Cleaning strategy generation method and device, electronic equipment and storage medium
By acquiring data from sound, radar, and door magnetic sensors, and using clustering algorithms to identify scene types and formulate cleaning strategies, the problem of mismatch between cleaning plans and actual needs in intelligent cleaning systems has been solved, achieving precise cleaning solutions.
Patent Information
- Application Number
- CN202511382811.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-25
- Publication Date
- 2026-02-03
AI Technical Summary
Existing intelligent cleaning systems struggle to accurately identify stain types and cleaning needs based on the varying levels of usage in different areas of a residential or office environment. This results in a mismatch between cleaning solutions and actual needs, preventing them from achieving precise adaptation.
By acquiring data from sound sensors, radar sensors, and door magnetic sensors, spatiotemporal characteristic data is determined. Clustering algorithms are used to identify target scene types, and cleaning strategies are formulated based on scene types, including path planning and equipment parameter adjustments.
It achieves precise cleaning based on the user's activity area and time period, avoids activity areas, and plans a unique cleaning plan for each area, improving cleaning effect and resource utilization efficiency.
Smart Images
Figure CN121454908A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of device control, in particular to a cleaning strategy generation method and device, electronic device and storage medium. BACKGROUND
[0002] In the current planning of intelligent cleaning systems, the activity information of users (such as movement trajectories, spatial use frequency, etc.) is generally relied on to identify the areas to be cleaned, and accordingly the corresponding cleaning mode is automatically matched. However, this behavior statistics-based scheme has significant limitations. Since there are obvious differences in the actual use of different areas in residential or office environments, for example, the kitchen and the living room may generate more oil stains or debris due to high-frequency use, while the activity intensity of the study or bedroom is relatively low, but the cleaning standard is not necessarily lower. If the system only uses "activity frequency" as the basis for judgment, it is often difficult to accurately reflect the real stain type and cleaning demand intensity of different functional areas.
[0003] As a result, the cleaning plan generated by the system is prone to mismatch with the actual demand: high-frequency activity areas may not be able to start deep cleaning due to not being identified as "heavy pollution areas", resulting in stubborn stains remaining; on the contrary, some low-activity areas may be over-cleaned, causing resource waste or equipment wear and tear. This "one-size-fits-all" cleaning logic cannot be finely adapted to diversified scene requirements, making it difficult to achieve the user's clean standard in the final cleaning effect. SUMMARY
[0004] In view of the above problems, a cleaning strategy generation method and device, electronic device and storage medium are provided to overcome the above problems or at least partially solve the above problems, comprising:
[0005] A cleaning strategy generation method, the method comprising:
[0006] acquiring sound data collected by a sound sensor, object movement trajectory data collected by a radar sensor, and door / window state data collected by a door magnetic sensor;
[0007] determining spatio-temporal feature data based on the sound data, the object movement trajectory data, and the door / window state data;
[0008] performing clustering processing on the spatio-temporal feature data using a preset clustering algorithm to determine a target scene type of a target cleaning device;
[0009] determining a target cleaning strategy of the target cleaning device according to the target scene type.
[0010] Optionally, the step of determining spatio-temporal feature data based on the sound data, the object movement trajectory data, and the door / window state data comprises:
[0011] Based on the sound data, the object movement trajectory data and the door and window state data, a duration statistics is performed to determine a duration distribution of a target behavior mode.
[0012] Optionally, the determining of the spatio-temporal feature data based on the sound data, the object movement trajectory data and the door and window state data comprises:
[0013] Based on the sound data, the object movement trajectory data and the door and window state data, an activity density distribution map of the target object is generated.
[0014] Based on the activity density distribution map, a high-frequency activity region of the target object is labeled.
[0015] Optionally, the determining of the spatio-temporal feature data based on the sound data, the object movement trajectory data and the door and window state data comprises:
[0016] Based on the sound data, the object movement trajectory data and the door and window state data, a trajectory entropy value is determined.
[0017] According to the trajectory entropy value, regularity of a movement mode is determined.
[0018] Optionally, the determining of the spatio-temporal feature data based on the sound data, the object movement trajectory data and the door and window state data comprises:
[0019] Based on the sound data, the object movement trajectory data and the door and window state data, an association matrix of device usage is established and a spatial correlation between door and window opening and closing and object movement is determined.
[0020] Optionally, the determining of the cleaning strategy of the target cleaning device according to the scene type comprises: obtaining a corresponding relationship between scene types and cleaning strategies pre-created for the target cleaning device.
[0021] Based on the corresponding relationship, a cleaning strategy corresponding to the scene type is determined.
[0022] Optionally, the method further comprises:
[0023] According to the cleaning strategy, path planning data and device parameter adjustment data of the target cleaning device are generated;
[0024] According to the device parameter adjustment data, device parameters of the target cleaning device are adjusted;
[0025] According to the path planning data, cleaning is performed on a region to be cleaned.
[0026] Optionally, the method further comprises:
[0027] acquire user feedback information performed by a user for the target cleaning strategy;
[0028] adjust the cleaning strategy based on the user feedback information.
[0029] Optionally, the adopting a preset clustering algorithm to perform clustering processing on the spatio-temporal feature data to determine a target scene type of the target cleaning device comprises:
[0030] determining a clustering parameter corresponding to the preset clustering algorithm;
[0031] determining a target activity area according to the clustering parameter and the spatio-temporal feature data;
[0032] performing clustering on target spatio-temporal feature data corresponding to the target activity area according to the clustering algorithm to determine a target scene type of the target cleaning device.
