Intelligent closestool water outlet intelligent control system and method

Abstract

The invention relates to the technical field of smart homes, and particularly discloses an intelligent water outlet control system and method for an intelligent closestool, and the system carries out the time sequence analysis of pressure data collected by a pressure sensor disposed on a closestool ring through the artificial intelligence technology based on deep learning, so as to capture the weight change of a user when the user uses the closestool. Meanwhile, by combining a user defecation sound signal collected by a sound sensor, the defecation sound features of the user are extracted, and then the defecation behavior mode of the user is intelligently recognized by fusing multi-modal data, so that the corresponding flushing water amount and flushing water pressure are executed based on the defecation behavior mode of the user. Thus, the comfort and satisfaction of the user can be remarkably improved, reasonable utilization of water resources can be greatly promoted, and green transformation of home life is achieved.

Description

Technical Field

[0001] The present application relates to the field of smart home technology, and more specifically, to an intelligent toilet water outlet intelligent control system and method. Background Art

[0002] In today's era of rapid technological development, the rapid development of science and technology has profoundly changed people's lives and improved people's spiritual and material needs. People are no longer satisfied with houses that can only meet basic living needs, but are beginning to pursue a safer, more comfortable and convenient living environment. With the continuous improvement of building intelligence and technical level, smart homes have been gradually promoted in the market. Smart home systems can provide people with a more convenient, efficient and comfortable life experience, making family life more intelligent and networked. People can control various devices in their homes, such as lighting, air conditioning, audio, etc. through smart home systems, realize remote control, and make life more convenient. In smart home systems, the application of smart toilets has gradually expanded to individual families. Smart toilets can not only provide basic use functions, but also can be intelligently adjusted according to user needs, such as temperature, seat temperature, water pressure, etc., to provide users with a more comfortable use experience. In addition, smart toilets also have functions such as automatic cleaning and drying, which can provide users with a more hygienic and convenient use environment.

[0003] Smart toilets have both environmental and medical properties and play an important role in home life. However, traditional smart toilets usually use fixed flushing modes, using infrared sensors to detect the user's departure, thereby triggering a fixed flushing mode without considering the user's actual usage. This may lead to a waste of water resources during the flushing process, which is not in line with the energy-saving and environmental protection concept of modern society. Therefore, a smart toilet water outlet intelligent control method is expected. Summary of the invention

[0004] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides an intelligent toilet water outlet intelligent control system and method, which uses artificial intelligence technology based on deep learning to perform time series analysis on the pressure data collected by the pressure sensor deployed on the toilet seat to capture the weight change when the user uses the toilet, and at the same time combines the user's toilet sound signal collected by the sound sensor to extract the user's toilet sound characteristics, and then through the fusion of multimodal data, intelligently identify the user's toilet behavior pattern, so as to execute the corresponding flushing water volume and flushing water pressure based on the user's toilet behavior pattern. In this way, it can not only significantly improve the user's comfort and satisfaction, but also greatly promote the rational use of water resources and realize the green transformation of home life.

[0005] Accordingly, according to one aspect of the present application, there is provided an intelligent toilet water outlet intelligent control system, which includes:

[0006] An information acquisition module, used to acquire a time series of pressure values ​​collected by a pressure sensor deployed on the toilet seat, and a sound signal of a user using the toilet collected by a sound sensor deployed on the toilet;

[0007] A pressure feature extraction module, used for performing time series analysis on the time series of the pressure values ​​to obtain a pressure feature vector;

[0008] A sound feature extraction module, used to extract the sound features of the sound signal of the user going to the toilet to obtain a sound feature vector of the user going to the toilet;

[0009] A toilet behavior analysis module, used to determine the toilet behavior pattern of the user based on the multimodal fusion features of the pressure feature vector and the user toilet sound feature vector;

[0010] The water output control module is used to execute corresponding water output control instructions based on the toilet behavior pattern of the user, wherein the water output control instructions include flushing water volume and flushing water pressure.

[0011] In the above-mentioned intelligent toilet water outlet intelligent control system, the pressure feature extraction module is used to: pass the time series of the pressure values ​​through a sequence encoder based on a bidirectional gated recurrent unit to obtain the pressure feature vector.

[0012] In the above-mentioned intelligent control system for water discharge of smart toilets, the sound feature extraction module includes: a Mel-spectrogram calculation unit, which is used to calculate the Mel-spectrogram of the toilet sound signal of the user's toilet sound signal; a sound feature encoding unit, which is used to pass the Mel-spectrogram of the toilet sound signal through a sound feature extractor based on a convolutional neural network model to obtain the user's toilet sound feature vector.

[0013] In the above-mentioned intelligent toilet water outlet intelligent control system, the sound feature extractor based on the convolutional neural network model includes an input layer, a convolution layer, an activation layer based on the ReLU function, a pooling layer, a batch normalization layer and an output layer.

