Self-cleaning technology evaluation method and system based on home robot

Abstract

The invention is suitable for the technical field of household cleaning, and provides a self-cleaning technology evaluation method and system based on a household robot, and the method comprises the steps: obtaining household equipment information, and obtaining three-dimensional modeling information; analyzing according to the micro biosensing information, comparing an analysis result with a safety threshold value, and carrying out differential calculation on an initial value and a final value; updating three-dimensional modeling information according to the periodically scanned data, performing trend analysis on pollution generation frequency of each region according to a cleaning historical record, and performing re-analysis on the processed target region; the emotional tendency is judged according to the semantic protector, the facial emotional state and the limb behavior are analyzed according to the expression behavior, and the execution deviation in the task process is accurately recognized according to the recognizable label, so that the problem that an existing household robot cannot comprehensively analyze the clean area during self-cleaning, and the cleaning efficiency is poor is solved. The problems that different areas of a family cannot be divided into different areas, and different levels of processing cannot be carried out on the cleaning frequencies of the different areas are solved.

Description

Technical Field

[0001] The present invention belongs to the field of household cleaning technology, and in particular relates to a method and system for evaluating a self-cleaning technology based on a household robot. Background Art

[0002] With the continuous advancement of home automation and intelligence, home service robots with automated sweeping, vacuuming, and wet mopping capabilities are now widely used. These robots typically use sensing methods such as lidar, visual cameras, or infrared sensors for mapping and path planning. They can cover most areas of the home and perform relatively efficient daily cleaning tasks.

[0003] Some high-end devices also feature self-cleaning features, such as automatic brush flushing, mopping, or filters, improving maintenance convenience and device durability. These robots are often controlled via apps or voice assistants, allowing users to set cleaning times, task areas, or cleaning modes, allowing them to complete basic floor cleaning tasks unattended.

[0004] However, existing home robots are unable to conduct a comprehensive analysis of the cleaning area during self-cleaning, are unable to partition different areas of the home, and are unable to handle the cleaning frequencies of different areas at different levels. Summary of the Invention

[0005] The purpose of the embodiment of the present invention is to provide a method for evaluating the self-cleaning technology of a household robot, aiming to solve the problem raised in the third part of the background technology.

[0006] The embodiment of the present invention is implemented as follows: a method for evaluating the self-cleaning technology of a household robot, the method comprising:

[0007] Obtain home device information, perform three-dimensional modeling based on the home device information, and obtain three-dimensional modeling information;

[0008] Analyze the microbial sensor information, compare the analysis results with the safety threshold, obtain the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, and perform a differential calculation on the initial and final values;

[0009] Update 3D modeling information based on periodic scan data, perform trend analysis on the frequency of contamination in each area based on cleaning history records, re-analyze the treated target areas, and determine the effectiveness of the treatment plan based on the secondary analysis results;

[0010] Discern emotional tendencies based on semantic protection, analyze facial emotional states and body movements based on facial expressions, and accurately identify execution deviations during tasks based on identifiable labels.

[0011] Preferably, the steps of analyzing the microbial load index information, comparing the analysis result with the safety threshold, obtaining the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, and performing differential calculation on the initial value and the final value specifically include:

[0012] Acquiring micro-biological sensing information, which is non-destructively sampling the target surface through a micro-biological sensing unit, wherein the micro-biological sensing unit includes a collection piece, a scanning module, and a micro gas sampling port;

[0013] Analyze the microbial sensor information to obtain the analysis results and the microbial safety threshold. The microbial safety threshold varies in different areas. Compare the analysis results with the safety threshold to obtain the comparison results. If the comparison result is unqualified, re-cleaning is performed.

[0014] Obtain the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the execution, perform differential calculation on the initial value and the final value, and obtain the differential calculation result. The larger the value of the calculation result, the more significant the effect of this round of cleaning task on reducing the microbial load.

[0015] Preferably, the steps of updating the three-dimensional modeling information based on the periodic scanning data, performing trend analysis on the pollution generation frequency of each area based on the cleaning history records, re-analyzing the treated target area, and judging the effectiveness of the treatment solution based on the secondary analysis results specifically include:

[0016] Acquiring periodic scanning data, including tool shifts, newly added obstacles, carpet area expansions, and narrowing of passageways, and updating 3D modeling information based on the periodic scanning data;

[0017] Acquire cleaning history records, perform trend analysis on the pollution generation frequency of each area based on the cleaning history records, and obtain analysis results, wherein the analysis structure includes a cleaning load increase and a cleaning load unchanged;

[0018] Obtain a treatment plan, which includes increasing the number of cleaning times, replacing the cleaning module, and adding a disinfection module. Re-analyze the treated target area to obtain secondary analysis results, and judge the effectiveness of the treatment plan based on the secondary analysis results.

[0019] Preferably, the steps of discriminating emotional tendencies based on semantic protective gear, analyzing facial emotional states and body movements based on facial expressions, and accurately identifying execution deviations during the task based on identifiable labels specifically include:

[0020] Acquire information from the perception module, capture semantics, facial expressions, and user-initiated feedback through the perception module, obtain semantic data, identify emotional tendencies based on semantic data, obtain the frequency and tone of keywords containing dissatisfaction, negation, or repetitive commands in the emotional tendencies, and change the cleaning task based on the emotional tendencies;

[0021] Acquire facial expression behavior analysis data, analyze facial emotional states and body movements based on the facial expressions, obtain features of facial emotional states and body movements including shaking heads and sighing, and adopt higher intensity, lower speed, or repetitive path patterns based on the facial expressions;

[0022] Active user feedback is converted into identifiable labels, and execution deviations during the task are accurately identified based on the identifiable labels, so that the cleaning task can be changed through execution deviations.

[0023] Preferably, the device information includes laser radar and ultrasonic ranging.

[0024] Another object of an embodiment of the present invention is to provide a self-cleaning technology evaluation system based on a household robot, the system comprising:

[0025] A three-dimensional modeling module obtains home device information, performs three-dimensional modeling based on the home device information, and obtains three-dimensional modeling information;

[0026] The microbial load module analyzes the microbial load information, compares the analysis results with the safety threshold, obtains the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, and performs differential calculation on the initial and final values;

[0027] The trend analysis module updates 3D modeling information based on periodic scanning data, performs trend analysis on the frequency of contamination generation in each area based on cleaning history records, re-analyzes the treated target areas, and determines the effectiveness of the treatment plan based on the secondary analysis results;

[0028] The perception module identifies emotional tendencies based on semantic information, analyzes facial emotional states and body movements based on facial expressions, and accurately identifies execution deviations during tasks based on identifiable labels.

[0029] Preferably, the microbial load module comprises:

[0030] A micro-biosensor information unit is used to obtain micro-biosensor information. The micro-biosensor information is non-destructively sampled from the target surface through the micro-biosensor unit. The micro-biosensor unit includes a collection sheet, a scanning module, and a micro gas sampling port.

[0031] The microbial safety threshold unit analyzes the microbial sensor information, obtains the analysis results, and obtains the microbial safety threshold. The microbial safety thresholds in different areas are different. The analysis results are compared with the safety threshold to obtain the comparison results. If the comparison result is unqualified, re-cleaning is performed;

[0032] The microbial load unit obtains the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the execution, performs a differential calculation on the initial value and the final value, and obtains the differential calculation result. The larger the calculation result value, the more significant the reduction effect of this round of cleaning task on the microbial load.

