistent, it indicates that the internal state of the terminal and the actual external performance corroborate each other. At this time, the fusion weight can be set according to the credibility of the two types of data. For example, data with higher credibility will have a larger weight in the comprehensive result, and when both meet the credibility requirements, the pass, failure or corresponding score result will be directly output. When the two are inconsistent, the test platform will not directly take either result as the final conclusion, but will first compare the credibility of the two types of data and locate the possible sources of data anomalies. For example, if there is frame loss, verification failure or timestamp interruption in the internal message, it can be considered that the reliability of the first judgment link is insufficient. If the audio data has a low signal-to-noise ratio, segment truncation or low recognition confidence, it can be considered that the second judgment link is interfered with. If only one type of data fails to meet the preset credibility condition, the judgment result of the other type that meets the condition can be used as the current test result, and the abnormal reason of the ignored data can be recorded simultaneously. If both types of data have high credibility but the conclusions still conflict, it means that the conflict is more likely to be due to the inconsistency between the status feedback of the terminal under test and the actual output. For example, the internal status code shows that the execution was successful but no corresponding feedback was detected externally, or the terminal has actually output the correct content but the internal message has not been updated. In this case, the current test case is marked as a judgment conflict test case, and the test is re-executed while keeping the original test data, environmental noise, target sound field parameters and device connection relationship unchanged. The two channels of data are re-collected for verification. If necessary, a preset number of retries can be set or the final conclusion can be determined by the consistency of multiple results.

[0075] This evaluation method, which verifies both internal status and external audio, can reduce misjudgments of test results caused by anomalies in a single data source. It can also distinguish between actual terminal function failures and test data acquisition failures, thereby improving the reliability of automatic judgment results and the efficiency of subsequent fault location.

[0076] Example 2 Please see Figure 3This invention provides an automated acoustic testing device for intelligent interactive terminals, applied to an automated acoustic testing system for intelligent interactive terminals. The system includes a testing platform, an acoustic signal output unit, a sound field acquisition unit, an acoustic calibration unit, a status data acquisition unit, and an acoustic interactive terminal under test. The testing platform is communicatively connected to the acoustic signal output unit, the sound field acquisition unit, the acoustic calibration unit, and the status data acquisition unit, respectively. The status data acquisition unit is communicatively connected to the acoustic interactive terminal under test. The device includes: The data acquisition module is used to acquire the test configuration data corresponding to the acoustic interactive terminal to be tested, wherein the test configuration data includes test corpus data, environmental noise data and test evaluation index data; The scene configuration module is used to configure the test scene based on the test configuration data through the acoustic calibration unit, and determine the target test sound field parameters corresponding to the acoustic interaction terminal to be tested. The interactive triggering module is used to control the sound signal output unit to output the test audio signal corresponding to the test corpus data and the environmental noise signal corresponding to the environmental noise data based on the target test sound field parameters, so as to construct the target test sound field corresponding to the acoustic interactive terminal under test, and trigger the acoustic interactive terminal under test to execute the corresponding acoustic interaction process. The data acquisition module is used to acquire the operating status data of the acoustic interaction terminal under test during the acoustic interaction process through the status data acquisition unit, and to acquire the audio feedback data output by the acoustic interaction terminal under test through the sound field acquisition unit. The test result analysis module is used to analyze the test results based on the running status data, the audio feedback data, and the test evaluation index data to obtain the corresponding test results.

[0077] It should be noted that each module and unit in the automated acoustic testing device for intelligent interactive terminals in this embodiment corresponds one-to-one with each step in the automated acoustic testing method for intelligent interactive terminals in the aforementioned embodiment. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned automated acoustic testing method for intelligent interactive terminals, and will not be repeated here.

[0078] Example 3 In addition, combined Figure 1 The automated acoustic testing method for intelligent interactive terminals described in this embodiment of the invention can be implemented by testing equipment. Figure 4 A schematic diagram of the hardware structure of the test equipment provided in an embodiment of the present invention is shown.

[0079] The testing equipment may include a processor and a memory storing computer program instructions.