[0014] In the above-mentioned intelligent toilet water discharge intelligent control system, the toilet behavior analysis module includes: a multimodal feature fusion unit, which is used to fuse the pressure feature vector and the user toilet sound feature vector to obtain the user toilet behavior feature vector; a feature distribution optimization unit, which is used to optimize the feature distribution of the user toilet behavior feature vector to obtain an optimized user toilet behavior feature vector; a behavior pattern recognition unit, which is used to pass the optimized user toilet behavior feature vector through a classifier-based behavior pattern recognizer to obtain a recognition result, and the recognition result is used to represent the user's toilet behavior pattern label.

[0015] In the above-mentioned intelligent toilet water discharge intelligent control system, the multimodal feature fusion unit is used to: perform a cascade operation on the pressure feature vector and the user toilet sound feature vector to obtain the user toilet behavior feature vector.

[0016] In the above-mentioned intelligent toilet water discharge intelligent control system, the feature distribution optimization unit is used to: input the user's toilet behavior feature vector into the viewpoint metric correlation network of backward machine learning to obtain the optimized user toilet behavior feature vector.

[0017] In the above-mentioned intelligent toilet water discharge intelligent control system, the feature distribution optimization unit includes: a probabilistic activation subunit, which is used to perform probabilistic activation on the user toilet behavior feature vector based on the Sigmoid function to obtain a probabilistic user toilet behavior feature vector; a pre-classification subunit, which is used to input the user toilet behavior feature vector into a classifier to obtain an initial class label probability value; an error measurement subunit, which is used to calculate the absolute value of the difference between the feature value of each position in the probabilistic user toilet behavior feature vector and the initial class label probability value to obtain a backward machine learning fine-grained error measurement vector; an attention optimization subunit, which is used to use the backward machine learning fine-grained error measurement vector as a key vector, the user toilet behavior feature vector as a query vector and a value vector, and The key vector, the query vector and the value vector are input into the converter structure to obtain the backward learning type attention optimized user toilet behavior feature vector; the optimization scaling factor calculation subunit is used to calculate the ratio between the sum of the backward machine learning fine-grained error metric vector and the length of the backward machine learning fine-grained error metric vector, and calculate the natural exponential function value with the ratio as the power as the optimization scaling factor; the feature optimization scaling subunit is used to calculate the product between the optimization scaling factor and the backward learning type attention optimized user toilet behavior feature vector to obtain the optimized user toilet behavior feature vector.

[0018] According to another aspect of the present application, a method for intelligently controlling water outlet of an intelligent toilet is provided, which comprises:

[0019] Obtain a time series of pressure values ​​collected by a pressure sensor deployed on a toilet seat, and a sound signal of a user using the toilet collected by a sound sensor deployed on the toilet;

[0020] Performing time series analysis on the time series of the pressure values ​​to obtain a pressure feature vector;

[0021] Extracting the sound features of the user toileting sound signal to obtain a user toileting sound feature vector;

[0022] Determining the user's toilet behavior pattern based on the multimodal fusion features of the pressure feature vector and the user's toilet sound feature vector;

[0023] Based on the toilet behavior pattern of the user, a corresponding water output control instruction is executed, wherein the water output control instruction includes flushing water volume and flushing water pressure.

[0024] Compared with the prior art, the intelligent toilet water discharge intelligent control system and method provided by the present application utilizes artificial intelligence technology based on deep learning to perform time series analysis on the pressure data collected by the pressure sensor deployed on the toilet seat to capture the weight change when the user uses the toilet, and at the same time, combines the user's toilet sound signal collected by the sound sensor to extract the user's toilet sound characteristics, and then through the fusion of multimodal data, intelligently identify the user's toilet behavior pattern, thereby executing the corresponding flushing water volume and flushing water pressure based on the user's toilet behavior pattern. In this way, it can not only significantly improve the user's comfort and satisfaction, but also greatly promote the rational use of water resources and realize the green transformation of home life. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0026] Figure 1 This is a block diagram of an intelligent toilet water outlet intelligent control system according to an embodiment of the present application.

[0027] Figure 2 This is a schematic diagram of the architecture of the smart toilet water outlet intelligent control system according to an embodiment of the present application.

[0028] Figure 3 This is a block diagram of a sound feature extraction module in an intelligent toilet water outlet intelligent control system according to an embodiment of the present application.

[0029] Figure 4 This is a block diagram of the toilet behavior analysis module in the smart toilet water outlet intelligent control system according to the embodiment of the present application.

[0030] Figure 5 This is a block diagram of a feature distribution optimization unit in an intelligent toilet water outlet intelligent control system according to an embodiment of the present application.