[0033] Preferably, the trend analysis module includes:

[0034] Updating the 3D modeling unit to obtain periodic scanning data, wherein the scanning data includes tool displacement, new obstacles, carpet area expansion, and narrowing of the passage path, and updating the 3D modeling information based on the periodic scanning data;

[0035] A trend analysis unit, which obtains cleaning history records, performs trend analysis on the pollution generation frequency of each area based on the cleaning history records, and obtains analysis results, wherein the analysis structure includes a cleaning load increase and a cleaning load unchanged;

[0036] The secondary analysis unit obtains a treatment plan, which includes increasing the number of cleaning times, replacing the cleaning module, and adding a disinfection module. The treated target area is re-analyzed to obtain a secondary analysis result, and the effectiveness of the treatment plan is judged based on the secondary analysis result.

[0037] Preferably, the perception module includes:

[0038] The semantic unit obtains information from the perception module, captures semantics, facial expressions, and user active feedback through the perception module, obtains semantic data, and discriminates emotional tendencies based on semantic data. It obtains the frequency and tone intensity of keywords containing dissatisfaction, negation, or repetitive commands in the emotional tendencies, and changes the cleaning task according to the emotional tendencies.

[0039] The facial expression behavior unit obtains facial expression behavior analysis data, analyzes the facial emotional state and body behavior based on the facial expression behavior, obtains features of the facial emotional state and body behavior including shaking head and sighing, and adopts higher intensity, lower speed or repeated path mode according to the facial expression behavior;

[0040] The user active feedback unit converts the user's active feedback into identifiable tags, accurately identifies the execution deviation during the task based on the identifiable tags, and changes the cleaning task through the execution deviation.

[0041] Preferably, the device information includes laser radar and ultrasonic ranging.

[0042] An embodiment of the present invention provides a method for evaluating a self-cleaning technology based on a household robot. The method includes obtaining household equipment information, performing three-dimensional modeling based on the household equipment information, obtaining three-dimensional modeling information, obtaining microbiological sensor information, performing non-destructive sampling of a target surface through a microbiological sensor unit, performing analysis based on the microbiological sensor information, and obtaining different microbial safety thresholds in different areas. The analysis result is compared with the safety threshold. If the comparison result is unqualified, re-cleaning is performed, and an initial value of the microbial load index before the task is executed and a final value of the microbial load index after the task is executed are obtained. The initial value and the final value are differentially calculated to obtain a differential calculation result. The larger the numerical value of the calculation result, the more significant the reduction effect of the current cleaning task on the microbial load is, and the periodic scanning is obtained. Scanning data, update three-dimensional modeling information according to periodic scanning data, conduct trend analysis on the frequency of pollution generation in each area according to the cleaning history records, obtain analysis results, re-analyze the treated target area, obtain secondary analysis results, judge the effectiveness of the treatment plan according to the secondary analysis results, obtain perception module information, capture semantics, facial expressions, and user active feedback through the perception module, judge emotional tendencies according to semantic protective gear, analyze facial emotional states and body behaviors according to facial expressions, and convert them into recognizable tags according to user active feedback, which solves the problems that existing home robots cannot conduct comprehensive analysis of cleaning areas during self-cleaning, cannot divide different areas of the home into different areas, and cannot perform different levels of processing on the cleaning frequency of different areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 A flowchart of a method for evaluating a self-cleaning technology based on a household robot provided by an embodiment of the present invention;

[0044] Figure 2 A flowchart of the steps of analyzing microbiological sensor information, comparing the analysis results with a safety threshold, and performing differential calculation between an initial value and a final value, provided in an embodiment of the present invention;

[0045] Figure 3 A flowchart of the steps of updating three-dimensional modeling information based on periodic scan data, reanalyzing the processed target area, and determining the effectiveness of the treatment solution based on the secondary analysis results, provided in an embodiment of the present invention;

[0046] Figure 4 A flowchart of the steps of discerning emotional tendencies based on semantic protective gear, analyzing facial emotional states and body movements based on facial expressions, and accurately identifying deviations during the task based on identifiable tags, provided by an embodiment of the present invention;

[0047] Figure 5This is an architecture diagram of a self-cleaning technology evaluation system based on a household robot provided by an embodiment of the present invention;

[0048] Figure 6 This is an architectural diagram of a microbial load module provided in an embodiment of the present invention;

[0049] Figure 7 An architectural diagram of a trend analysis module provided in an embodiment of the present invention;

[0050] Figure 8 This is an architectural diagram of the perception module provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, 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 used to explain the present invention and are not intended to limit the present invention.

[0052] It is understood that the terms "first," "second," etc., used herein may be used to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script without departing from the scope of this application.

[0053] like Figure 1 As shown in FIG, a method for evaluating a self-cleaning technology based on a household robot is provided in an embodiment of the present invention, and the method includes:

[0054] S100, obtaining home device information, performing three-dimensional modeling based on the home device information, and obtaining three-dimensional modeling information.

[0055] In this step, information about household devices is collected. Multimodal sensing technologies such as LiDAR, depth cameras, and image recognition are used to capture the device's geometry, location characteristics, and material information. This information is then combined with an AI model to identify specific items such as sofas, refrigerators, and TV cabinets. The recognition results are then integrated with spatial point cloud data to construct a 3D model with semantic attributes. This allows for the demarcation of functional areas such as device boundaries, risk zones, and priority cleaning zones, providing support for navigation and path planning.

[0056] During the modeling process, pollution-sensitive areas and obstacle avoidance zones are marked based on the equipment's usage status and spatial distribution. For example, the area beneath the air conditioner is designated as a high-humidity cleaning sensitive zone, and the area behind the sofa is designated as a no-go zone for roller brushes. When furniture shifts or new equipment are detected, the model is updated in real time to ensure that cleaning strategies adapt to environmental changes. Through dynamic, semantically informed 3D modeling, the robot not only improves cleaning efficiency and obstacle avoidance capabilities, but also enables precise scheduling and optimization of self-cleaning and maintenance strategies for different scenarios.

[0057] S200, analyze the microbial sensor information, compare the analysis result with the safety threshold, obtain the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, and perform differential calculation on the initial value and the final value.

[0058] In this step, the robot analyzes microbial sensor information. Before the cleaning task begins, the robot samples the target area using its microbial sensor unit. The collected information includes multiple microbial contamination characteristics, such as fluorescence response, odor concentration, and electrochemical signals. The system inputs this sampled data into a microbial load analysis model, calculates a normalized microbial load index (MLI), and compares it to a preset safety threshold for the area to determine whether there is a risk of biocontamination. If the initial MLI value exceeds the threshold, the system can preemptively adjust the cleaning intensity or trigger a self-cleaning mechanism to ensure response to high contamination conditions.

[0059] After the task is completed, the system collects microbial information from the same area again, calculates the final MLI value at the end of the task, and performs a differential analysis (ΔMLI) with the initial value to form a microbial removal efficiency index. A significant ΔMLI value indicates effective cleaning; a close or negative difference indicates poor cleaning performance or a failure. The system records this value for subsequent evaluation and determines whether additional cleaning, self-cleaning maintenance, or route optimization are needed, achieving truly closed-loop control of cleaning quality based on actual contamination effectiveness.