[0080] Specifically, the processor may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement embodiments of the present invention.

[0081] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0082] Computer-readable media include both permanent and non-permanent, removable and non-removable media, which can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient media, such as modulated communication signals and carrier waves.

[0083] The processor reads and executes computer program instructions stored in memory to implement any of the above-described automated acoustic testing methods for intelligent interactive terminals.

[0084] In one example, the test device may also include a communication interface and a bus. For example, Figure 4 As shown, the processor 401, memory 402, and communication interface 403 are connected through bus 410 and complete communication with each other.

[0085] The communication interface is mainly used to enable communication between various modules, devices, units and / or equipment in the embodiments of the present invention.

[0086] A bus, including hardware, software, or both, couples components of a test device together. For example, and not limitingly, a bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, a bus may include one or more buses. While specific buses are described and illustrated in embodiments of the invention, the invention contemplates any suitable bus or interconnect.

[0087] In summary, the embodiments of the present invention provide an automated acoustic testing method, apparatus, and device for intelligent interactive terminals.

[0088] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0089] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0090] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0091] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0092] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0093] It should also be noted that the exemplary embodiments mentioned in this invention describe methods or systems based on a series of steps or apparatus. However, this invention is not limited to the order of the steps described above; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0094] The above description is merely a specific embodiment of the present invention. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the protection scope of the present invention.

Claims

1. An automated acoustic testing method for intelligent interactive terminals, characterized in that, An automated acoustic testing system for intelligent interactive terminals is provided. The system includes a test platform, an acoustic signal output unit, a sound field acquisition unit, an acoustic calibration unit, a status data acquisition unit, and an acoustic interactive terminal under test. The test platform is communicatively connected to the acoustic signal output unit, the sound field acquisition unit, the acoustic calibration unit, and the status data acquisition unit. The status data acquisition unit is communicatively connected to the acoustic interactive terminal under test. The method includes: Obtain the test configuration data corresponding to the acoustic interactive terminal to be tested, wherein the test configuration data includes test corpus data, environmental noise data, and test evaluation index data; Based on the test configuration data, the test scenario is configured through the acoustic calibration unit to determine the target test sound field parameters corresponding to the acoustic interaction terminal to be tested. Based on the target test sound field parameters, the sound signal output unit is controlled to output the test audio signal corresponding to the test corpus data and the environmental noise signal corresponding to the environmental noise data, so as to construct the target test sound field corresponding to the acoustic interaction terminal under test and trigger the acoustic interaction terminal under test to execute the corresponding acoustic interaction process. The status data acquisition unit collects the operating status data of the acoustic interaction terminal under test during the acoustic interaction process, and the sound field acquisition unit collects the audio feedback data output by the acoustic interaction terminal under test. Based on the operational status data, the audio feedback data, and the test evaluation index data, the test results are analyzed to obtain the corresponding test results.

2. The automated acoustic testing method for intelligent interactive terminals according to claim 1, characterized in that, The step of configuring the test scene based on the test configuration data and determining the target test sound field parameters corresponding to the acoustic interaction terminal under test through the acoustic calibration unit includes: Obtain the target test sound pressure level, target background noise, and target signal-to-noise ratio parameters corresponding to the test configuration data; The control sound signal output unit outputs a preset calibration audio signal and a preset calibration noise signal; The actual test sound pressure data and actual background noise data in the test area are collected through the acoustic calibration unit; Calculate the actual signal-to-noise ratio parameter based on the actual test sound pressure data and actual background noise data; Calculate the first deviation between the actual test sound pressure data and the target test sound pressure parameter, the second deviation between the actual background noise data and the target background noise parameter, and the third deviation between the actual signal-to-noise ratio parameter and the target signal-to-noise ratio parameter, respectively. The output gain parameter of the acoustic signal output unit is adjusted according to the first deviation value, the second deviation value and the third deviation value. When the first deviation value, the second deviation value and the third deviation value all meet the preset threshold condition, the target test acoustic field parameter is determined.