[0033] A cleaning strategy device, the device comprising:
[0034] a data acquisition module configured to acquire sound data collected by a sound sensor, object moving track data collected by a radar sensor, and door / window state data collected by a door magnetic sensor;
[0035] a spatio-temporal feature extraction module configured to determine spatio-temporal feature data based on the sound data, the object moving track data, and the door / window state data;
[0036] a target scene type determination module configured to adopt a preset clustering algorithm to perform clustering processing on the spatio-temporal feature data to determine a target scene type of a target cleaning device;
[0037] a target cleaning strategy determination module configured to determine a target cleaning strategy of the target cleaning device according to the target scene type.
[0038] An electronic device comprising a processor, a memory, and a computer program stored on the memory and capable of running on the processor, the computer program being implemented when executed by the processor to realize the above cleaning strategy generation method.
[0039] A computer readable storage medium, the computer readable storage medium storing a computer program, the computer program being implemented when executed by a processor to realize the above cleaning strategy generation method.
[0040] A robot vacuum cleaner, the robot vacuum cleaner performing the above cleaning strategy generation method.
[0041] Embodiments of the present application have the following advantages:
[0042] In the embodiment of the present application, sound data collected by a sound sensor, object moving track data collected by a radar sensor, and door / window state data collected by a door magnetic sensor are acquired; spatio-temporal feature data is determined based on the sound data, the object moving track data, and the door / window state data; a preset clustering algorithm is used to perform clustering processing on the spatio-temporal feature data, to determine a target scene type of a target cleaning device; a target cleaning strategy of the target cleaning device is determined according to the target scene type; and the target cleaning device is controlled to execute the target cleaning strategy, so that the user's activity area is avoided and a unique cleaning plan is planned for each area based on the user's activity area and time period. BRIEF DESCRIPTION OF DRAWINGS
[0043] In order to more clearly illustrate the technical solutions of the present application, the drawings needed to be used in the description of the present application will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and all other drawings obtained by those skilled in the art without creative labor on the basis of these drawings also belong to the protection scope of the present application.
[0044] Figure 1 is a step flow chart of a cleaning strategy generation method provided by an embodiment of the present application;
[0045] Figure 2 is a step flow chart of another cleaning strategy generation method provided by an embodiment of the present application;
[0046] Figure 3 is a step flow chart of another cleaning strategy generation method provided by an embodiment of the present application;
[0047] Figure 4 is a step flow chart of another cleaning strategy generation method provided by an embodiment of the present application;
[0048] Figure 5 is a step flow chart of another cleaning strategy generation method provided by an embodiment of the present application;
[0049] Figure 6 is a structural schematic diagram of a cleaning strategy generation device provided by an embodiment of the present application. DETAILED DESCRIPTION
[0050] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments. Obviously, the described embodiments are some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor also belong to the protection scope of the present application.
[0051] Currently, in the existing scheme, the user activity information is used to realize planning of the cleaning mode corresponding to the to-be-cleaned area; due to different practical degrees of different areas, the cleaning scheme is not matched. Therefore, it is difficult to meet the demand of cleaning.
[0052] For example, in a device control method, user activity information in a to-be-cleaned area can be acquired; a cleaning mode corresponding to the to-be-cleaned area is determined according to the user activity information; and the to-be-cleaned area is cleaned according to the determined cleaning mode. This method can determine the cleaning mode of the intelligent dust removal device according to the current activity state of the user, provide a quiet living environment and working environment for the user, and reduce noise pollution.
[0053] In another cleaning method, life body information and / or target object position information in a specified area can be detected; and a cleaning mode corresponding to the specified area is determined according to the life body information and / or target object position information in the specified area. Then, the cleaning mode can be dynamically adjusted according to the actual use condition in the specified area, so as to ensure that the specified area is cleaned completely while the energy consumption is reduced to the greatest extent.
[0054] In the embodiment of the application, sound data collected by a sound sensor, object moving track data collected by a radar sensor, and door / window state data collected by a door magnetic sensor are acquired; spatio-temporal feature data is determined based on the sound data, the object moving track data, and the door / window state data; a preset clustering algorithm is used to perform clustering processing on the spatio-temporal feature data, to determine a target scene type of a target cleaning device; a target cleaning strategy of the target cleaning device is determined according to the target scene type; and the target cleaning device is controlled to execute the target cleaning strategy, so that the activity area of the user is avoided and a unique cleaning scheme is planned for each area based on the activity area and time period of the user.
[0055] Referring to Figure 1 , a step flowchart of a cleaning strategy generation method provided by an embodiment of the application is shown, and specifically can include the following steps:
[0056] In step S101, sound data collected by a sound sensor, object moving track data collected by a radar sensor, and door / window state data collected by a door magnetic sensor are acquired;
[0057] In actual application, a plurality of different types of sensor data for detecting environmental data can be pre-set in the to-be-cleaned area, such as a sound sensor, a radar sensor, a door magnetic sensor and the like. Among them, the sound sensor can be used to collect sound data in the environment, the radar sensor can be used to collect object moving track data in the environment, and the door magnetic sensor can be used to collect door and window state data. The object moving track can be a moving track of a movable object in the to-be-cleaned area, such as a moving track of a living body such as a human or a pet. The door and window state data can be specifically an opening state or a closing state of the door and window.
[0058] Specifically: the millimeter wave radar can be installed on the ceiling to monitor the human moving track (without recording image); the environmental sound sensor can collect sound pressure level spectrum (50-4000Hz), without storing specific voice content; the door magnetic sensor can record the opening and closing state of the door and window; and the intelligent electric meter can monitor the use of high-power electrical appliances (such as air conditioners and televisions).
[0059] In actual application, each type of data can be provided with corresponding collection specifications. In an example, the data collection specifications are as follows: the sampling frequency of the millimeter wave radar is 0.5HZ, and the data retention time length is 24 hours of rolling coverage; the sampling frequency of the sound sensor is 1Hz, and the data retention time length is local real-time analysis; the sampling frequency of the door magnetic sensor is event-driven, and the data retention time length is 30 days of historical record.