[0060] S300, updating the three-dimensional modeling information based on the periodic scanning data, performing a trend analysis on the pollution generation frequency of each area based on the cleaning history record, re-analyzing the treated target area, and judging the effectiveness of the treatment plan based on the secondary analysis results.

[0061] In this step, the 3D modeling information is updated based on periodic scan data. This dynamic update of the home's 3D modeling is achieved through periodic scanning, identifying spatial structural adjustments such as furniture movement and changes in area accessibility. Cleaning paths and zoning are then optimized in real time. The system also draws on cleaning history records for each area, analyzing contamination trends based on microbial load index, re-cleaning frequency, and failed task markers. This analysis determines whether the area exhibits persistent high levels of contamination, and adjusts cleaning intensity, self-cleaning frequency, or prioritization strategies accordingly.

[0062] After implementing a new cleaning strategy, the system conducts a secondary analysis of the treated target areas, recollecting microbial load data and comparing it with the pre-adjustment results to assess the reduction in contamination and the improvement in cleaning efficiency. If the test results show a significant improvement, the system marks the strategy as valid and writes it as the preferred execution plan for the current environment. If the improvement is not significant, the system triggers a strategy replacement or issues a recommendation to the user. This mechanism enables the identification of contamination trends, closed-loop verification of treatment plans, and dynamic self-optimization of cleaning strategies.

[0063] S400, identifies emotional tendencies based on semantic protection, analyzes facial emotional states and body movements based on expression behaviors, and accurately identifies execution deviations during tasks based on identifiable labels.

[0064] In this step, semantic tools are used to identify emotional tendencies, using semantic recognition and facial expression analysis modules to determine the user's emotional tendencies during the cleaning process in real time. The semantic module identifies keywords, vocal inflections, and semantic structure in the user's voice or text to detect dissatisfaction, doubt, or anxiety. The facial expression analysis module uses the camera to identify facial expressions (such as frowning and shaking heads) and body movements (such as pointing to the floor and approaching specific areas), extracting emotional features from non-verbal signals and focusing on the area of ​​interest based on the user's behavioral trajectory.

[0065] Combining emotional cues with cleaning task execution data, the system analyzes whether there are execution deviations, such as insufficient path coverage, ineffective reduction of microbial load, or insufficient mop dampness. If the system identifies an anomaly that matches the user's focus area with the task record, it is identified as an "execution deviation" and triggers a partial re-sweep or self-cleaning operation. This information is also recorded as part of the user's preference profile and used for optimized scheduling and personalized cleaning strategy recommendations in subsequent tasks, achieving a more intelligent and context-aware human-computer interaction experience.

[0066] like Figure 2 As shown, as a preferred embodiment of the present invention, the steps of parsing the microbial load index information, comparing the parsing result with the safety threshold, obtaining the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, and performing differential calculation on the initial value and the final value specifically include:

[0067] S201, obtaining micro-biological sensing information, wherein the micro-biological sensing information is non-destructively sampled on a target surface by a micro-biological sensing unit, wherein the micro-biological sensing unit includes a collection sheet, a scanning module, and a micro gas sampling port.

[0068] In this step, microbiological sensing information is obtained using a non-destructive sampling method that does not damage the floor or furniture surface. It also enables high-frequency, rapid, and low-interference data acquisition. The sensing unit integrates three functional components: a collection sheet, a scanning module, and a micro-gas sampling port. The collection sheet is made of microstructured nanofiber material with high adhesion and selective adsorption capabilities. It can briefly contact the target surface during the robot's roller brush or mop wiping process, capturing bacteria, fungal spores, or organic biological residues from the surface.

[0069] The scanning module combines a fluorescent excitation source with an optical sensor to excite and detect the collection piece or directly the target area on the ground, identifying the characteristic spectral response of typical bacterial proteins or enzyme substances; the micro-gas sampling port periodically extracts the air around the target area through a micro air pump to detect volatile organic compounds produced by microbial metabolism.

[0070] S202, analyze the microbial sensor information, obtain the analysis results, and obtain the microbial safety threshold. The microbial safety threshold is different in different areas. Compare the analysis results with the safety threshold and obtain the comparison results. If the comparison result is unqualified, re-cleaning is performed.

[0071] In this step, the microbial load analysis model is used to analyze the microbial load information and process the data to generate a standardized microbial contamination index. The analysis result is usually expressed as a normalized microbial load index (MLI), which is used to intuitively reflect the current microbial residual level in the target area.

[0072] Automatically match the corresponding microbial safety threshold based on the semantic label of the analyzed area (such as kitchen, bathroom, living room, etc.). Because the contamination tolerance and usage risk of different areas vary greatly, the threshold standards for each area are also different. For example, kitchens and bathrooms are highly sensitive areas, and their safety thresholds are usually set lower (such as MLI ≤ 60), while the thresholds for relatively clean areas such as living rooms and bedrooms can be relaxed to 70-80. The current detection area can be accurately identified through device mapping information and task records, and the threshold standard T for that area can be automatically called;

[0073] After comparing the analysis result (MLI) with the safety threshold (T), the system generates a comparison result. If MLI ≤ T, the microbial contamination in the area is within an acceptable range and the cleaning result is satisfactory. If MLI > T, residual contamination still exceeds the standard after cleaning, resulting in a failed comparison. The task is marked as a cleaning failure and automatically triggers a re-cleaning mechanism or an increased cleaning level for re-cleaning. This judgment logic serves as an important reference input for self-cleaning control strategies.

[0074] S203, obtaining the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, performing a differential calculation on the initial value and the final value, and obtaining a differential calculation result. The larger the numerical value of the calculation result, the more significant the effect of reducing the microbial load of this round of cleaning task is.

[0075] In this step, the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed are obtained. Before the cleaning task begins, the initial microbial load data of the target area is obtained through the micro-biosensor unit. After data analysis, the initial value of the microbial load index is generated, recorded as MLI1. After the cleaning task is completed, the robot samples the same location or area again and generates the final value MLI2. The difference between the two is calculated: ΔMLI = MLI1-MLI2. This difference represents the amount of microbial contamination removed by this cleaning task. The larger the ΔMLI value, the more sufficient the pollution reduction and the more significant the cleaning effect. If the difference is small or even negative, it means that the pollution has not been effectively removed, and there is even a risk of rebound or increased residue.

[0076] For example, during a kitchen cleaning task, the robot sampled the area in front of the stove and obtained an initial microbial load index (MLI) of 85. After the cleaning was complete, a second test revealed an MLI of 34, resulting in a ΔMLI of 51. The system determined that this cleaning task was effective in removing contaminants and recorded the cleaning parameters as a "high-efficiency task." However, during another task in the bedroom doorway, the MLI of 63, MLI of 59, and ΔMLI of 4 were found. The system determined the cleaning effect to be average and recorded it as an "low-efficiency task," prompting the user to increase the frequency of cleaning in this area or to check whether the self-cleaning module was operating abnormally.

