Intelligent learning machine seating wake-up pressure measurement method and system

Abstract

The embodiment of the invention provides an intelligent learning machine seating wake-up pressure test method and system, and the method can respond to a test instruction, generates a test plan, controls a mechanical arm to execute a seating simulation action according to the test plan, and synchronously collects test data. And obtaining an ideal data logic chain of a seating track mode in the test plan, thereby generating pressure test result information according to the ideal data logic chain and the test data. Wherein the pressure test result information comprises a seating wake-up result generated according to the screen state event and a key test event determined by comparing the test data with the ideal data logic chain. According to the method, a mechanical arm can be used for executing multiple pre-stored seating track modes to simulate the seating behavior of a user, and multi-dimensional test data are combined to perform seating wake-up test and root cause analysis, so that multi-modal data cooperative verification and adaptive test strategy adjustment are realized, logic defects in the wake-up process are accurately found, and the test efficiency is improved. And the accuracy of the test result and the test efficiency are improved.

Description

Technical Field

[0001] This application relates to the field of intelligent educational equipment technology, and in particular to a method and system for stress testing the wake-up function of an intelligent learning machine when the user sits down. Background Technology

[0002] A smart learning machine is an electronic device designed for educational information interaction, utilizing computer, internet, automation, and artificial intelligence technologies. It includes a processor, storage module, display module, input / output module, and power module. The input / output module incorporates various sensors, such as cameras, proximity sensors, and infrared sensors, to handle data input, output, and device connectivity.

[0003] The seat-activated wake-up function is an intelligent feature of smart learning machines designed to enhance the user experience. It utilizes built-in cameras, proximity sensors, and other detection components to detect whether a user is seated in the designated seating area in front of the machine. Upon detection, the screen automatically illuminates to facilitate user interaction. As an intelligent interaction method, the seat-activated wake-up function requires stress testing before official application to maintain a high probability of successful wake-up and minimize false wake-ups.

[0004] Because manual seat-based wake-up stress testing is inefficient, costly, and lacks consistency, automated seat-based wake-up stress testing can be performed using robotic arms and a fixed frame. This involves mounting the smart learning machine on the frame and using a robotic arm to simulate the user's spatial movement during the seat-based process, recording the number of times the machine is woken up and the wake-up probability to calculate the pass / fail rate of the seat-based wake-up function. However, this automated testing method cannot simulate various user seat-based behaviors, resulting in a simplistic testing logic that cannot handle complex scenarios. Furthermore, when tests fail, there is a lack of effective means to quickly pinpoint the problem's origin—whether it stems from hardware sensors, data fusion algorithms, or upper-level application logic—leading to lengthy troubleshooting cycles and inaccurate test results. Summary of the Invention

[0005] In view of this, embodiments of this application provide a method and system for stress testing the seat-and-wake function of an intelligent learning machine, in order to solve the problem of inaccurate stress test results for the seat-and-wake function.

[0006] According to a first aspect of this application, a method for stress testing the wake-up function of a smart learning machine upon seating is provided, the method comprising: In response to a test command, a test plan is generated, the test plan including at least one seating trajectory pattern from a pre-stored pattern library; According to the test plan, control the robotic arm to perform a seating simulation action; During the process of the robotic arm performing the sitting simulation action, test data is collected synchronously. The test data includes the real-time spatial coordinates of the robotic arm, the continuous distance readings of the position sensor, the target confidence of the head sensor, the ambient light value of the ambient light sensor, and the screen status events of the learning machine under test, all recorded synchronously with a unified timestamp. Obtain the ideal data logic chain for the seating trajectory pattern described in the test plan; The stress test results are generated based on the ideal data logic chain and the test data. The stress test results include seat wake-up results and key test events. The seat wake-up results are generated based on the screen state events. The key test events are determined by comparing the test data with the ideal data logic chain.

[0007] In some embodiments, the method further includes: Obtain multiple load test results within a preset test period; Frequent test items are read from multiple stress test result information, wherein the frequent test items are the seat wake-up results and / or the key test events that occur more than or equal to the number of occurrences of a threshold. The test plan is dynamically adjusted based on the frequently tested items.

[0008] In some embodiments, dynamically adjusting the test plan based on the frequent test items includes: Based on the frequent test items, test items to be adjusted are matched in the pre-stored pattern library. The test items to be adjusted include test items to be added and / or test items to be deleted. The test items to be added are determined based on abnormally frequent test items whose occurrence frequency is greater than or equal to a frequency threshold. The test items to be deleted are determined based on normally frequent test items whose occurrence frequency is greater than or equal to a frequency threshold. Modify the test plan according to the test items to be adjusted, and generate an updated trajectory pattern based on the modified test plan; Add the updated trajectory pattern to the pre-stored pattern library.

[0009] In some embodiments, according to the test plan, controlling the robotic arm to perform a seat-down simulation action includes: Read the seating trajectory pattern from the test plan; Based on the seating trajectory pattern, at least one decomposed action parameter is set, and the action parameter includes at least one of the action orientation, action speed, pause parameter and number of cycles associated with the seating trajectory pattern. Generate trajectory simulation commands based on the motion parameters, and generate a set of control commands by combining the trajectory simulation commands; According to the set of control commands, the trajectory simulation commands are sent to the robotic arm in sequence to control the robotic arm to perform a sitting simulation action according to the trajectory simulation commands.