[0060] In step S102, space-time feature data is determined based on the sound data, the object moving track data and the door and window state data.
[0061] After obtaining the sound data, the object moving track data and the door and window state data, space-time feature data can be extracted based on the sound data, the object moving track data and the door and window state data, wherein the space-time feature data is feature data associated with time and space based on the sound data, the object moving track data and the door and window state data.
[0062] Specifically, the space-time feature data can include any one or more of the following:
[0063] Duration distribution of target behavior mode, high-frequency activity area of target object, regularity of moving mode, spatial correlation.
[0064] In an embodiment of the present application, the determination of the space-time feature data based on the sound data, the object moving track data and the door and window state data includes: performing duration statistics based on the sound data, the object moving track data and the door and window state data to determine the duration distribution of the target behavior mode.
[0065] In actual application, the 24 hours can be divided into 6 time period modules (such as early morning / morning / noon / afternoon / evening / deep night) for time period division, and then duration statistics can be performed, and specifically, duration distribution of a specific behavior mode (such as duration of continuous static state / duration of high-frequency activity) can be calculated.
[0066] In an embodiment of the present application, the determining of the spatio-temporal feature data based on the sound data, the object moving trajectory data and the door / window state data comprises: generating an activity density distribution map of the target object based on the sound data, the object moving trajectory data and the door / window state data; and labeling a high-frequency activity region of the target object based on the activity density distribution map.
[0067] In an embodiment of the present application, the target object is a movable object, and the spatio-temporal feature data can comprise an activity density distribution map of the target object and a high-frequency activity region of the target object.
[0068] In actual application, the activity density distribution map of the target object can be determined through comprehensive analysis of the sound data, the moving trajectory and the door / window state, wherein the activity density distribution map is a data visualization chart, and a certain specific index (such as the number of people, the number of times, the duration, the frequency) can be intuitively displayed on a certain spatial or time dimension in terms of the concentration degree and the distribution mode through the color depth, the point density or the size of the heat region.
[0069] After obtaining the activity density distribution map, the high-frequency activity region of the target object can be determined from the activity density distribution map, for example, when the frequency of the target object appearing in a certain position or region (the entire cleaning region can be pre-divided into small regions of a preset size) is higher than a preset frequency, the region or the position can be determined as the high-frequency activity region.
[0070] In an embodiment of the present application, the determining of the spatio-temporal feature data based on the sound data, the object moving trajectory data and the door / window state data comprises: determining a trajectory entropy value based on the sound data, the object moving trajectory data and the door / window state data; and determining the regularity of the moving mode according to the trajectory entropy value.
[0071] In an embodiment of the present application, the trajectory entropy value is calculated according to the formula H = -∑p(x)logp(x).
[0072] In an embodiment of the present application, H represents the entropy value. In the moving path analysis, the entropy value is a quantitative index of the “complexity” or “regularity” calculated in the embodiment of the present application. The higher the H value is, the more complex the moving path is, the more unpredictable the moving path is, and the worse the regularity is. The lower the H value is, the simpler the moving path is, the more regular the moving path is, and the stronger the predictability is.
[0073] p(x): represents the frequency of the target object selecting route x among several fixed routes.
[0074] In an embodiment of the present application, the determination of the space-time feature data based on the sound data, the object moving track data and the door and window state data comprises: establishing a correlation matrix of device usage based on the sound data, the object moving track data and the door and window state data and determining the spatial correlation between door and window opening and closing and object movement.
[0075] In actual application, behavior correlation mining can be performed. Specifically, the correlation mining can be divided into appliance usage correlation: a device usage correlation matrix is established, that is, when one device is used, the usage of another device is correlated (for example, opening the air conditioner is often accompanied by the behavior of closing the window); the correlation mining can also include door and window linkage mode: the spatial correlation between door and window opening and closing and personnel movement can be analyzed (for example, opening the balcony door is often accompanied by personnel going out).
[0076] In actual application, the behavior correlation mining can be performed in the following manner:
[0077] 1. The data source is obtained by recording user operation and device running state through the intelligent socket and the home platform.
[0078] 2. The correlation strength between different devices is evaluated by support (the frequency of simultaneous occurrence of different devices), confidence (the probability of device B being turned on under the condition that device A is turned on), and lift (the significance of correlation). For example, support data, confidence data and lift data and other index data for representing correlation strength can be obtained. In order to avoid too large difference between the bases of various index data, the normalization processing is performed on the various index data to obtain data within a preset range, and then the weight information of each index data is obtained. The correlation strength score between devices is obtained by weighting processing according to each index data and the corresponding weight data. The correlation strength score can directly reflect the correlation between devices. The higher the correlation strength score, the stronger the correlation between devices, and the lower the correlation strength score, the weaker the correlation between devices.
[0079] That is, the correlation strength score = index data 1 * weight data 1 +... + index data n * weight data n. Wherein, weight data 1 +... + weight data n = 1.
[0080] When the correlation strength score between devices is higher, the corresponding device can be triggered after the corresponding device is triggered in the environment, thereby forming a corresponding scene. The corresponding cleaning strategy can be started for the scene.
[0081] Similarly, when the device is associated with the user behavior, the correlation strength between devices can be referred to for correlation score calculation.
[0082] 3. Devices are arranged in rows, and values are associated strengths.
[0083] 4. The association matrix is updated periodically according to user behavior.
[0084] Step S103, using a preset clustering algorithm to cluster the spatio-temporal feature data, and determining the target scene type of the target cleaning device.
[0085] After obtaining the spatio-temporal feature data, a clustering algorithm can be used for clustering, so as to determine the target scene type existing in the target cleaning area. The target scene type can be composed of multiple features, and a trigger condition can be set.
[0086] For example, the target scene type can be divided into a home office scene, a family gathering scene, and a sleep period scene.