[0077] like Figure 3 As shown, as a preferred embodiment of the present invention, the steps of updating the three-dimensional modeling information based on the periodic scanning data, performing trend analysis on the pollution generation frequency of each area based on the cleaning history, re-analyzing the treated target area, and judging the effectiveness of the treatment solution based on the secondary analysis results specifically include:

[0078] S301, acquiring periodic scanning data, wherein the scanning data includes tool displacement, newly added obstacles, carpet area expansion, and narrowing of the passage path, and updating the three-dimensional modeling information according to the periodic scanning data.

[0079] In this step, periodic scanning data is acquired in real time using multimodal sensing devices such as LiDAR, visual SLAM, and depth cameras. The scan data not only contains structural information about the static environment, but also focuses on furniture displacement, new obstacles, carpet area expansion, and changes in traffic paths.

[0080] The 3D modeling information is updated based on periodic scanning data to ensure the robot's path planning, self-cleaning strategy, and task efficiency evaluation.

[0081] S302 , obtaining cleaning history records, performing trend analysis on the pollution generation frequency of each area based on the cleaning history records, and obtaining analysis results, wherein the analysis structure includes cleaning load increase and cleaning load unchanged.

[0082] In this step, the cleaning history records are obtained. After each cleaning task is completed, key data such as task execution time, target area, microbial load index (MLI) changes, number of re-sweep triggers, and self-cleaning module response status are recorded and stored by area to form a cleaning history database.

[0083] During the analysis process, the system retrieves multiple cleaning records for the same area at different time points. By comparing key parameters such as the initial microbial load index (MLI1) and ΔMLI (cleaning reduction efficiency) during each task, it calculates the trend of pollutant accumulation and changes in cleaning effectiveness per unit time in that area. If the MLI1 in a particular area shows an upward trend over multiple consecutive tasks, or the ΔMLI continues to decrease, the system identifies that area as having an increasing cleaning load. Conversely, if the MLI1 remains relatively stable or decreasing, and the ΔMLI is essentially constant, the system determines that the cleaning load is unchanged.

[0084] S303, obtaining a treatment plan, which includes increasing the number of cleaning times, replacing the cleaning module, and adding a disinfection module, re-analyzing the treated target area to obtain a secondary analysis result, and judging the effectiveness of the treatment plan based on the secondary analysis result.

[0085] In this step, a treatment plan is obtained and executed based on the analysis results. This treatment plan generally includes three types of strategies: increasing the number of cleanings, replacing cleaning modules, and adding disinfection modules. These strategies are used to specifically address different types of contamination or causes of self-cleaning failures, and subsequent analysis verifies the effectiveness of the plan.

[0086] Increasing the frequency of cleaning is suitable for areas with a high frequency of contamination accumulation, but the type of contamination can be removed by conventional cleaning methods; replacing the cleaning module is suitable for areas where the robot's self-cleaning system shows signs of degradation; adding a disinfection module is suitable for areas with a high microbial load or infection risk.

[0087] After the treatment plan is implemented, the system will reanalyze the treated target area. This means that during the next or several consecutive cleaning cycles, the system will recollect parameters such as the microbial load index, cleaning difference ΔMLI, and rescan frequency in that area and compare them with the pre-treatment data. If the test results show a significant decrease in MLI, an increase in ΔMLI, and a decrease in rescan records, the current treatment plan will be judged to be effective and recorded as the recommended response plan for the current environment. Conversely, if the secondary analysis results do not change significantly or remain abnormal, the treatment plan will be considered ineffective, and the system will attempt to switch to an alternative plan or prompt the user for manual intervention.

[0088] like Figure 4 As shown, as a preferred embodiment of the present invention, the steps of discriminating emotional tendencies based on semantic protective gear, analyzing facial emotional states and body movements based on expression behaviors, and accurately identifying execution deviations during the task based on identifiable labels specifically include:

[0089] S401, obtain perception module information, capture semantics, facial expressions, and user active feedback through the perception module, obtain semantic data, judge emotional tendencies based on semantic data, obtain the frequency and tone intensity of keywords containing dissatisfaction, negation, or repeated commands in the emotional tendencies, and change the cleaning task according to the emotional tendencies.

[0090] In this step, the perception module acquires information. The perception module is responsible for real-time perception of user emotions and intentions. Through multimodal interaction, it acquires semantic data, facial expression and behavioral characteristics, and user-initiated feedback, thereby dynamically responding to changes in user cleaning satisfaction and needs. This perception module typically integrates a microphone array, camera, speech recognition system, and natural language processing model to capture user language content, intonation characteristics, facial expressions, and movements in real time.

[0091] During the semantic recognition process, the system converts user speech into text data and analyzes sentiment using a semantic protection model. This model identifies whether user statements contain typical dissatisfaction, negation, or repetitive command keywords, such as "It's still not clean even after sweeping," "Do it again," and "It's still dirty." It also analyzes these words for frequency, rising intonation, and increased speech speed, among other indicators of tone intensity. If these negative signals occur frequently, the system will label the current user's sentiment as "mild dissatisfaction," "clear negation," or "task rejection," and, based on their behavioral path, identify areas of dissatisfaction or points of concern.

[0092] When the system determines that the emotional tendency is negative and directly associated with a specific area or cleaning task, it will proactively change the cleaning task parameters or path. For example, if the user expresses "It's still dirty under the sofa" twice in a row, accompanied by a frown or pointing gesture, the system, after recognizing this semantics and behavior, will mark the area under the sofa as a "key area" and schedule the robot to clean repeatedly at a low speed, using high suction and increasing the path coverage density. At the same time, this adjustment behavior will be written into the user preference model to optimize the default cleaning strategy for similar scenarios in the future.

[0093] S402, obtaining facial expression behavior analysis data, analyzing the facial emotional state and body behavior based on the facial expression behavior, obtaining features of the facial emotional state and body behavior including shaking head and sighing, and executing with higher intensity, lower speed or repeated path mode based on the facial expression behavior.

[0094] In this step, facial expression analysis data is collected. The facial expression analysis module uses the camera to capture the user's facial expressions and body movements, identifying typical emotional signals such as frowning, shaking the head, sighing, gazing at a specific area, and pointing at the ground. When facial expression analysis identifies behavioral characteristics such as "shaking the head" or "sighing" that clearly indicate dissatisfaction or disappointment, these signals are classified as "potential cleaning dissatisfaction feedback." The system also uses the user's gaze direction or standing position at the time of the behavior to infer the user's current focus area.

[0095] For example, if a user approaches a coffee table in the living room and frequently shakes their head, sighs, and looks at the floor, the system will mark that area as an "emotional hotspot." The system will then compare the task log with the area's cleaning parameters. If it detects that the area's execution speed is too fast, the path coverage is low, or the self-cleaning module has not been triggered, the system will initially determine that the task execution intensity is insufficient.

[0096] In response to the identified negative facial expression behavior signals, the robot will automatically enable a cleaning mode with higher intensity, lower speed or repeated path as a response strategy: first, increase the suction level or mop pressure to ensure increased cleaning depth; second, reduce the movement speed, extend the processing time per unit area, and improve the decontamination efficiency; third, repeat the task path, especially for corners, edges or the bottom of furniture, etc., to perform secondary or even tertiary coverage of areas that are easily missed by conventional modes.