[0010] In some embodiments, at least one decomposed action parameter is set based on the seating trajectory pattern, including: If the seating trajectory pattern includes the standard seating pattern, the movement direction is set to a smooth movement from the upper corner of the learning machine under test to the detection area directly in front, and the movement speed is a first speed. If the sitting trajectory mode includes a rapid impact mode, the action speed is set to a second speed, which is greater than the first speed; If the seating trajectory mode includes a hesitant approach mode, the pause parameter is set to include at least one pause. If the seating trajectory mode includes a non-frontal oblique entry mode, the action orientation is set to enter at a preset angle to the central axis of the learning machine under test. If the seating trajectory mode includes a multi-person sequence mode, the number of loops is set to be greater than or equal to 2.

[0011] In some embodiments, during the process of the robotic arm performing a seat-down simulation action, test data is collected synchronously, including: Obtain the data collection frequency; Data acquisition requests are sent to the target data source according to the data acquisition frequency, the target data source including the robotic arm and the learning machine under test; the data acquisition request includes a first acquisition request and a second acquisition request. Receive the real-time spatial coordinates of the robotic arm in response to the first acquisition request; The test learning machine receives the continuous distance readings, target confidence, ambient light values, and screen status events fed back via ADB connection in response to the second acquisition request.

[0012] In some embodiments, generating stress test result information based on the ideal data logic chain and the test data includes: Read the screen state events from the test data; If the screen state event is a screen light-up event, generate a normal wake-up result; If the screen state event is a screen not lit event, a wake-up abnormality sitting wake-up result is generated, and key test events are determined by comparing the ideal data logic chain and the test data.

[0013] In some embodiments, generating stress test result information based on the ideal data logic chain and the test data includes: Based on the timestamps of the test data, the test data is sorted to generate a logical chain of test data; The test data logic chain is verified using the ideal data logic chain to obtain a verification result, which includes at least one abnormal data. Record the abnormal test events associated with the abnormal data in the verification results.

[0014] In some embodiments, the method further includes: Set a critical time window based on the aforementioned abnormal test events; Record the real-time spatial coordinates of the robotic arm and the raw data stream collected by the sensors in the learning machine under test within the key time window; Capture device system logs within the specified critical time window; A test snapshot data package is generated based on the real-time spatial coordinates, the raw data stream, and the device system log.

[0015] According to a second aspect of this application, a seating and wake-up stress testing system for an intelligent learning machine is provided, the system comprising: a test execution mechanism, a data acquisition module, and an intelligent control module; The test execution mechanism includes an adjustable support, a robotic arm, a fixed corner bracket, and a seated human model; the seated human model is set at the end of the robotic arm via the fixed corner bracket; the learning machine under test is set on the adjustable support. The data acquisition module is connected to the learning machine under test via an ADB interface to acquire in real time the screen status events, continuous distance readings of the position sensor, target confidence of the head sensor, ambient light values ​​of the ambient light sensor, and complete system logs of the learning machine under test; the data acquisition module is also connected to the robotic arm to acquire the real-time spatial coordinates of the robotic arm. The intelligent control module connects the robotic arm and the data acquisition module; the intelligent control module is configured to execute the intelligent learning machine seating and wake-up stress test method described in the first aspect.

[0016] According to a third aspect of this application, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the program to implement the above-described intelligent learning machine seat wake-up stress test method.

[0017] According to a fourth aspect of this application, a storage medium is provided that stores a computer program thereon, which, when executed by a processor, implements the above-described intelligent learning machine seating and wake-up stress test method.

[0018] By employing the above technical solutions, this application provides a method and system for stress testing the seat-and-wake-up mechanism of an intelligent learning machine. The method responds to test commands, generates a test plan, and controls a robotic arm to perform a seat-and-wake simulation while simultaneously collecting test data. It then obtains the ideal data logic chain of the seat-and-wake trajectory pattern in the test plan, thereby generating stress test result information based on the ideal data logic chain and the test data. The stress test result information includes seat-and-wake-up results generated based on screen state events and key test events determined by comparing the test data with the ideal data logic chain. This method can utilize a robotic arm to execute multiple pre-stored seat-and-wake trajectory patterns to simulate user seat-and-wake behavior, and combines multi-dimensional test data for seat-and-wake-up testing and root cause analysis. This enables multi-modal data collaborative verification and adaptive test strategy adjustment, thereby accurately identifying logical defects in the wake-up process and improving the accuracy and efficiency of test results.

[0019] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0020] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a schematic diagram of the intelligent learning machine seating and wake-up stress testing system provided in an embodiment of this application; Figure 2 This is a schematic diagram of the adjustable support structure provided in an embodiment of this application; Figure 3 This is a schematic diagram of the robotic arm and seated human figure model provided in an embodiment of this application; Figure 4 This is a schematic diagram of the intelligent learning machine seating and wake-up stress test method provided in an embodiment of this application; Figure 5 This is a schematic diagram of the process for generating a test snapshot data packet provided in an embodiment of this application; Figure 6 This is a schematic diagram of the dynamic adjustment test plan process provided in an embodiment of this application. Detailed Implementation

[0021] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0022] In this embodiment, the intelligent learning machine is an intelligent mobile terminal, an electronic device that uses computer, internet, automation, and artificial intelligence technologies for the purpose of educational information interaction. The term "intelligent learning machine" has a broad meaning; it can specifically refer to a dedicated learning machine used in a smart teaching system, or it can refer to other electronic devices used in the smart teaching process, such as mobile terminals like mobile phones and tablets, as well as all-in-one personal computers and workstations.

[0023] In some embodiments, the intelligent learning machine includes a processor, a storage module, a display module, an input / output module, and a power supply module. The processor is the core hardware of the intelligent learning machine, used for computation and instruction control. For example, the processor may be a multi-core architecture central processing unit (CPU), a system-on-chip (SoC), or a microcontroller unit (MCU) to meet complex computational needs, such as image recognition, voice interaction, and intelligent tutoring tasks.