[0087] Among them, the characteristics of the home office scene are low movement entropy + continuous appliance use, and the trigger condition of this scene can be computer power consumption > 200W from 9:00 to 17:00 on weekdays.
[0088] The characteristics of the family gathering scene are high frequency sound pressure mutation + multi-region movement, and the trigger condition of this scene can be sound pressure fluctuation > 20dB from 19:00 to 22:00 on weekends.
[0089] The characteristics of the sleep period scene are zero movement + continuous door magnet closure, and the trigger condition of this scene can be no door and window changes from 23:00 to 6:00.
[0090] Step S104, determining the target cleaning strategy of the target cleaning device according to the target scene type.
[0091] In actual application, after determining the scene type, the cleaning strategy that meets the characteristics of the scene type can be set according to the scene type. Thus, a cleaning plan unique to each area can be realized. As shown in Table 1, it is a setting mode of a cleaning strategy in the embodiment of the application.
[0092]
[0093] In actual application, the priority of the cleaning strategy is set as follows:
[0094] (1) Prioritize deep cleaning: corresponding to the deep cleaning period (such as weekends or holidays), full house coverage is required, which is a high priority task.
[0095] (2) Environmentally intelligent avoidance: corresponding to the high activity period (such as daytime activity), dynamic adjustment is required, which is a medium priority task.
[0096] (3) Daily personal cleaning: corresponding to the sleep period (such as night), it needs to run with low interference, and it belongs to a low-priority task.
[0097] In the embodiment of the application, sound data collected by a sound sensor, object moving track data collected by a radar sensor, and door / window state data collected by a door magnetic sensor are acquired; spatio-temporal feature data is determined based on the sound data, the object moving track data, and the door / window state data; a preset clustering algorithm is used to perform clustering processing on the spatio-temporal feature data to determine a target scene type of the target cleaning device; and a target cleaning strategy of the target cleaning device is determined according to the target scene type, so that the user's activity area is avoided and a unique cleaning plan for each area is planned based on the user's activity area and time period.
[0098] Referring to Figure 2 , a step flowchart of another cleaning strategy generation method provided by an embodiment of the application is shown, which can specifically include the following steps:
[0099] In step S201, sound data collected by a sound sensor, object moving track data collected by a radar sensor, and door / window state data collected by a door magnetic sensor are acquired.
[0100] In actual application, a plurality of different types of sensor data for detecting environmental data can be pre-set in the area to be cleaned, such as a sound sensor, a radar sensor, a door magnetic sensor, etc. Among them, the sound sensor is used to collect sound data in the environment, the radar sensor is used to collect object moving track data in the environment, and the door magnetic sensor is used to collect door / window state data.
[0101] Specifically, the millimeter wave radar can be installed on the ceiling to monitor the human moving track (without recording the image); the environmental sound sensor can collect the sound pressure level spectrum (50-4000Hz) without storing specific voice content; the door magnetic sensor can record the door / window opening and closing state; and the intelligent electric meter can monitor the use of high-power electrical appliances (such as air conditioners and televisions).
[0102] In actual application, each type of data is provided with a corresponding data collection specification. In an example, the data collection specification is as follows: the sampling frequency of the millimeter wave radar is 0.5HZ, and the data retention time length is 24 hours of rolling coverage; the sampling frequency of the sound sensor is 1Hz, and the data retention time length is local real-time analysis; the sampling frequency of the door magnetic sensor is event-driven, and the data retention time length is 30 days of historical record.
[0103] In step S202, spatio-temporal feature data is determined based on the sound data, the object moving track data, and the door / window state data.
[0104] After obtaining the sound data, the object moving track data and the door and window state data, spatio-temporal feature data can be extracted based on the sound data, the object moving track data and the door and window state data, wherein the spatio-temporal feature data is feature data associated with time and space based on the sound data, the object moving track data and the door and window state data.
[0105] Specifically, the spatio-temporal feature data can include any one or more of the following:
[0106] The duration distribution of the target behavior mode, the high-frequency activity area of the target object, the regularity of the moving mode, and the spatial correlation.
[0107] In an embodiment of the present application, determining the spatio-temporal feature data based on the sound data, the object moving track data and the door and window state data includes: performing duration statistics based on the sound data, the object moving track data and the door and window state data to determine the duration distribution of the target behavior mode.
[0108] In step S203, a preset clustering algorithm is used to cluster the spatio-temporal feature data to determine the target scene type of the target cleaning device.
[0109] In step S204, a corresponding relationship between the scene type and the cleaning strategy pre-created for the target cleaning device is obtained.
[0110] In actual application, a cleaning strategy can be pre-created for a possible scene type in a cleaning area. Specifically, the corresponding relationship between the scene type and the cleaning strategy can be generated according to the cleaning records of previous users. In the corresponding relationship, the corresponding cleaning strategy that can be adapted in a certain specific scene type can be recorded. The cleaning strategy can be set for the characteristics of the scene type, thereby effectively improving the cleaning efficiency.
[0111] In step S205, the cleaning strategy corresponding to the scene type is determined based on the corresponding relationship.
[0112] In an embodiment of the present application, the sound data collected by the sound sensor, the object moving track data collected by the radar sensor and the door and window state data collected by the door magnetic sensor are obtained; the spatio-temporal feature data is determined based on the sound data, the object moving track data and the door and window state data; the spatio-temporal feature data is clustered using a preset clustering algorithm to determine the target scene type of the target cleaning device; and the target cleaning strategy of the target cleaning device is determined according to the target scene type, thereby realizing the avoidance of the activity area of the user and the planning of an exclusive cleaning plan for each area based on the activity area and the time period of the user.