[0097] S403, converting the user's active feedback into a recognizable label, accurately identifying the execution deviation during the task based on the recognizable label, and changing the cleaning task based on the execution deviation.

[0098] In this step, user active feedback is converted into recognizable labels. User active feedback includes statements such as "the floor is still dirty," "corners not swept clean," "it took too long," "there's still a smell," and "there's hair on the carpet." Through semantic parsing and keyword extraction, these are converted into standardized labels such as "cleaning residue," "corners missed," "task takes too long," "odor remains," and "hair not removed." These labels are then associated with the corresponding area and task timestamp.

[0099] After acquiring the tag, the system will retrieve and analyze the operating parameters of the cleaning task, including key indicators such as path coverage, operating speed, roller brush pressure, suction level, microbial reduction index ΔMLI, and self-cleaning trigger frequency. If it is found that the path in a certain area is skipped, the cleaning power is insufficient, or the self-cleaning is not triggered, the system will determine that the "execution deviation" corresponding to the tag is real;

[0100] After identifying execution deviations, the system automatically adjusts cleaning task parameters as a response, including switching to enhanced mode, adjusting path priorities, performing partial repeat cleaning, lifting carpet restrictions, extending cleaning time, and updating regional cleaning strategies. For example, during the next cleaning, the system will prioritize the area with execution deviations in the task list and enable low-speed, high-coverage mode to ensure that user feedback is addressed.

[0101] like Figure 5 As shown in FIG, a self-cleaning technology evaluation system based on a household robot is provided in an embodiment of the present invention, and the system includes:

[0102] The three-dimensional modeling module 100 is used to obtain home device information, perform three-dimensional modeling based on the home device information, and obtain three-dimensional modeling information.

[0103] In this system, the 3D modeling module 100 acquires household device information, using multimodal sensing technologies such as LiDAR, depth cameras, and image recognition to capture the device's geometry, location characteristics, and material information. This information is then combined with AI models to identify specific items such as sofas, refrigerators, and TV cabinets. The recognition results are then integrated with spatial point cloud data to construct a 3D model with semantic attributes. This allows for the demarcation of functional areas such as device boundaries, risk zones, and priority cleaning zones, providing support for navigation and path planning.

[0104] During the modeling process, pollution-sensitive areas and obstacle avoidance zones are marked based on the equipment's usage status and spatial distribution. For example, the area beneath the air conditioner is designated as a high-humidity cleaning sensitive zone, and the area behind the sofa is designated as a no-go zone for roller brushes. When furniture shifts or new equipment are detected, the model is updated in real time to ensure that cleaning strategies adapt to environmental changes. Through dynamic, semantically informed 3D modeling, the robot not only improves cleaning efficiency and obstacle avoidance capabilities, but also enables precise scheduling and optimization of self-cleaning and maintenance strategies for different scenarios.

[0105] The microbial load module 200 is used to analyze the microbial load information, compare the analysis results with the safety threshold, obtain the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, and perform differential calculation on the initial value and the final value.

[0106] In this system, the microbial load module 200 analyzes microbial sensor information. Before a cleaning task begins, the robot samples the target area using its microbial sensor units. The collected information includes multiple microbial contamination characteristics, such as fluorescence response, odor concentration, and electrochemical signals. The system inputs this sampled data into a microbial load analysis model, calculates a normalized microbial load index (MLI), and compares it to a preset safety threshold for the area to determine whether there is a risk of biocontamination. If the initial MLI value exceeds the threshold, the system can proactively adjust the cleaning intensity or trigger a self-cleaning mechanism to ensure response to high contamination conditions.

[0107] After the task is completed, the system collects microbial information from the same area again, calculates the final MLI value at the end of the task, and performs a differential analysis (ΔMLI) with the initial value to form a microbial removal efficiency index. A significant ΔMLI value indicates effective cleaning; a close or negative difference indicates poor cleaning performance or a failure. The system records this value for subsequent evaluation and determines whether additional cleaning, self-cleaning maintenance, or route optimization are needed, achieving truly closed-loop control of cleaning quality based on actual contamination effectiveness.

[0108] The trend analysis module 300 is used to update the three-dimensional modeling information based on the periodic scanning data, perform trend analysis on the pollution generation frequency of each area based on the cleaning history records, re-analyze the treated target area, and judge the effectiveness of the treatment plan based on the secondary analysis results.

[0109] In this system, the trend analysis module 300 updates 3D modeling information based on periodic scan data. This dynamic update of the home's 3D modeling information through periodic scans identifies spatial structural adjustments such as furniture movement and changes in area accessibility, and optimizes cleaning paths and zoning accordingly. The system also draws on cleaning history records for each area, combining microbial load index, re-cleaning frequency, and failed task markers to analyze contamination trends, determine whether the area exhibits persistent high levels of contamination, and adjust cleaning intensity, self-cleaning frequency, or prioritization strategies accordingly.

[0110] After implementing a new cleaning strategy, the system conducts a secondary analysis of the treated target areas, recollecting microbial load data and comparing it with the pre-adjustment results to assess the reduction in contamination and the improvement in cleaning efficiency. If the test results show a significant improvement, the system marks the strategy as valid and writes it as the preferred execution plan for the current environment. If the improvement is not significant, the system triggers a strategy replacement or issues a recommendation to the user. This mechanism enables the identification of contamination trends, closed-loop verification of treatment plans, and dynamic self-optimization of cleaning strategies.

[0111] The perception module 400 is used to identify emotional tendencies based on semantic protection, analyze facial emotional states and body movements based on expression behaviors, and accurately identify execution deviations during the task based on identifiable labels.

[0112] In this system, the perception module 400 uses semantics to identify emotional tendencies. Using the semantic recognition and facial expression analysis modules, it assesses the user's emotional tendencies during the cleaning process in real time. The semantic module identifies keywords, vocal inflections, and semantic structure in the user's voice or text to detect dissatisfaction, doubt, or anxiety. The facial expression analysis module uses a camera to identify facial expressions (such as frowning and shaking the head) and body movements (such as pointing to the floor and approaching a specific area), extracting emotional features from non-verbal signals and, based on the user's behavioral trajectory, identifying areas of interest.

[0113] Combining emotional cues with cleaning task execution data, the system analyzes whether there are execution deviations, such as insufficient path coverage, ineffective reduction of microbial load, or insufficient mop dampness. If the system identifies an anomaly that matches the user's focus area with the task record, it is identified as an "execution deviation" and triggers a partial re-sweep or self-cleaning operation. This information is also recorded as part of the user's preference profile and used for optimized scheduling and personalized cleaning strategy recommendations in subsequent tasks, achieving a more intelligent and context-aware human-computer interaction experience.

[0114] like Figure 6 As shown, as a preferred embodiment of the present invention, the microbial load module 200 includes:

[0115] The micro-biosensor information unit 201 is used to obtain micro-biosensor information. The micro-biosensor information is non-destructively sampled on the target surface through the micro-biosensor unit. The micro-biosensor unit includes a collection piece, a scanning module and a micro gas sampling port.