3. The automated acoustic testing method for intelligent interactive terminals according to claim 2, characterized in that, The step of adjusting the output gain parameter of the acoustic signal output unit according to the first deviation value, the second deviation value, and the third deviation value, and determining the target test sound field parameter when the first deviation value, the second deviation value, and the third deviation value all meet the preset threshold condition, includes: Establish a test sound pressure error model between the target test sound pressure parameter and the actual test sound pressure data, and a noise sound pressure error model between the target background noise parameter and the actual background noise data; Based on the test sound pressure error model, calculate the first gain correction parameter corresponding to the test audio output channel; Based on the noise sound pressure error model, calculate the second gain correction parameter corresponding to the noise output channel; Using the first gain correction parameter, adjust the output gain parameter corresponding to the test audio signal, and using the second gain correction parameter, adjust the output gain parameter corresponding to the ambient noise signal. The new actual test sound pressure data and new actual background noise data are re-acquired and adjusted. The process returns to the step of calculating the actual signal-to-noise ratio parameter based on the actual test sound pressure data and actual background noise data, until the first deviation value, the second deviation value and the third deviation value all meet the preset threshold conditions, and then the target test sound field parameter is determined.

4. The automated acoustic testing method for intelligent interactive terminals according to claim 2, characterized in that, After adjusting the output gain parameter corresponding to the acoustic signal output unit according to the first deviation value, the second deviation value, and the third deviation value, and determining the target test acoustic field parameter when the first deviation value, the second deviation value, and the third deviation value all meet the preset threshold condition, the method further includes: Obtain the target output gain parameter of the acoustic signal output unit when the preset threshold condition is met; A standard sound source output profile is generated based on the target output gain parameters, and a mapping relationship is established between the standard sound source output profile and the target test sound field parameters. The standard sound source output file is invoked, and the sound signal output unit is controlled to output a reference calibration audio signal with a sound pressure level corresponding to the target test sound pressure parameter. The reference calibration audio signal is acquired by the acoustic calibration unit to obtain reference sound pressure data; By comparing the reference sound pressure data with the target test sound pressure parameter, when the deviation between the reference sound pressure data and the target test sound pressure parameter meets the preset calibration threshold condition, the current sound field is determined as the reference sound field.

5. The automated acoustic testing method for intelligent interactive terminals according to claim 4, characterized in that, After comparing the reference sound pressure data with the target test sound pressure parameter, and determining the current sound field as the reference sound field when the deviation between the reference sound pressure data and the target test sound pressure parameter meets a preset calibration threshold condition, the process further includes: In the reference sound field, the sound field acquisition unit is controlled to acquire the reference calibration audio signal to obtain the corresponding sampling electrical signal; Based on the sound pressure conversion function of the sound field acquisition unit, the sampled electrical signal is converted into sampled sound pressure data; Calculate the sound pressure deviation between the sampled sound pressure data and the reference sound pressure data, wherein the reference sound pressure data is the standard reference sound pressure data measured by the acoustic calibration unit; The sensitivity correction coefficient of the corresponding channel of the sound field acquisition unit is calculated based on the sound pressure deviation value; Establish a mapping relationship between the sensitivity correction coefficient and the acquisition channel of the sound field acquisition unit, and store the sensitivity correction coefficient in the parameter library of the test platform.

6. The automated acoustic testing method for intelligent interactive terminals according to claim 1, characterized in that, Before controlling the sound signal output unit to output the test audio signal corresponding to the test corpus data and the environmental noise signal corresponding to the environmental noise data based on the target test sound field parameters, so as to construct the target test sound field corresponding to the acoustic interaction terminal under test, and trigger the acoustic interaction terminal under test to execute the corresponding acoustic interaction process, the method further includes: Acquire the device clock information corresponding to the acoustic signal output unit, the acoustic field acquisition unit, and the status data acquisition unit; Based on the global time reference corresponding to the device clock information, a unified clock synchronization relationship is established between the sound signal output unit, the sound field acquisition unit, and the status data acquisition unit. Based on the timing control parameters corresponding to the unified clock synchronization relationship, test the execution timing of audio signal output, environmental noise signal output, audio feedback data acquisition, and running status data acquisition, and generate corresponding timing control results. Based on the timing control results, each hardware unit is controlled to execute corresponding automated test tasks.