[0113] Reference Figure 3Fig. 3 shows a flow chart of steps of another method for generating a cleaning strategy according to an embodiment of the present application, which can include the following steps:
[0114] In step S301, sound data collected by a sound sensor, object moving track data collected by a radar sensor, and door / window state data collected by a door magnetic sensor are acquired.
[0115] In actual applications, a plurality of sensors for detecting environmental data of different types can be pre-set in the area to be cleaned, such as a sound sensor, a radar sensor, a door magnetic sensor, etc. Among them, the sound sensor is used to collect sound data in the environment, the radar sensor is used to collect object moving track data in the environment, and the door magnetic sensor is used to collect door / window state data.
[0116] Specifically, the millimeter wave radar can be installed on the ceiling to monitor human moving tracks (without recording images); the environmental sound sensor can collect sound pressure level spectrum (50-4000Hz) without storing specific voice content; the door magnetic sensor can record the opening and closing state of doors and windows; and the smart meter can monitor the use of high-power electrical appliances (such as air conditioners and televisions).
[0117] In actual applications, each type of data is provided with corresponding collection specifications. In an example, the data collection specifications are as follows: the sampling frequency of the millimeter wave radar is 0.5HZ, and the data retention time is 24 hours of rolling coverage; the sampling frequency of the sound sensor is 1Hz, and the data retention time is local real-time analysis; the sampling frequency of the door magnetic sensor is event-driven, and the data retention time is 30 days of historical records.
[0118] In step S302, spatio-temporal feature data is determined based on the sound data, the object moving track data, and the door / window state data.
[0119] After acquiring the sound data, the object moving track data, and the door / window state data, spatio-temporal feature data can be extracted based on the sound data, the object moving track data, and the door / window state data, wherein the spatio-temporal feature data is feature data associated with time and space based on the sound data, the object moving track data, and the door / window state data.
[0120] Specifically, the spatio-temporal feature data can include any one or more of the following:
[0121] The duration distribution of the target behavior mode, the high-frequency activity area of the target object, the regularity of the moving mode, and the spatial correlation.
[0122] In step S303, a preset clustering algorithm is used to cluster the spatio-temporal feature data to determine a target scene type of a target cleaning device.
[0123] After obtaining the spatio-temporal feature data, a clustering algorithm can be used for clustering, so that the target scene type existing in the target cleaning area can be determined. The target scene type can be composed of multiple features, and a trigger condition can be set.
[0124] For example, the target scene type can be divided into a home office scene, a family gathering scene, and a sleep period scene.
[0125] The characteristics of the home office scene are low movement entropy + continuous appliance use, and the trigger condition of this scene can be computer power consumption > 200W from 9:00 to 17:00 on weekdays.
[0126] The characteristics of the family gathering scene are high-frequency sound pressure mutation + multi-zone movement, and the trigger condition of this scene can be sound pressure fluctuation > 20dB from 19:00 to 22:00 on weekends.
[0127] The characteristics of the sleep period scene are zero movement + continuous door magnet closure, and the trigger condition of this scene can be no door / window change from 23:00 to 6:00.
[0128] Step S304, determining a target cleaning strategy of the target cleaning device according to the target scene type;
[0129] In actual application, after determining the scene type, a cleaning strategy that meets the characteristics of the scene type can be set according to the scene type. Thus, a cleaning plan unique to each area can be realized.
[0130] Step S305, generating path planning data and device parameter adjustment data of the target cleaning device according to the cleaning strategy;
[0131] After determining the cleaning strategy, an executable scheme can be generated based on the cleaning strategy to clean the area to be cleaned. Specifically, according to the cleaning strategy, the distribution of obstacles in the area to be cleaned and the like can be considered, and the path planning data executable by the target cleaning device and the specific device parameter adjustment data of the target cleaning device can be executed, wherein the target cleaning device can be a cleaning device such as a sweeping robot. The target cleaning device can be connected with a central control device, the central control device can collect multiple sensor data and analyze and execute the cleaning strategy, and then issue specific control instructions to the target cleaning device to realize precise cleaning.
[0132] After determining the cleaning strategy, the cleaning strategy is actually a cleaning target for the to-be-cleaned area. According to the cleaning target, the actual situation in the to-be-cleaned area (such as the positions of fixed objects and moving objects in the space of the to-be-cleaned area) can be combined to generate a scheme executable by the target cleaning device, that is, to determine the path planning data and device parameter adjustment data of the target cleaning device, wherein the path planning data can include a global target cleaning route, an avoidance mode triggered according to a real-time dynamic environment in an actual cleaning process, path planning for cleaning an avoidance area, a route frequency, and the like.
[0133] For example, when the target cleaning device cleans according to the preset target cleaning route, the dynamic changes of surrounding objects can be learned at a preset frequency. When it is detected that there is a target object in front of the target cleaning route, the route for avoiding the target object can be re-planned according to the specific situation of the target object (such as the size of the object or the predicted moving direction of the object) on the basis of the target cleaning route, so that the target object can be dynamically avoided, safe cleaning can be realized, and the avoided area can be recorded. When it is detected that there is no target object in the area, the route can be re-planned according to the current position of the target cleaning device and the unfinished target cleaning route to complete the cleaning in the avoided area.
[0134] In the embodiments of the present application, the device parameter adjustment data can include device movement parameters (such as cleaning movement speed, steering / angle speed, acceleration, and the like), and can also include cleaning system parameters (such as main brush rotation speed, suction fan power, side brush control, cleaning power duration, and the like), perception parameters (such as pedestrian trajectory prediction, sensor refresh rate, and the like).
[0135] Different cleaning strategies can be provided with corresponding device parameter adjustment information, which can control the target cleaning device to complete the cleaning target corresponding to the cleaning strategy.
[0136] Step S306: adjusting the device parameters of the target cleaning device according to the device parameter adjustment data;
[0137] After obtaining the device parameter adjustment data, the obtained device parameter adjustment data can be used to cover the original device parameters of the target cleaning device, so as to adjust the device parameters of the target cleaning device according to the device parameter adjustment data.