[0116] In this module, the micro-biological sensing information unit 201 acquires micro-biological sensing information using a non-destructive sampling method that does not damage the floor or furniture surface. It also enables high-frequency, rapid, and low-interference data acquisition. The sensing unit integrates three functional components: a collection sheet, a scanning module, and a micro-gas sampling port. The collection sheet is made of microstructured nanofiber material with high adhesion and selective adsorption capabilities. It can briefly contact the target surface during the robot's roller brush or mop wiping process, capturing bacteria, fungal spores, or organic biological residues from the surface.

[0117] The scanning module combines a fluorescent excitation source with an optical sensor to excite and detect the collection piece or directly the target area on the ground, identifying the characteristic spectral response of typical bacterial proteins or enzyme substances; the micro-gas sampling port periodically extracts the air around the target area through a micro air pump to detect volatile organic compounds produced by microbial metabolism.

[0118] The microbial safety threshold unit 202 is used to analyze the microbial sensor information, obtain the analysis results, and obtain the microbial safety threshold. The microbial safety thresholds in different areas are different. The analysis results are compared with the safety threshold to obtain the comparison results. If the comparison result is unqualified, re-cleaning is performed.

[0119] In this module, the microbial safety threshold unit 202 analyzes the microbial sensor information, processes the data using a built-in microbial load analysis model, and generates a standardized microbial contamination index. This analysis result is typically expressed as a normalized microbial load index (MLI), which is used to intuitively reflect the current level of microbial residue in the target area.

[0120] Automatically match the corresponding microbial safety threshold based on the semantic label of the analyzed area (such as kitchen, bathroom, living room, etc.). Because the contamination tolerance and usage risk of different areas vary greatly, the threshold standards for each area are also different. For example, kitchens and bathrooms are highly sensitive areas, and their safety thresholds are usually set lower (such as MLI ≤ 60), while the thresholds for relatively clean areas such as living rooms and bedrooms can be relaxed to 70-80. The current detection area can be accurately identified through device mapping information and task records, and the threshold standard T for that area can be automatically called;

[0121] After comparing the analysis result (MLI) with the safety threshold (T), the system generates a comparison result. If MLI ≤ T, the microbial contamination in the area is within an acceptable range and the cleaning result is satisfactory. If MLI > T, residual contamination still exceeds the standard after cleaning, resulting in a failed comparison. The task is marked as a cleaning failure and automatically triggers a re-cleaning mechanism or an increased cleaning level for re-cleaning. This judgment logic serves as an important reference input for self-cleaning control strategies.

[0122] The microbial load unit 203 is used to obtain the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, perform differential calculation on the initial value and the final value, and obtain the differential calculation result. The larger the numerical value of the calculation result, the more significant the effect of reducing the microbial load of this round of cleaning task is.

[0123] In this module, the microbial load unit 203 obtains the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the execution. Before the cleaning task begins, the initial microbial load data of the target area is obtained through the micro-biosensor unit. After data analysis, the initial value of the microbial load index is generated, recorded as MLI1. After the cleaning task is completed, the robot samples again at the same location or area and generates the final value MLI2. The difference between the two is calculated: ΔMLI = MLI1-MLI2. The difference represents the amount of microbial contamination removed by this cleaning task. The larger the ΔMLI value, the more sufficient the pollution reduction and the more significant the cleaning effect. If the difference is small or even negative, it means that the pollution has not been effectively removed, and there is even a risk of rebound or increased residue.

[0124] For example, during a kitchen cleaning task, the robot sampled the area in front of the stove and obtained an initial microbial load index (MLI) of 85. After the cleaning was complete, a second test revealed an MLI of 34, resulting in a ΔMLI of 51. The system determined that this cleaning task was effective in removing contaminants and recorded the cleaning parameters as a "high-efficiency task." However, during another task in the bedroom doorway, the MLI of 63, MLI of 59, and ΔMLI of 4 were found. The system determined the cleaning effect to be average and recorded it as an "low-efficiency task," prompting the user to increase the frequency of cleaning in this area or to check whether the self-cleaning module was operating abnormally.

[0125] like Figure 7 As shown, as a preferred embodiment of the present invention, the trend analysis module 300 includes:

[0126] The updating 3D modeling unit 301 is configured to obtain periodic scanning data including tool displacement, newly added obstacles, carpet area expansion, and narrowing of passage paths, and update the 3D modeling information according to the periodic scanning data.

[0127] In this module, the 3D modeling unit 301 is updated to obtain periodic scanning data. This data is acquired in real time using multimodal sensing devices such as lidar, visual SLAM, and depth cameras. The scanning data not only contains structural information about the static environment, but also focuses on furniture displacement, new obstacles, carpet area expansion, and changes in the passage path.

[0128] The 3D modeling information is updated based on periodic scanning data to ensure the robot's path planning, self-cleaning strategy, and task efficiency evaluation.

[0129] The trend analysis unit 302 is used to obtain cleaning history records, perform trend analysis on the pollution generation frequency of each area based on the cleaning history records, and obtain analysis results. The analysis structure includes cleaning load increase and cleaning load unchanged.

[0130] In this module, the trend analysis unit 302 obtains cleaning history records. After each cleaning task is completed, key data such as task execution time, target area, microbial load index (MLI) changes, number of re-sweep triggers, and self-cleaning module response status are recorded and stored by area to form a cleaning history database.

[0131] During the analysis process, the system retrieves multiple cleaning records for the same area at different time points. By comparing key parameters such as the initial microbial load index (MLI1) and ΔMLI (cleaning reduction efficiency) during each task, it calculates the trend of pollutant accumulation and changes in cleaning effectiveness per unit time in that area. If the MLI1 in a particular area shows an upward trend over multiple consecutive tasks, or the ΔMLI continues to decrease, the system identifies that area as having an increasing cleaning load. Conversely, if the MLI1 remains relatively stable or decreasing, and the ΔMLI is essentially constant, the system determines that the cleaning load is unchanged.

[0132] The secondary analysis unit 303 is used to obtain a treatment plan, which includes increasing the number of cleaning times, replacing the cleaning module and adding a disinfection module, re-analyzing the treated target area, obtaining a secondary analysis result, and judging the effectiveness of the treatment plan based on the secondary analysis result.

[0133] In this module, the secondary analysis unit 303 obtains a treatment plan and, based on the analysis results, obtains and executes the corresponding treatment plan. This treatment plan generally includes three types of strategies: increasing the number of cleanings, replacing the cleaning module, and adding a disinfection module. These strategies are used to specifically address different types of contamination or causes of self-cleaning failures, and subsequent analysis verifies the effectiveness of the plan.

[0134] Increasing the frequency of cleaning is suitable for areas with a high frequency of contamination accumulation, but the type of contamination can be removed by conventional cleaning methods; replacing the cleaning module is suitable for areas where the robot's self-cleaning system shows signs of degradation; adding a disinfection module is suitable for areas with a high microbial load or infection risk.

[0135] After the treatment plan is implemented, the system will reanalyze the treated target area. This means that during the next or several consecutive cleaning cycles, the system will recollect parameters such as the microbial load index, cleaning difference ΔMLI, and rescan frequency in that area and compare them with the pre-treatment data. If the test results show a significant decrease in MLI, an increase in ΔMLI, and a decrease in rescan records, the current treatment plan will be judged to be effective and recorded as the recommended response plan for the current environment. Conversely, if the secondary analysis results do not change significantly or remain abnormal, the treatment plan will be considered ineffective, and the system will attempt to switch to an alternative plan or prompt the user for manual intervention.