[0024] Storage modules, comprising memory and storage chips, are used for data storage, data retrieval, and caching acceleration. For example, storage modules can store learning materials, applications, and user data. They can also detect and correct errors during data transmission and storage using verification technologies, and support different interfaces and specifications. Users can increase storage capacity as needed to meet their data storage requirements.

[0025] The display module may include a display screen and display driver components for controlling the display process. The display module can be used for information presentation, displaying learning content, user interfaces, and other information to the user in the form of images and text, allowing the user to intuitively view learning materials, operation prompts, etc. For example, a smart learning machine may be equipped with a 10.95-inch 2K resolution touchscreen.

[0026] The input / output module integrates various sensors such as a camera, proximity sensor, and infrared sensor to achieve functions such as data input, data output, device connection, and signal conversion. Examples of input / output modules include buttons, microphone, speaker, USB interface, SD card interface, Bluetooth, and Wi-Fi modules.

[0027] The power module is responsible for power supply, voltage conversion, power management, and charging management of the entire smart learning machine. The power module can use a rechargeable battery and supports an external power interface to ensure the device's battery life and stability.

[0028] Based on the aforementioned functional components, the intelligent learning machine can achieve user interaction by running an operating system and applications to meet users' functional needs for intelligent teaching. For example, the operating system of the intelligent learning machine can be based on Android or other customized operating systems, and support the installation and running of applications to provide a user-friendly interface.

[0029] The smart learning machine supports a seat-activated wake-up function. This function uses built-in cameras, proximity sensors, and other detection components to detect if a user is seated in the designated seating area in front of the machine. When a user is detected, the screen automatically lights up to facilitate user interaction. As an intelligent interaction method, the seat-activated wake-up function requires stress testing before official application to ensure a high probability of successful wake-up and minimize false wake-ups.

[0030] In some embodiments, the seat-based wake-up stress test can be performed manually. That is, before the smart learning machine leaves the factory or before its operating system (or control program) is released, a stress test can be conducted on the seat-based wake-up function. During the test, the smart learning machine is fixed in front of a seat in its normal usage configuration, and the seat-based wake-up function is activated. Then, a tester performs the sitting action and records the wake-up status of the smart learning machine.

[0031] Since manual seat-to-wake stress testing is inefficient, costly, and difficult to guarantee test consistency, in some embodiments, automated seat-to-wake stress testing can be performed using a robotic arm and a fixed frame. That is, after the smart learning machine is installed on the fixed frame, the robotic arm simulates the user's spatial position movement during the seat-to-wake process and records the number of times the smart learning machine is woken up and the wake-up probability to calculate the pass rate of the seat-to-wake function.

[0032] However, this automated testing method cannot simulate various user seating behaviors, resulting in simplistic test logic that cannot handle complex scenarios. Furthermore, when tests fail, there is a lack of effective means to quickly pinpoint the problem's origin—whether it stems from hardware sensors, data fusion algorithms, or upper-layer application logic—leading to lengthy troubleshooting cycles and inaccurate test results.

[0033] To address the issue of inaccurate stress test results for the seat-activated wake-up function, this application provides a seat-activated wake-up stress test method for intelligent learning machines in some embodiments. This method can automate the testing of the seat-activated wake-up function and introduces dynamic multi-scenario behavior simulation, multi-modal test data collaborative verification, and intelligent adaptive testing strategies. This enables the accurate discovery of deep, intermittent logical defects that are difficult to reach with conventional testing, improving the accuracy and efficiency of test results, and enhancing scenario coverage, testing depth, and problem diagnosis efficiency.

[0034] The method can be applied to a smart learning machine seat-and-wake stress testing system (hereinafter referred to as the stress testing system), specifically to the intelligent control module of the smart learning machine seat-and-wake stress testing system. To execute the smart learning machine seat-and-wake stress testing method, as follows... Figure 1 As shown, the stress testing system includes: a test execution mechanism, a data acquisition module, and an intelligent control module.

[0035] The testing mechanism includes an adjustable support, a robotic arm, fixed corner brackets, and a seated human model. The seated human model is mounted on the end effector of the robotic arm via fixed corner brackets, and the learning machine under test is mounted on the adjustable support during the test.

[0036] For example, the test execution mechanism includes a smart learning tablet fixed on an adjustable stand, such as... Figure 2 As shown. The RoArm-M2-Pro robotic arm, controlled by a Raspberry Pi 5, and the mannequin plate fixed to the end of the robotic arm via fixed angle brackets, are shown. Figure 3 As shown, a human image, including a head, is affixed to the model board. The adjustable stand is a flat panel stand with adjustable height and angle, allowing the learning tablet under test to be fixed in place and its posture adjusted to maintain the required height during testing. To ensure stability during testing, hot melt adhesive can be used to reinforce key connection points.

[0037] The data acquisition module connects to the learning machine under test via the Android Debug Bridge (ADB) interface to acquire, in real time, the learning machine's screen status events, continuous distance readings from the position sensor (P-sensor), target confidence from the head sensor (TOF-sensor), ambient light values ​​from the ambient light sensor, and complete system logs. The data acquisition module also connects to a robotic arm to obtain its real-time spatial coordinates.

[0038] For example, the data acquisition module connects via ADB to acquire in real time the screen on / off status of the learning tablet, the proximity signal and distance value of the distance sensor (P-sensor), the recognition results of the head sensor (TOF-sensor) (such as the recognition confidence of faces or human-shaped targets), ambient light sensor data, and complete system logs (such as aplog, bugreport, etc.).