[0138] Step S307: performing cleaning on the to-be-cleaned area according to the path planning data.
[0139] In the embodiment of the present application, sound data collected by a sound sensor, object movement trajectory data collected by a radar sensor, and door / window state data collected by a door magnetic sensor are acquired; spatio-temporal feature data is determined based on the sound data, the object movement trajectory data, and the door / window state data; a preset clustering algorithm is used to perform clustering processing on the spatio-temporal feature data to determine a target scene type of a target cleaning device; and a target cleaning strategy of the target cleaning device is determined according to the target scene type, so that the user activity area and time period are avoided and a unique cleaning plan for each area is planned.
[0140] Referring to Figure 4 , a step flowchart of another cleaning strategy generation method provided by an embodiment of the present application is shown, which can specifically include the following steps:
[0141] In step S401, sound data collected by a sound sensor, object movement trajectory data collected by a radar sensor, and door / window state data collected by a door magnetic sensor are acquired.
[0142] In actual application, a plurality of different types of sensor data for detecting environmental data can be pre-set in the area to be cleaned, such as a sound sensor, a radar sensor, a door magnetic sensor, etc. Among them, the sound sensor is used to collect sound data in the environment, the radar sensor is used to collect object movement trajectory data in the environment, and the door magnetic sensor is used to collect door / window state data.
[0143] Specifically, the millimeter wave radar can be installed on the ceiling to monitor the human movement trajectory (without recording the image); the environmental sound sensor can collect the sound pressure level spectrum (50-4000Hz) without storing specific voice content; the door magnetic sensor can record the door / window opening and closing state; and the intelligent electric meter can monitor the use of high-power electrical appliances (such as air conditioners and televisions).
[0144] In actual application, each type of data is provided with corresponding collection specifications. In an example, the data collection specifications are as follows: the sampling frequency of the millimeter wave radar is 0.5HZ, and the data retention time length is 24 hours of rolling coverage; the sampling frequency of the sound sensor is 1Hz, and the data retention time length is local real-time analysis; the sampling frequency of the door magnetic sensor is event-driven, and the data retention time length is 30 days of historical record.
[0145] In step S402, spatio-temporal feature data is determined based on the sound data, the object movement trajectory data, and the door / window state data.
[0146] After obtaining the sound data, the object moving track data and the door and window state data, spatio-temporal feature data can be extracted based on the sound data, the object moving track data and the door and window state data, wherein the spatio-temporal feature data is feature data associated with time and space based on the sound data, the object moving track data and the door and window state data.
[0147] Specifically, the spatio-temporal feature data can include any one or more of the following:
[0148] a duration distribution of a target behavior mode, a high-frequency activity area of a target object, regularity of a moving mode, and spatial correlation.
[0149] In an embodiment of the present application, determining the spatio-temporal feature data based on the sound data, the object moving track data and the door and window state data includes: performing duration statistics based on the sound data, the object moving track data and the door and window state data to determine a duration distribution of a target behavior mode.
[0150] Step S403: performing clustering processing on the spatio-temporal feature data using a preset clustering algorithm to determine a target scene type of the target cleaning device.
[0151] Step S404: determining a target cleaning strategy of the target cleaning device according to the target scene type.
[0152] Step S405: obtaining user feedback information executed by a user for the target cleaning strategy.
[0153] In actual application, in order to make the cleaning plan more suitable for user needs, user feedback information executed by a user for the current cleaning strategy can be obtained after each execution of the cleaning strategy.
[0154] Step S406: adjusting the cleaning strategy based on the user feedback information.
[0155] After obtaining the user feedback information, the type of the feedback information can be determined, for example, if the user feedback information is positive feedback, the cleaning strategy can be continued to be used, and if the user feedback is negative feedback, the cleaning strategy needs to be adjusted accordingly.
[0156] In an embodiment of the present application, when the user feedback is negative feedback, the reason type of the negative feedback can be determined according to the negative feedback, and the existing implementation mode corresponding to the cleaning strategy can be adjusted according to the reason type, in an example, the path planning data and the device parameter adjustment data of the target cleaning device corresponding to the cleaning strategy can be adjusted to avoid the negative feedback.
[0157] Specifically, when the negative feedback of the user is of the 'cleaning cleanliness' type (for example, not clean enough, etc.), the cleaning strategy can be adjusted by increasing the effective cleaning operation time and effect of the cleaning device under the current cleaning strategy, for example, adjusting from the path planning data can include adding the number of cleaning tasks (i.e. cleaning multiple times), and adjusting from the device parameter adjustment data can include extending the cleaning power duration.
[0158] When the negative feedback of the user is of the 'affecting user activity' type, the cleaning strategy can be adjusted by improving the intelligent obstacle avoidance capability of the cleaning device, for example, from the path planning data, the obstacle avoidance mode can be adjusted and optimized (such as adjusting the safety distance of the detected target object), and from the device parameter adjustment data, the speed buffer during obstacle avoidance can be set, that is, when a target object is detected, the speed of the device is slowly reduced from the current value to the target value, reducing the sense of strangeness, in addition, a pre-prepared voice can be played when triggering obstacle avoidance to remind the user that the target cleaning device is currently in an obstacle avoidance state.
[0159] In the embodiment of the application, sound data collected by a sound sensor, object movement trajectory data collected by a radar sensor, and door / window state data collected by a door magnetic sensor are acquired; spatio-temporal feature data is determined based on the sound data, the object movement trajectory data, and the door / window state data; a preset clustering algorithm is used to perform clustering processing on the spatio-temporal feature data to determine a target scene type of a target cleaning device; a target cleaning strategy of the target cleaning device is determined according to the target scene type; and the target cleaning device is controlled to execute the target cleaning strategy, thereby avoiding the activity area of the user and planning a unique cleaning plan for each area based on the activity area and time period of the user.