[0136] like Figure 8 As shown, as a preferred embodiment of the present invention, the perception module 400 includes:

[0137] The semantic unit 401 is used to obtain perception module information, capture semantics, facial expressions, and user active feedback through the perception module, obtain semantic data, judge emotional tendencies based on semantic data, obtain the frequency and tone of keywords containing dissatisfaction, negation, or repeated commands in the emotional tendencies, and change the cleaning tasks according to the emotional tendencies.

[0138] In this module, the semantic unit 401 obtains information from the perception module. The perception module is responsible for real-time perception of user emotions and intentions. Through multimodal interaction, it obtains semantic data, facial expression and behavioral characteristics, and user-initiated feedback information, thereby dynamically responding to changes in user cleaning satisfaction and needs. This perception module typically integrates a microphone array, camera, speech recognition system, and natural language processing model to capture user language content, intonation characteristics, facial expressions, and movements in real time.

[0139] During the semantic recognition process, the system converts user speech into text data and analyzes sentiment using a semantic protection model. This model identifies whether user statements contain typical dissatisfaction, negation, or repetitive command keywords, such as "It's still not clean even after sweeping," "Do it again," and "It's still dirty." It also analyzes these words for frequency, rising intonation, and increased speech speed, among other indicators of tone intensity. If these negative signals occur frequently, the system will label the current user's sentiment as "mild dissatisfaction," "clear negation," or "task rejection," and, based on their behavioral path, identify areas of dissatisfaction or points of concern.

[0140] When the system determines that the emotional tendency is negative and directly associated with a specific area or cleaning task, it will proactively change the cleaning task parameters or path. For example, if the user expresses "It's still dirty under the sofa" twice in a row, accompanied by a frown or pointing gesture, the system, after recognizing this semantics and behavior, will mark the area under the sofa as a "key area" and schedule the robot to clean repeatedly at a low speed, using high suction and increasing the path coverage density. At the same time, this adjustment behavior will be written into the user preference model to optimize the default cleaning strategy for similar scenarios in the future.

[0141] The facial expression behavior unit 402 is used to obtain facial expression behavior analysis data, analyze the facial emotional state and body behavior based on the facial expression behavior, obtain the features of the facial emotional state and body behavior including shaking head and sighing, and adopt higher intensity, lower speed or repeated path mode according to the facial expression behavior.

[0142] In this module, the facial expression and behavior unit 402 acquires facial expression and behavior analysis data. Based on the user's facial expressions and body movements captured by the camera, the facial expression and behavior analysis module identifies typical emotional signals such as frowning, shaking the head, sighing, gazing at a specific area, and pointing at the ground. When behavioral characteristics such as "shaking the head" or "sighing" are identified from the facial expression and behavior analysis, these signals are classified as "potential cleaning dissatisfaction feedback." The system also infers the user's current focus area based on the user's gaze direction or standing position at the time of the behavior.

[0143] For example, if a user approaches a coffee table in the living room and frequently shakes their head, sighs, and looks at the floor, the system will mark that area as an "emotional hotspot." The system will then compare the task log with the area's cleaning parameters. If it detects that the area's execution speed is too fast, the path coverage is low, or the self-cleaning module has not been triggered, the system will initially determine that the task execution intensity is insufficient.

[0144] In response to the identified negative facial expression behavior signals, the robot will automatically enable a cleaning mode with higher intensity, lower speed or repeated path as a response strategy: first, increase the suction level or mop pressure to ensure increased cleaning depth; second, reduce the movement speed, extend the processing time per unit area, and improve the decontamination efficiency; third, repeat the task path, especially for corners, edges or the bottom of furniture, etc., to perform secondary or even tertiary coverage of areas that are easily missed by conventional modes.

[0145] The user active feedback unit 403 is used to convert the user active feedback into a recognizable label, accurately identify the execution deviation during the task according to the recognizable label, and change the cleaning task according to the execution deviation.

[0146] In this module, the user active feedback unit 403 converts user active feedback into identifiable tags. User active feedback includes sentences such as "the floor is still dirty", "corners are not cleaned clean", "it takes too long", "there is still a smell", "there are hairs on the carpet", etc. Through semantic parsing and keyword extraction mechanisms, it converts them into standardized tags such as "cleaning residue", "corners are missing", "task takes too long", "odor remains", "hair is not removed", etc., and associates them with the corresponding area and task timestamp;

[0147] After acquiring the tag, the system will retrieve and analyze the operating parameters of the cleaning task, including key indicators such as path coverage, operating speed, roller brush pressure, suction level, microbial reduction index ΔMLI, and self-cleaning trigger frequency. If it is found that the path in a certain area is skipped, the cleaning power is insufficient, or the self-cleaning is not triggered, the system will determine that the "execution deviation" corresponding to the tag is real;

[0148] After identifying execution deviations, the system automatically adjusts cleaning task parameters as a response, including switching to enhanced mode, adjusting path priorities, performing partial repeat cleaning, lifting carpet restrictions, extending cleaning time, and updating regional cleaning strategies. For example, during the next cleaning, the system will prioritize the area with execution deviations in the task list and enable low-speed, high-coverage mode to ensure that user feedback is addressed.

[0149] In one embodiment, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are performed:

[0150] Obtain home device information, perform three-dimensional modeling based on the home device information, and obtain three-dimensional modeling information;

[0151] Analyze the microbial sensor information, compare the analysis results with the safety threshold, obtain the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, and perform a differential calculation on the initial and final values;

[0152] Update 3D modeling information based on periodic scan data, perform trend analysis on the frequency of contamination in each area based on cleaning history records, re-analyze the treated target areas, and determine the effectiveness of the treatment plan based on the secondary analysis results;

[0153] Discern emotional tendencies based on semantic protection, analyze facial emotional states and body movements based on facial expressions, and accurately identify execution deviations during tasks based on identifiable labels.

[0154] In one embodiment, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the processor performs the following steps:

[0155] Obtain home device information, perform three-dimensional modeling based on the home device information, and obtain three-dimensional modeling information;

[0156] Analyze the microbial sensor information, compare the analysis results with the safety threshold, obtain the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, and perform a differential calculation on the initial and final values;

[0157] Update 3D modeling information based on periodic scan data, perform trend analysis on the frequency of contamination in each area based on cleaning history records, re-analyze the treated target areas, and determine the effectiveness of the treatment plan based on the secondary analysis results;

[0158] Discern emotional tendencies based on semantic protection, analyze facial emotional states and body movements based on facial expressions, and accurately identify execution deviations during tasks based on identifiable labels.

[0159] It should be understood that, although the various steps in the flow chart of each embodiment of the present invention are shown in sequence according to the indication of the arrows, these steps are not necessarily performed in sequence according to the order indicated by the arrows. Unless otherwise specified herein, the execution of these steps is not strictly limited in order, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0160] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When executed, the program can include the processes of the above-described method embodiments. Any reference to memory, storage, database, or other media used in the various embodiments provided herein may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM).

[0161] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned 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.