[0039] The intelligent control module connects the robotic arm and the data acquisition module. Furthermore, the intelligent control module is configured to execute the program steps corresponding to the intelligent learning machine's seating and wake-up stress test method, serving as the main control component. It runs the core control script and is responsible for directing the robotic arm's movement, sending ADB commands, and performing data synchronization analysis and intelligent decision-making.

[0040] For example, the intelligent control module can be a Raspberry Pi 5, which can control the RoArm-M2-Pro robotic arm during testing. The Raspberry Pi 5 is a single-board computer released by the Raspberry Pi Foundation in October 2023. The RoArm-M2-S Pro is a high-performance desktop robotic arm from Hiwonder, designed for robotics education, maker development, automated experiments, and light industrial applications. It combines high-precision servo motors, an open-source control architecture, and a modular design, making it suitable for scientific research, STEAM education, and industrial prototyping.

[0041] The method can also be applied to electronic devices with data processing capabilities. These electronic devices include, but are not limited to, computers, servers, mobile terminals, smart wearable devices, industrial control computers, and single-board computers. For ease of description, this application embodiment uses a pressure testing system or intelligent control module as the execution subject of the method. It should be understood that the method can also be applied to other types of execution subjects, which are not shown one by one in this application embodiment. Figure 4 As shown, the method includes: S101. In response to the test command, generate a test plan.

[0042] The load testing system can receive test commands input by the user after it starts running. These test commands control the system to begin the seating and wake-up load test. Therefore, the test commands can include relevant parameters for controlling the test process. For example, test commands can include parameters such as test time, test object, test content, and test mode.

[0043] Upon receiving a test command from the user, the load testing system can respond to the command and generate a test plan accordingly. The test plan is a collection of control program data that can be executed in the intelligent control module, such as control scripts related to the test program.

[0044] The test plan can be generated based on the test instructions input by the user. During the test plan generation process, the stress testing system can read relevant test parameters from the test instructions and write test programs or call pre-written test programs based on these parameters, forming a control script. Then, by running the control script, the system controls the robotic arm and the learning machine under test during the seating and wake-up stress test.

[0045] The test plan may include at least one seating trajectory pattern from a pre-stored pattern library. To achieve dynamic behavior simulation of seating actions in multiple scenarios, multiple seating trajectory patterns can be pre-stored in the storage unit of the load testing system to form a pre-stored pattern library.

[0046] For example, the seating trajectory mode can be any one of mode A (standard seating mode), mode B (rapid impact mode), mode C (hesitant approach mode), mode D (non-frontal oblique entry mode), and mode E (multi-person sequence mode) to expand the coverage of test scenarios by simulating rich and varied real user seating behaviors, thereby assessing the robustness and adaptability of the intelligent learning machine under different behavioral modes.

[0047] In addition to pre-stored seating trajectory patterns, the pre-stored pattern library can also store action parameters related to the seating trajectory patterns, as well as preset test items and related judgment criteria when performing seating wake-up stress tests. For example, the pre-stored pattern library can be a comprehensive database stored in the stress test system's storage unit. This includes the decomposed actions and action parameters for each seating trajectory pattern, as well as the reasonable detection parameter range when performing that seating trajectory pattern.

[0048] To generate a test plan that includes a seating trajectory pattern, the load testing system, after receiving the user's input test instructions, can read the user-specified seating trajectory pattern from the test instructions. Based on the user-specified seating trajectory pattern, it extracts the relevant motion parameters of the seating trajectory pattern from the pre-stored pattern library and arranges control scripts based on the motion parameters. The arranged control scripts can include control scripts for controlling the robotic arm to perform simulated seating behavior and control scripts for controlling the data acquisition module to acquire test data, thus forming the test plan.

[0049] S102. According to the test plan, control the robotic arm to perform a sitting simulation action.

[0050] After generating the test plan, the interactive actions corresponding to the user-specified seating trajectory pattern can be simulated according to the test plan, that is, the robotic arm can be controlled to perform simulated seating actions. Obviously, the simulated seating actions performed by the robotic arm will be different for different seating trajectory patterns.

[0051] For example, in Mode A, the robotic arm can simulate a standard sitting motion, smoothly moving a human figure from the upper right corner of the learning machine under test to the seated position directly in front at a uniform speed. In Mode B, the robotic arm can simulate a rapid, impact-driven sitting motion, entering the detection area at a speed exceeding the standard to test the system's response limits. In Mode C, the robotic arm can simulate a hesitant approach sitting motion, adding one or more pauses during entry into the detection area to simulate user hesitation. In Mode D, the robotic arm can simulate a non-frontal, oblique entry sitting motion, moving in from an angle with the central axis of the tablet to test the sensor's field of view coverage capability in the learning machine under test. In Mode E, the robotic arm can simulate a multi-user sequence sitting motion, simulating an interaction scenario where one user leaves and another immediately follows to sit in the seat directly in front.

[0052] In some embodiments, the stress testing system can control the robotic arm through a series of trajectory simulation commands. Specifically, when executing a simulated sitting motion according to the test plan, the system first reads the sitting trajectory pattern from the test plan, and then sets motion parameters for at least one decomposed motion based on the sitting trajectory pattern. These motion parameters include at least one of the following: motion orientation, motion speed, pause parameters, and number of cycles associated with the sitting trajectory pattern.

[0053] Therefore, different motion parameters can be set for different sitting trajectory patterns. That is, after reading the sitting trajectory pattern from the test plan, the specific sitting trajectory pattern can be judged. If the sitting trajectory pattern includes the standard sitting pattern, the motion direction is set to a smooth movement from the upper corner of the learning machine under test to the detection area directly in front, and the motion speed is set to the first speed to simulate the sitting motion of moving smoothly from the upper right corner of the learning machine under test at a constant speed (the first speed) to the sitting position directly in front.