7. The automated acoustic testing method for intelligent interactive terminals according to claim 1, characterized in that, The step of acquiring the operating status data of the acoustic interaction terminal under test through the status data acquisition unit during the acoustic interaction process, and acquiring the audio feedback data output by the acoustic interaction terminal under test through the sound field acquisition unit includes: Identify the physical communication interface type of the acoustic interactive terminal under test, and load a communication protocol stack and message parsing rules that match the physical communication interface type. The physical communication interface type includes CAN bus interface, RS232 / RS485 serial port interface and Ethernet interface. A low-level data connection is established between the status data acquisition unit and the acoustic interaction terminal under test based on the communication protocol stack. Listen to and capture the original interactive messages generated by the acoustic interactive terminal under test during the response to test commands; The original interactive message is decoded according to the message parsing rules to extract running status data including wake-up flag, acoustic recognition text, instruction execution status code and timestamp; Simultaneously, the sound field acquisition unit acquires the audio feedback data output by the acoustic interactive terminal under test.

8. The automated acoustic testing method for intelligent interactive terminals according to any one of claims 1-7, characterized in that, The test results obtained by analyzing the operating status data, the audio feedback data, and the test evaluation index data include: Acquire the running status data, the audio feedback data, the test corpus data, and the test evaluation index data; Based on the test corpus data and the test evaluation index data, determine the corresponding expected recognition results, expected execution results, and evaluation rules; The identification result data in the running status data is matched and analyzed with the expected identification result to obtain the identification judgment result; The execution result data in the running status data is matched and analyzed with the expected execution result to obtain the execution determination result; Calculate the response latency parameters based on the timestamp data and the playback time data corresponding to the test audio signal, and generate the response latency evaluation results; A wake-up evaluation result is generated based on the wake-up status data in the operation status data; Based on the audio feedback data, audio quality evaluation parameters are extracted, and audio quality evaluation results are generated. A comprehensive evaluation calculation is performed based on the recognition judgment result, the execution judgment result, the response latency evaluation result, the wake-up evaluation result, and the audio quality evaluation result, and the corresponding test result is output.

9. An automated acoustic testing device for an intelligent interactive terminal, characterized in that, An automated acoustic testing system for intelligent interactive terminals is provided. The system includes a test platform, an acoustic signal output unit, a sound field acquisition unit, an acoustic calibration unit, a status data acquisition unit, and an acoustic interactive terminal under test. The test platform is communicatively connected to the acoustic signal output unit, the sound field acquisition unit, the acoustic calibration unit, and the status data acquisition unit. The status data acquisition unit is communicatively connected to the acoustic interactive terminal under test. The device includes: The data acquisition module is used to acquire the test configuration data corresponding to the acoustic interactive terminal to be tested, wherein the test configuration data includes test corpus data, environmental noise data and test evaluation index data; The scene configuration module is used to configure the test scene based on the test configuration data through the acoustic calibration unit, and determine the target test sound field parameters corresponding to the acoustic interaction terminal to be tested. The interactive triggering module is used to control the sound signal output unit to output the test audio signal corresponding to the test corpus data and the environmental noise signal corresponding to the environmental noise data based on the target test sound field parameters, so as to construct the target test sound field corresponding to the acoustic interactive terminal under test, and trigger the acoustic interactive terminal under test to execute the corresponding acoustic interaction process. The data acquisition module is used to acquire the operating status data of the acoustic interaction terminal under test during the acoustic interaction process through the status data acquisition unit, and to acquire the audio feedback data output by the acoustic interaction terminal under test through the sound field acquisition unit. The test result analysis module is used to analyze the test results based on the running status data, the audio feedback data, and the test evaluation index data to obtain the corresponding test results.

10. A testing device, characterized in that, include: At least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the method as described in any one of claims 1-8.

Images