[0162] The above-described embodiments merely illustrate several implementations of the present invention, and while their descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

[0163] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for evaluating the self-cleaning technology of a household robot, characterized in that: The method comprises: Acquire home device information, perform three-dimensional modeling based on the home device information, and acquire three-dimensional modeling information; Analyze the microbial sensor information, compare the analysis results with the safety threshold, obtain the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, and perform a differential calculation on the initial and final values; Update 3D modeling information based on periodic scan data, perform trend analysis on the frequency of contamination in each area based on cleaning history records, re-analyze the treated target areas, and determine the effectiveness of the treatment plan based on the secondary analysis results; Discern emotional tendencies based on semantic protection, analyze facial emotional states and body movements based on facial expressions, and accurately identify execution deviations during tasks based on identifiable labels.

2. A method for evaluating the self-cleaning technology of a household robot according to claim 1, characterized in that: The steps of analyzing the microbial sensor information, comparing the analysis result with the safety threshold, obtaining the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, and performing differential calculation on the initial value and the final value specifically include: Acquiring micro-biological sensing information, which is non-destructively sampling the target surface through a micro-biological sensing unit, wherein the micro-biological sensing unit includes a collection piece, a scanning module, and a micro gas sampling port; Analyze the microbial sensor information to obtain the analysis results and the microbial safety threshold. The microbial safety threshold varies in different areas. Compare the analysis results with the safety threshold to obtain the comparison results. If the comparison result is unqualified, re-cleaning is performed. Obtain the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the execution, perform differential calculation on the initial value and the final value, and obtain the differential calculation result. The larger the value of the calculation result, the more significant the effect of this round of cleaning task on reducing the microbial load.

3. The method for evaluating the self-cleaning technology of a household robot according to claim 1, wherein: The steps of updating the three-dimensional modeling information based on the periodic scanning data, performing trend analysis on the pollution generation frequency of each area based on the cleaning history records, re-analyzing the treated target area, and judging the effectiveness of the treatment plan based on the secondary analysis results specifically include: Acquiring periodic scanning data, including tool shifts, newly added obstacles, carpet area expansions, and narrowing of passageways, and updating 3D modeling information based on the periodic scanning data; Acquire cleaning history records, perform trend analysis on the pollution generation frequency of each area based on the cleaning history records, and obtain analysis results, wherein the analysis structure includes a cleaning load increase and a cleaning load unchanged; Obtain a treatment plan, which includes increasing the number of cleaning times, replacing the cleaning module, and adding a disinfection module. Re-analyze the treated target area to obtain secondary analysis results, and judge the effectiveness of the treatment plan based on the secondary analysis results.

4. The method for evaluating the self-cleaning technology of a household robot according to claim 1, wherein: The steps of discriminating emotional tendencies based on semantic protective gear, analyzing facial emotional states and body movements based on facial expressions, and accurately identifying execution deviations during the task based on identifiable labels specifically include: Acquire information from the perception module, capture semantics, facial expressions, and user-initiated feedback through the perception module, obtain semantic data, identify emotional tendencies based on semantic data, obtain the frequency and intensity of keywords containing dissatisfaction, negation, or repetitive commands in the emotional tendencies, and change the cleaning task based on the emotional tendencies; Acquire facial expression behavior analysis data, analyze facial emotional states and body movements based on the facial expressions, obtain features of facial emotional states and body movements including shaking heads and sighing, and adopt higher intensity, lower speed, or repetitive path patterns based on the facial expressions; Active user feedback is converted into identifiable labels, and execution deviations during the task are accurately identified based on the identifiable labels, so that the cleaning task can be changed through execution deviations.

5. The method for evaluating the self-cleaning technology of a household robot according to claim 1, wherein: The device information includes laser radar and ultrasonic ranging.

6. A self-cleaning technology evaluation system based on a household robot, characterized in that: The system comprises: A three-dimensional modeling module obtains home device information, performs three-dimensional modeling based on the home device information, and obtains three-dimensional modeling information; The microbial load module analyzes the microbial load information, compares the analysis results with the safety threshold, obtains the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, and performs differential calculation on the initial and final values; The trend analysis module updates 3D modeling information based on periodic scanning data, performs trend analysis on the frequency of contamination generation in each area based on cleaning history records, re-analyzes the treated target areas, and determines the effectiveness of the treatment plan based on the secondary analysis results; The perception module identifies emotional tendencies based on semantic information, analyzes facial emotional states and body movements based on facial expressions, and accurately identifies execution deviations during tasks based on identifiable labels.

7. A household robot-based self-cleaning technology evaluation system according to claim 6, characterized in that: The microbial load module comprises: A micro-biosensor information unit is used to obtain micro-biosensor information. The micro-biosensor information is non-destructively sampled from the target surface through the micro-biosensor unit. The micro-biosensor unit includes a collection sheet, a scanning module, and a micro gas sampling port. The microbial safety threshold unit analyzes the microbial sensor information, obtains the analysis results, and obtains the microbial safety threshold. The microbial safety thresholds in different areas are different. The analysis results are compared with the safety threshold to obtain the comparison results. If the comparison result is unqualified, re-cleaning is performed; The microbial load unit obtains the initial value of the microbial load index before the task is executed and the final value of the microbial load index after the task is executed, performs a differential calculation on the initial value and the final value, and obtains the differential calculation result. The larger the calculation result value, the more significant the reduction effect of this round of cleaning task on the microbial load.

8. The self-cleaning technology evaluation system based on a household robot according to claim 7, characterized in that: The trend analysis module includes: Updating the 3D modeling unit to obtain periodic scanning data, wherein the scanning data includes tool displacement, new obstacles, carpet area expansion, and narrowing of the passage path, and updating the 3D modeling information based on the periodic scanning data; A trend analysis unit, which obtains cleaning history records, performs trend analysis on the pollution generation frequency of each area based on the cleaning history records, and obtains analysis results, wherein the analysis structure includes a cleaning load increase and a cleaning load unchanged; The secondary analysis unit obtains a treatment plan, which includes increasing the number of cleaning times, replacing the cleaning module, and adding a disinfection module. The treated target area is re-analyzed to obtain a secondary analysis result, and the effectiveness of the treatment plan is judged based on the secondary analysis result.

9. The self-cleaning technology evaluation system based on a household robot according to claim 8, characterized in that: The perception module includes: The semantic unit obtains information from the perception module, captures semantics, facial expressions, and user active feedback through the perception module, obtains semantic data, and discriminates emotional tendencies based on semantic data. It obtains the frequency and tone intensity of keywords containing dissatisfaction, negation, or repetitive commands in the emotional tendencies, and changes the cleaning task according to the emotional tendencies. The facial expression behavior unit obtains facial expression behavior analysis data, analyzes the facial emotional state and body behavior based on the facial expression behavior, obtains features of the facial emotional state and body behavior including shaking head and sighing, and adopts higher intensity, lower speed or repeated path mode according to the facial expression behavior; The user active feedback unit converts the user's active feedback into identifiable tags, accurately identifies the execution deviation during the task based on the identifiable tags, and changes the cleaning task through the execution deviation.

10. A household robot-based self-cleaning technology evaluation system according to claim 9, characterized in that: The device information includes laser radar and ultrasonic ranging.

Images