[0054] Similarly, if the seating trajectory mode includes a rapid impact mode, the stress testing system can set the movement direction to a smooth movement from the upper corner of the learning machine under test towards the detection area directly in front, and then set the movement speed to a second speed. The second speed is greater than the first speed, simulating a faster-than-standard speed to quickly enter the detection area, thereby testing the response limits of the learning machine's seating wake-up function.

[0055] If the seating trajectory mode includes a hesitant approach mode, the stress testing system can, on the basis of setting the action orientation to smoothly move from the upper corner of the learning machine under test to the detection area directly in front and setting the action speed to the first speed, additionally set a pause parameter including at least one pause, so as to add one or more pauses during the process of entering the detection area to simulate the user's hesitant behavior.

[0056] If the seating trajectory mode includes a non-frontal oblique entry mode, the stress testing system can set the action orientation to enter from a direction with a preset angle to the central axis of the learning machine under test, and set the action speed to the first speed to simulate the user entering from a direction with a certain angle to the central axis of the learning machine under test, thereby testing the sensor field of view coverage capability of the learning machine under test.

[0057] If the seating trajectory mode includes a multi-person sequence mode, the stress testing system can, on the basis of setting the action direction to move smoothly from the upper corner of the learning machine under test to the front detection area and setting the action speed to the first speed, additionally set the number of loops, that is, set the number of loops to be greater than or equal to 2, to simulate the scenario where one user leaves and another user immediately enters the front detection area to sit down.

[0058] After setting the motion parameters for the decomposed motion based on the seating trajectory pattern, trajectory simulation commands can be generated according to the motion parameters, and a set of control commands can be generated by combining the trajectory simulation commands. Then, according to the set of control commands, the trajectory simulation commands are sent to the robotic arm in sequence to control the robotic arm to execute the seating simulation motion according to the trajectory simulation commands.

[0059] For example, the test plan generated by the stress testing system based on the test instructions may include first executing the standard sitting mode, followed by the rapid impact mode. The stress testing system can then generate two trajectory simulation commands based on the sitting trajectory pattern in the test plan: Trajectory Simulation Command A and Trajectory Simulation Command B. Trajectory Simulation Command A includes motion parameters: a smooth movement from the upper corner of the learning machine under test towards the detection area in front, and a first speed. Trajectory Simulation Command B includes motion parameters: a smooth movement from the upper corner of the learning machine under test towards the detection area in front, and a second speed. Trajectory Simulation Command A and Trajectory Simulation Command B are then sent sequentially to the robotic arm to control it to perform a simulated smooth movement from the upper right corner to the sitting position in front, followed by a simulated rapid entry into the detection area at a speed higher than the standard.

[0060] S103. During the process of the robotic arm performing the sitting simulation action, test data is collected synchronously.

[0061] After controlling the robotic arm to perform a seat-simulation action, the stress testing system can enter data acquisition mode to simultaneously collect test data during the seat-simulation process. This test data can include multiple dimensions to comprehensively reflect the performance of the seat-wake-up function of the learning machine under test.

[0062] Therefore, test data can include the real-time spatial coordinates of the robotic arm recorded synchronously with a uniform timestamp, continuous distance readings from the position sensor, target confidence from the head sensor, ambient light values ​​from the ambient light sensor, and screen status events of the learning machine under test.

[0063] To acquire test data, in some embodiments, when the load testing system synchronously collects test data during the robotic arm's simulated sitting motion, it can first obtain the data acquisition frequency and then send data acquisition requests to the target data source according to the data acquisition frequency. Here, the target data source generally refers to the device used to generate and collect test data; that is, the target data source includes the robotic arm and the learning machine under test. Therefore, the load testing system can generate different types of data acquisition requests for the robotic arm and the learning machine under test. Thus, the data acquisition requests include a first acquisition request and a second acquisition request. The first acquisition request is used to collect test data from the robotic arm; the second acquisition request is used to collect test data from the learning machine under test.

[0064] After sending a data acquisition request to the target data source according to the data acquisition frequency, the stress testing system can receive the real-time spatial coordinates of the robotic arm in response to the first acquisition request, as well as the continuous distance readings, target confidence, ambient light values, and screen status events fed back by the learning machine under test via the ADB connection in response to the second acquisition request.

[0065] For example, to achieve synchronization of multimodal sensor data, the pressure testing system can initiate high-frequency data synchronous acquisition when performing the simulated sitting action corresponding to the aforementioned sitting trajectory mode. This means the data acquisition frequency is greater than or equal to a preset frequency threshold. Then, using a unified timestamp, the system synchronously records the real-time spatial coordinates of the robotic arm, the continuous distance readings of the P-sensor, the target confidence level of the TOF-sensor, the ambient light sensor values, and screen status events to obtain multimodal test data.

[0066] S104. Obtain the ideal data logic chain of the seating trajectory pattern in the test plan.

[0067] Based on the synchronous acquisition of multimodal test data, the load testing system can also perform collaborative verification. To this end, the load testing system can obtain the ideal data logic chain for each seating trajectory pattern in the test plan. This ideal data logic chain can be built into a pre-stored pattern library and is associated with each seating trajectory pattern. When the user specifies that the load testing system executes the test according to any seating trajectory pattern, the load testing system can extract the ideal data logic chain related to the current seating trajectory pattern from the pre-stored pattern library.

[0068] An ideal data logic chain can include multiple logical nodes, which are arranged in chronological order. These nodes represent the standard test steps under the current seating trajectory mode, as well as the action type, data existence status, corresponding reasonable data range, and data change status of each test step.