[0160] Referring to Figure 5 , a step flowchart of a cleaning strategy generation method in an embodiment of the application is shown, and specifically includes the following steps:
[0161] (1) Sensor network deployment: In the area to be cleaned, multi-source sensors such as sound sensors, millimeter wave radars, and door magnetic sensors can be deployed, and privacy protection modules are set for the multi-source sensors for data desensitization, local storage, and encrypted transmission of collected data.
[0162] (2) Multi-source data collection is performed using sensors of various types, and then the collected data can be preprocessed, and after preprocessing, feature extraction can be performed.
[0163] (3) The extracted features are analyzed using a spatio-temporal clustering method to determine the scene type of cleaning.
[0164] (4) Further, according to the scene type, a cleaning strategy is determined, such as a high-activity period strategy, an unattended home strategy, and a sleep period strategy. A dynamic path planning and device parameter adjustment are generated according to the cleaning strategy, and the dynamic path planning and device parameter adjustment are adjusted.
[0165] (5) After the strategy is executed, an effect is evaluated, and user feedback is collected, so that the scene recognition process can be optimized based on the user feedback to optimize the cleaning strategy.
[0166] It should be noted that, for the method embodiment, in order to simply describe, it is expressed as a series of action combinations, but those skilled in the art should know that the embodiments of the present application are not limited by the order of the described actions, because according to the embodiments of the present application, certain steps can be performed in other order or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily necessary for the embodiments of the present application.
[0167] Referring to Figure 6 , a structure schematic diagram of a cleaning strategy generation device provided by an embodiment of the present application is shown, which can specifically include the following modules:
[0168] The data acquisition module 601 is configured to acquire sound data collected by a sound sensor, object movement trajectory data collected by a radar sensor, and door / window state data collected by a door magnetic sensor.
[0169] The space-time feature extraction module 602 is configured to determine space-time feature data based on the sound data, the object movement trajectory data, and the door / window state data.
[0170] The target scene type determination module 603 is configured to perform clustering processing on the space-time feature data by using a preset clustering algorithm, and determine a target scene type of a target cleaning device.
[0171] The target cleaning strategy determination module 604 is configured to determine a target cleaning strategy of the target cleaning device according to the target scene type.
[0172] In an embodiment of the present application, the space-time feature extraction module 602 can include:
[0173] The first feature extraction sub-module is configured to perform duration statistics based on the sound data, the object movement trajectory data, and the door / window state data, and determine a duration distribution of a target behavior mode.
[0174] In an embodiment of the present application, the space-time feature extraction module 602 can include:
[0175] a second feature extraction submodule configured to generate an activity density distribution map of the target object based on the sound data, the object moving trajectory data, and the door / window state data;
[0176] a third feature extraction submodule configured to label a high-frequency activity region of the target object based on the activity density distribution map.
[0177] In an embodiment of the present application, the spatio-temporal feature extraction module 602 can include:
[0178] a fourth feature extraction submodule configured to determine a trajectory entropy value based on the sound data, the object moving trajectory data, and the door / window state data;
[0179] a fifth feature extraction submodule configured to determine regularity of a moving pattern according to the trajectory entropy value.
[0180] In an embodiment of the present application, the spatio-temporal feature extraction module 602 can include:
[0181] a sixth feature extraction submodule configured to establish an association matrix of device usage and determine spatial correlation between door / window opening / closing and object moving based on the sound data, the object moving trajectory data, and the door / window state data.
[0182] In an embodiment of the present application, the target cleaning strategy determination module 603 can include:
[0183] a corresponding relationship acquisition submodule configured to acquire a corresponding relationship between a scene type and a cleaning strategy pre-created for the target cleaning device;
[0184] a cleaning strategy determination submodule configured to determine a cleaning strategy corresponding to the scene type based on the corresponding relationship.
[0185] In an embodiment of the present application, the apparatus further includes:
[0186] a path and parameter data determination module configured to generate path planning data and device parameter adjustment data of the target cleaning device according to the cleaning strategy;
[0187] a device parameter adjustment module configured to adjust device parameters of the target cleaning device according to the device parameter adjustment data;
[0188] a cleaning execution module configured to perform cleaning on a region to be cleaned according to the path planning data.
[0189] In an embodiment of the present application, the apparatus can further include:
[0190] a user feedback information acquisition module configured to acquire user feedback information generated by a user for the target cleaning strategy.
[0191] an adjusting module, configured to adjust the cleaning strategy based on the user feedback information.
[0192] In an embodiment of the present application, the target scene type determination module 603 can include:
[0193] a clustering parameter determination sub-module, configured to determine a clustering parameter corresponding to the preset clustering algorithm;
[0194] a target activity area determination sub-module, configured to determine a target activity area according to the clustering parameter and the spatio-temporal feature data;
[0195] a target scene type determination sub-module, configured to perform clustering on target spatio-temporal feature data corresponding to the target activity area according to the clustering algorithm, and determine a target scene type of the target cleaning device.
[0196] In an embodiment of the present application, sound data collected by a sound sensor, object moving track data collected by a radar sensor, and door / window state data collected by a door magnetic sensor are acquired; spatio-temporal feature data is determined based on the sound data, the object moving track data, and the door / window state data; a preset clustering algorithm is used to perform clustering processing on the spatio-temporal feature data, to determine a target scene type of the target cleaning device; a target cleaning strategy of the target cleaning device is determined according to the target scene type; and the target cleaning device is controlled to execute the target cleaning strategy, thereby avoiding the activity area of a user and planning a unique cleaning plan for each area based on the activity area and time period of the user.