[0069] For example, in a standard seating mode, a successful seating wake-up should follow an ideal data logic chain: "robotic arm enters - P-sensor detects continuously decreasing distance - TOF-sensor identifies target and confidence steadily increases - screen lights up." Therefore, the stress testing system not only checks the final test results but also needs to analyze the completeness and rationality of the logic chain corresponding to the test data in real time based on the ideal data logic chain.

[0070] S105. Generate stress test result information based on the ideal data logic chain and test data.

[0071] After obtaining the test data and the ideal data logic chain, the load testing system can generate load testing results based on these two sets of data. These results include the seat-and-wake-up results and key test events. The seat-and-wake-up results are generated based on screen state events, while key test events are determined by comparing the test data with the ideal data logic chain. In other words, the load testing system can determine whether the seat-and-wake-up was successfully executed by analyzing the test data, and it can identify key test events during the testing process by comparing the ideal data logic chain with the test data.

[0072] In order to generate the seat wake-up result and key test events, in some embodiments, when the stress testing system generates stress test result information based on the ideal data logic chain and test data, it can first read the screen state event from the test data. If the screen state event is a screen lighting event, it means that the learning machine under test is normally woken up and the seat wake-up function is normal. Therefore, a normal wake-up result can be generated.

[0073] If the screen status event is "screen not lit," it indicates that the learning machine under test has not been properly woken up, and the seat-to-wake function is malfunctioning. In this case, a seat-to-wake result indicating an abnormal wake-up can be generated. When the seat-to-wake function is determined to be malfunctioning, the load testing system can mark the event as an "abnormal wake-up event" instead of simply outputting a success or failure test result. Therefore, the load testing system can identify key test events by comparing the ideal data logic chain with the test data.

[0074] To identify critical test events, in some embodiments, when the load testing system generates load testing result information based on the ideal data logic chain and test data, it can first sort the test data according to the timestamps to generate a test data logic chain. Then, the ideal data logic chain is used to verify the test data logic chain to obtain a verification result. The verification result includes at least one type of abnormal data. Critical test events are then identified by recording the abnormal test events associated with the abnormal data in the verification result.

[0075] For example, the load testing system can mark the current learning machine under test as having an abnormal seat-to-wake function and record the abnormal wake-up event when one of the following situations occurs: Situation 1: If the screen is confirmed to be lit through screen status events, but the target is not identified throughout the data detected by the TOF-sensor, it may be due to a false triggering of the P-sensor or a logical error. Situation 2: If the TOF-sensor shows high-confidence recognition, but the P-sensor shows no proximity signal, the abnormality may be due to a false TOF recognition. Situation 3: If the test data logic chain is normal, but the screen status events indicate that the screen is not lit, it can be determined that there is an application-layer logic problem in the current seat-to-wake process.

[0076] It is evident that by recording the potential problems corresponding to the verification results, key test events can be identified, thereby achieving a "penetration" of the internal working state of the equipment. This allows for the accurate identification of potential, intermittent logical defects and sensor coordination issues from massive amounts of testing, something that purely result-based testing cannot achieve.

[0077] By applying the technical solutions of the above embodiments, the intelligent learning machine seating and wake-up stress test method described in the above embodiments can realize automated intelligent learning machine seating and wake-up stress test by using multimodal test data synchronously recorded with a unified timestamp and combining it with the ideal data logic chain of the seating trajectory pattern. Furthermore, by introducing dynamic multi-scenario behavior simulation and multimodal sensor data collaborative verification, it can accurately discover deep and intermittent logical defects that are difficult to reach by conventional testing, thereby improving the accuracy and efficiency of test results.

[0078] In some embodiments, as a refinement and extension of the specific implementation of the above embodiments, and to fully illustrate the specific implementation process of this embodiment, some embodiments of this application also provide a method for stress testing the seating and wake-up of an intelligent learning machine. The difference between this method and the above embodiments is that a root cause analysis package can be integrated when generating stress test result information, such as... Figure 5 As shown, the method includes: S201. Set critical time windows based on abnormal test events; S202. Record the real-time spatial coordinates of the robotic arm and the raw data stream collected by the sensors in the learning machine under test within the key time window. S203. Capture device system logs within critical time windows; S204. Generate a test snapshot data package based on real-time spatial coordinates, raw data stream, and device system logs.

[0079] When generating stress test results, if the seat-and-wake function of the current intelligent learning machine is found to be abnormal, a critical time window can be set based on the abnormal test event, and the real-time spatial coordinates of the robotic arm and the raw data stream collected by the sensors in the learning machine under test can be recorded within the critical time window. Then, by capturing the device system log within the critical time window, a test snapshot data package can be generated based on the real-time spatial coordinates, raw data stream, and device system log.

[0080] For example, when the generated load test results contain any failures or anomalies, the load testing system automatically triggers a snapshot mechanism. This snapshot mechanism is not a simple log file, but a rich data package—a test snapshot data package. This package can contain raw data streams from all sensors within a specific time window before and after the event, precise motion trajectory data of the robotic arm during the corresponding time period, and system logs from the tested learning machine, such as aplog and bugreport. The generated test snapshot data package is then stored for review by relevant R&D personnel.

[0081] By applying the technical solutions of the above embodiments, the intelligent learning machine seating and wake-up stress testing method described in the above embodiments can integrate a root cause analysis package when generating stress test result information. That is, it generates a test snapshot data package based on the real-time spatial coordinates, raw data stream, and device system logs within the time window corresponding to the abnormal event. The generated test snapshot data package can provide R&D personnel with a complete context of the problem occurrence, shortening the time for R&D personnel to locate and fix the problem.

[0082] In some embodiments, as a refinement and extension of the specific implementation of the above embodiments, and to fully illustrate the specific implementation process of this embodiment, some embodiments of this application also provide a method for stress testing the seating and wake-up of an intelligent learning machine. The difference between this method and the above embodiments is that the stress testing system can dynamically adjust the test plan based on real-time test results according to an adaptive testing strategy, such as... Figure 6 As shown, the method includes: S301. Obtain multiple stress test results within a preset test cycle; S302. Read frequently tested items from multiple stress test results; S303. Dynamically adjust the test plan based on frequently tested items.

[0083] To achieve intelligent adaptive load testing, the load testing system can analyze the load testing results within a preset test period after generating the results. Specifically, it acquires multiple load testing results within the preset test period and then reads frequently occurring test items from these results. These frequently occurring test items are wake-up results and / or key test events that occur more than or equal to a specified frequency threshold. The test plan is then dynamically adjusted based on these frequently occurring test items.

[0084] When dynamically adjusting the test plan based on the frequent test items, the load testing system can also match test items to be adjusted in a pre-stored pattern library based on the frequent test items. The test items to be adjusted include test items to be added and / or test items to be deleted. Test items to be added are determined based on abnormally frequent test items whose occurrence frequency is greater than or equal to a threshold, while test items to be deleted are determined based on normally frequent test items whose occurrence frequency is greater than or equal to the threshold. The test plan is then modified based on the test items to be adjusted, and an updated trajectory pattern is generated based on the modified test plan, thereby adding the updated trajectory pattern to the pre-stored pattern library.

[0085] For example, when several "abnormal wake-up events" or seated wake-up failures occur in succession, the stress testing system will automatically increase the test density near the current trajectory parameters or switch to a more extreme test mode to actively "chase" and reproduce the defects.

[0086] By applying the technical solutions of the above embodiments, the intelligent learning machine seating and wake-up stress testing method described in the above embodiments can, after generating stress test result information, determine frequently tested items through statistical analysis of multiple stress test result information, and then dynamically adjust the test plan based on the frequently tested items. The method can achieve multi-intelligent adaptive stress testing, and dynamically adjust the intensity, density, and strategy of subsequent tests based on real-time test results, dynamically update the pre-stored pattern library, and improve the adaptability and personalized stress testing of the stress testing process.

[0087] In some embodiments, as a specific implementation of the intelligent learning machine seat-wake-up stress testing method described in the above embodiments, some embodiments of this application also provide an intelligent learning machine seat-wake-up stress testing system, such as... Figure 1 As shown, the system includes: a test execution mechanism, a data acquisition module, and an intelligent control module; The test execution mechanism includes an adjustable support, a robotic arm, a fixed corner bracket, and a seated human model; the seated human model is set at the end of the robotic arm via the fixed corner bracket; the learning machine under test is set on the adjustable support. The data acquisition module is connected to the learning machine under test via an ADB interface to acquire in real time the screen status events, continuous distance readings of the position sensor, target confidence of the head sensor, ambient light values ​​of the ambient light sensor, and complete system logs of the learning machine under test; the data acquisition module is also connected to the robotic arm to acquire the real-time spatial coordinates of the robotic arm. The intelligent control module connects the robotic arm and the data acquisition module; the intelligent control module is configured to execute the intelligent learning machine seating and wake-up stress test method described in any of the above embodiments.

[0088] By applying the technical solutions of the above embodiments, the intelligent learning machine seat-and-wake stress testing method and system described in the above embodiments can respond to test commands, generate test plans, and control a robotic arm to perform seat-and-wake simulation actions and simultaneously collect test data according to the test plans. Then, it obtains the ideal data logic chain of the seat-and-wake trajectory patterns in the test plans, thereby generating stress test result information based on the ideal data logic chain and the test data. The stress test result information includes seat-and-wake results generated based on screen state events and key test events determined by comparing test data with the ideal data logic chain. By utilizing a robotic arm to execute multiple pre-stored seat-and-wake trajectory patterns to simulate user seat-and-wake behavior, and combining multi-dimensional test data for seat-and-wake testing and root cause analysis, it achieves multi-modal data collaborative verification and adaptive test strategy adjustment, thereby accurately identifying logical defects in the wake-up process and improving the accuracy and efficiency of test results.

[0089] It should be noted that other corresponding descriptions of the functional units involved in the intelligent learning machine seating and wake-up stress testing system provided in this application embodiment can be found in the corresponding descriptions in the intelligent learning machine seating and wake-up stress testing method provided in the above embodiment, and will not be repeated here.

[0090] This application also provides a computer device, specifically a personal computer, server, network device, etc. The computer device includes a bus, processor, memory, and communication interface, and may also include input / output interfaces and a display device. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores location information. The network interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the various method embodiments.

[0091] Those skilled in the art will understand that the structure of the computer device described above is only a partial structure related to the solution of this application and does not constitute a limitation on the computer device to which the solution of this application is applied. A specific computer device may include more or fewer components, or combine certain components, or have different component arrangements.

[0092] In one embodiment, a computer-readable storage medium is also provided, which may be non-volatile or volatile, and a computer program is stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0093] In one embodiment, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0094] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0095] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods.

[0096] Any references to memory, database, or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc.

[0097] Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can take many forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0098] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchain. The processors involved in the embodiments provided in this application may be, but are not limited to, general-purpose processors, graphics processors, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc.