[0197] An embodiment of the present application further provides an electronic device, which can include a processor, a memory, and a computer program stored on the memory and capable of running on the processor, and the computer program is executed by the processor to implement the above cleaning strategy generation method.
[0198] An embodiment of the present application further provides a computer readable storage medium, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the above cleaning strategy generation method.
[0199] An embodiment of the present application further provides a sweeping robot, which executes the above cleaning strategy generation method.
[0200] For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and relevant parts refer to the part of the method embodiment.
[0201] The various embodiments described in this specification are intended to be illustrative only. Each embodiment was chosen for illustration only, and not as a limitation of the scope of the disclosure. Numerous alternatives not specifically set forth herein will be apparent in view of the teachings herein.
[0202] Those skilled in the art will appreciate that embodiments of the present application can be readily used as software, hardware, or a combination of software and hardware. In a software embodiment, various embodiments of the present application can be implemented by computer software programs or
[0203] Embodiments of the present application are described herein with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing device or other programmable data processing terminal devices to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal devices, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.
[0204] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing terminal devices to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.
[0205] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal devices to cause a series of operational steps to be performed on the computer or other programmable terminal devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable terminal devices provide steps for implementing the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.
[0206] While preferred embodiments of the application have been described, those skilled in the art will appreciate that other modifications than those specifically described can be made within the scope of the present application. Accordingly, the appended claims are intended to embrace all such alternatives as well as the embodiments specifically described.
[0207] Finally, it should be noted that, in the present document, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily implying any actual relationship or order between such entities or actions. Also, the terms "comprises", "comprising", or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the presence of additional identical elements in the process, method, article, or apparatus that comprises the element.
[0208] The above provides a cleaning strategy generation method and device, electronic equipment and storage medium, the principle and implementation mode of the present application are described in the present document by applying specific examples, the above example is only used to help understand the method and core idea of the present application; At the same time, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed; In view of the above, the content of the specification should not be understood as the limitation of the present application.
Claims
1. A cleaning policy generation method characterized by comprising: The method comprises: acquiring sound data collected by a sound sensor, object movement trajectory data collected by a radar sensor, and door / window state data collected by a door magnetic sensor; determining spatio-temporal feature data based on the sound data, the object movement trajectory data, and the door / window state data; performing clustering processing on the spatio-temporal feature data using a preset clustering algorithm to determine a target scene type of a target cleaning device; determining a target cleaning strategy of the target cleaning device according to the target scene type.
2. The method of claim 1, wherein, The determination of the spatio-temporal feature data based on the sound data, the object movement trajectory data, and the door / window state data comprises: performing duration statistics based on the sound data, the object movement trajectory data, and the door / window state data to determine a duration distribution of a target behavior mode.
3. The method of claim 1, wherein, The determination of the spatio-temporal feature data based on the sound data, the object movement trajectory data, and the door / window state data comprises: generating an activity density distribution map of a target object based on the sound data, the object movement trajectory data, and the door / window state data; annotating a high-frequency activity region of the target object based on the activity density distribution map.
4. The method of claim 1, wherein, The determination of the spatio-temporal feature data based on the sound data, the object movement trajectory data, and the door / window state data comprises: determining a trajectory entropy value based on the sound data, the object movement trajectory data, and the door / window state data; determining regularity of a movement mode according to the trajectory entropy value.
5. The method of claim 1, wherein, The determination of the spatio-temporal feature data based on the sound data, the object movement trajectory data, and the door / window state data comprises: establishing an association matrix of device usage and determining spatial correlation between door / window opening and closing and object movement based on the sound data, the object movement trajectory data, and the door / window state data.
6. The method of claim 1, wherein, The determination of the cleaning strategy of the target cleaning device according to the scene type comprises: acquiring a corresponding relationship between a scene type and a cleaning strategy that is pre-created for the target cleaning device; determining a cleaning strategy corresponding to the scene type based on the corresponding relationship.
7. The method of claim 1, wherein, Further comprising: generating path planning data and device parameter adjustment data of the target cleaning device according to the cleaning strategy; adjusting device parameters of the target cleaning device according to the device parameter adjustment data; performing cleaning on a cleaning area according to the path planning data.
8. The method of claim 1, wherein, Further comprising: acquiring user feedback information performed by a user for the target cleaning strategy; adjusting the cleaning strategy based on the user feedback information.
9. The method of claim 1, wherein, The determination of the target scene type of the target cleaning device by performing clustering processing on the spatio-temporal feature data using a preset clustering algorithm comprises: determining clustering parameters corresponding to the preset clustering algorithm; determining a target activity region according to the clustering parameters and the spatio-temporal feature data; performing clustering on target spatio-temporal feature data corresponding to the target activity region according to the clustering algorithm to determine a target scene type of a target cleaning device.
10. A cleaning policy device, characterized by The apparatus comprises: a data acquisition module configured to acquire sound data collected by a sound sensor, object movement trajectory data collected by a radar sensor, and door / window state data collected by a door magnetic sensor; The space-time feature extraction module is configured to determine space-time feature data based on the sound data, the object moving track data, and the door and window state data; The target scene type determination module is configured to determine a target scene type of the target cleaning device by performing clustering processing on the space-time feature data using a preset clustering algorithm. The target cleaning strategy determination module is configured to determine a target cleaning strategy of the target cleaning device according to the target scene type.
11. An electronic device, comprising: A computer program is stored on the memory and executable on the processor, and when the computer program is executed by the processor, the cleaning strategy generation method in any one of claims 1 to 9 is implemented.
12. A computer-readable storage medium, characterized in that, A computer program is stored on the computer readable storage medium and executable on the processor, and when the computer program is executed by the processor, the cleaning strategy generation method in any one of claims 1 to 9 is implemented.
13. A robot vacuum cleaner characterised in that The sweeping machine executes the cleaning strategy generation method in any one of claims 1 to 9.