[0099] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0100] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for stress testing the wake-up function of an intelligent learning machine upon seating, characterized in that, The method includes: In response to a test command, a test plan is generated, the test plan including at least one seating trajectory pattern from a pre-stored pattern library; According to the test plan, control the robotic arm to perform a seating simulation action; During the process of the robotic arm performing the sitting simulation action, test data is collected synchronously. The test data includes the real-time spatial coordinates of the robotic arm, the continuous distance readings of the position sensor, the target confidence of the head sensor, the ambient light value of the ambient light sensor, and the screen status events of the learning machine under test, all recorded synchronously with a unified timestamp. Obtain the ideal data logic chain for the seating trajectory pattern described in the test plan; The stress test results are generated based on the ideal data logic chain and the test data. The stress test results include seat wake-up results and key test events. The seat wake-up results are generated based on the screen state events. The key test events are determined by comparing the test data with the ideal data logic chain.

2. The method according to claim 1, characterized in that, The method further includes: Obtain multiple load test results within a preset test period; Frequent test items are read from multiple stress test result information, wherein the frequent test items are the seat wake-up results and / or the key test events that occur more than or equal to the number of occurrences threshold; The test plan is dynamically adjusted based on the frequently tested items.

3. The method according to claim 2, characterized in that, Dynamically adjust the test plan based on the frequently tested items, including: Based on the frequent test items, test items to be adjusted are matched in the pre-stored pattern library. The test items to be adjusted include test items to be added and / or test items to be deleted. The test items to be added are determined based on abnormally frequent test items whose occurrence frequency is greater than or equal to a frequency threshold. The test items to be deleted are determined based on normally frequent test items whose occurrence frequency is greater than or equal to a frequency threshold. Modify the test plan according to the test items to be adjusted, and generate an updated trajectory pattern based on the modified test plan; Add the updated trajectory pattern to the pre-stored pattern library.

4. The method according to claim 1, characterized in that, According to the test plan, the robotic arm is controlled to perform a seat-down simulation action, including: Read the seating trajectory pattern from the test plan; Based on the seating trajectory pattern, at least one decomposed action parameter is set, and the action parameter includes at least one of the action orientation, action speed, pause parameter and number of cycles associated with the seating trajectory pattern. Generate trajectory simulation commands based on the motion parameters, and generate a set of control commands by combining the trajectory simulation commands; According to the set of control commands, the trajectory simulation commands are sent to the robotic arm in sequence to control the robotic arm to perform a sitting simulation action according to the trajectory simulation commands.

5. The method according to claim 5, characterized in that, Based on the seating trajectory pattern, at least one decomposed action parameter is set, including: If the seating trajectory pattern includes the standard seating pattern, the movement direction is set to a smooth movement from the upper corner of the learning machine under test to the detection area directly in front, and the movement speed is a first speed. If the sitting trajectory mode includes a rapid impact mode, the action speed is set to a second speed, which is greater than the first speed; If the seating trajectory mode includes a hesitant approach mode, the pause parameter is set to include at least one pause. If the seating trajectory mode includes a non-frontal oblique entry mode, the action orientation is set to enter at a preset angle to the central axis of the learning machine under test. If the seating trajectory mode includes a multi-person sequence mode, the number of loops is set to be greater than or equal to 2.

6. The method according to claim 1, characterized in that, During the process of the robotic arm performing the seating simulation action, test data is collected simultaneously, including: Obtain the data collection frequency; Data acquisition requests are sent to the target data source according to the data acquisition frequency, the target data source including the robotic arm and the learning machine under test; the data acquisition request includes a first acquisition request and a second acquisition request. Receive the real-time spatial coordinates of the robotic arm in response to the first acquisition request; The test learning machine receives the continuous distance readings, target confidence, ambient light values, and screen status events fed back via ADB connection in response to the second acquisition request.

7. The method according to claim 1, characterized in that, Based on the ideal data logic chain and the test data, stress test result information is generated, including: Read the screen state events from the test data; If the screen state event is a screen light-up event, generate a normal wake-up result; If the screen state event is a screen not lit event, a wake-up abnormality sitting wake-up result is generated, and key test events are determined by comparing the ideal data logic chain and the test data.

8. The method according to claim 1, characterized in that, Based on the ideal data logic chain and the test data, stress test result information is generated, including: Based on the timestamps of the test data, the test data is sorted to generate a logical chain of test data; The test data logic chain is verified using the ideal data logic chain to obtain a verification result, which includes at least one abnormal data. Record the abnormal test events associated with the abnormal data in the verification results.

9. The method according to claim 8, characterized in that, The method further includes: Set a critical time window based on the aforementioned abnormal test events; Record the real-time spatial coordinates of the robotic arm and the raw data stream collected by the sensors in the learning machine under test within the key time window; Capture device system logs within the specified critical time window; A test snapshot data package is generated based on the real-time spatial coordinates, the raw data stream, and the device system log.

10. A seating wake-up pressure testing system for an intelligent learning machine, characterized in that, The system includes: a test execution mechanism, a data acquisition module, and an intelligent control module; The test execution mechanism includes an adjustable support, a robotic arm, a fixed corner bracket, and a seated human model; the seated human model is set at the end of the robotic arm via the fixed corner bracket; the learning machine under test is set on the adjustable support. The data acquisition module is connected to the learning machine under test via an ADB interface to acquire in real time the screen status events, continuous distance readings of the position sensor, target confidence of the head sensor, ambient light values ​​of the ambient light sensor, and complete system logs of the learning machine under test; the data acquisition module is also connected to the robotic arm to acquire the real-time spatial coordinates of the robotic arm. The intelligent control module connects the robotic arm and the data acquisition module; the intelligent control module is configured to perform the intelligent learning machine seating and wake-up pressure test method according to any one of claims 1-9.

Images