[0031] Figure 6 This is a flow chart of the intelligent toilet water outlet intelligent control method according to the embodiment of the present application. DETAILED DESCRIPTION

[0032] Below, the embodiments of the present application will be described in more detail in conjunction with the accompanying drawings, and the above and other purposes, features and advantages of the present application will become more apparent. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein.

[0033] Figure 1 This is a block diagram of an intelligent toilet water outlet intelligent control system according to an embodiment of the present application. Figure 2 FIG. 1 is a schematic diagram of the structure of the intelligent toilet water outlet intelligent control system according to the embodiment of the present application. Figure 1 and Figure 2 As shown, according to the embodiment of the present application, the intelligent control system 100 for intelligent toilet water discharge includes: an information acquisition module 110, which is used to obtain the time series of pressure values ​​collected by the pressure sensor deployed on the toilet seat, and the user's toilet sound signal collected by the sound sensor deployed on the toilet; a pressure feature extraction module 120, which is used to perform time series analysis on the time series of the pressure values ​​to obtain a pressure feature vector; a sound feature extraction module 130, which is used to extract the sound features of the user's toilet sound signal to obtain the user's toilet sound feature vector; a toilet behavior analysis module 140, which is used to determine the user's toilet behavior pattern based on the multimodal fusion features of the pressure feature vector and the user's toilet sound feature vector; a water discharge control module 150, which is used to execute corresponding water discharge control instructions based on the user's toilet behavior pattern, wherein the water discharge control instructions include flushing water volume and flushing water pressure.

[0034] As mentioned in the background technology above, traditional smart toilets usually adopt a fixed flushing mode, and use infrared sensors to detect the user's departure, thereby triggering a fixed flushing mode without considering the user's actual usage. This may lead to a waste of water resources during the flushing process, which is not in line with the energy-saving and environmental protection concept of modern society. In response to the above technical problems, the technical concept of this application is to use artificial intelligence technology based on deep learning to perform time series analysis on the pressure data collected by the pressure sensor deployed on the toilet seat to capture the weight change when the user uses the toilet, and at the same time combine the user's toilet sound signal collected by the sound sensor to extract the user's toilet sound characteristics, and then through the fusion of multimodal data, intelligently identify the user's toilet behavior pattern, so as to execute the corresponding flushing water volume and flushing water pressure based on the user's toilet behavior pattern. In this way, it can not only significantly improve the user's comfort and satisfaction, but also greatly promote the rational use of water resources and realize the green transformation of home life.

[0035] In the above-mentioned intelligent toilet water discharge intelligent control system 100, the information acquisition module 110 is used to obtain the time series of pressure values ​​collected by the pressure sensor deployed on the toilet seat, and the user's toilet sound signal collected by the sound sensor deployed on the toilet. It should be understood that the pressure sensor can capture the pressure changes generated when the user sits on the toilet seat in real time, which can not only effectively confirm the user's presence, but also reveal the user's weight fluctuations and sitting posture changes through in-depth data time series analysis, so as to accurately distinguish a variety of toilet behavior patterns, including but not limited to urination, defecation and long-term sitting. At the same time, the sound sensor deployed around the toilet or in an appropriate position is responsible for capturing various sounds generated by the user during the toilet, such as urine sound, exhaust sound, etc., as another important basis for identifying the user's toilet behavior pattern. In this way, by combining pressure data and sound data, the accuracy and reliability of the user's toilet behavior recognition can be significantly improved, so as to more reasonably adjust the water control strategy. For example, for urination behavior, the flushing water volume and water pressure can be reduced to save water; while for defecation behavior, the flushing water volume and water pressure can be appropriately increased to ensure the cleaning effect.

[0036] In the above-mentioned intelligent toilet water discharge intelligent control system 100, the pressure feature extraction module 120 is used to perform a time series analysis on the time series of the pressure value to obtain a pressure feature vector. In a specific example of the present application, the encoding method for performing a time series analysis on the time series of the pressure value is to pass the time series of the pressure value through a sequence encoder based on a bidirectional gated recurrent unit to obtain the pressure feature vector. It should be understood that, considering that the user's behavior on the toilet is a dynamic and continuous process, the pressure data generated by it also has temporal continuity and correlation. In other words, the time series of the pressure value contains complete information about the pressure change over time when the user sits on the toilet. Therefore, in order to fully explore the potential user behavior information in the time series of the pressure value, in the technical solution of the present application, a bidirectional gated recurrent unit (Bi-directional Gated Recurrent Unit, Bi-GRU) is used as a sequence encoder to capture the temporal dependency and contextual information of the pressure data. Compared with the traditional unidirectional RNN network, the Bi-GRU model can more fully extract the contextual dependencies of pressure data through the gating mechanism and bidirectional information propagation capabilities, and capture the complex temporal changes in the user's toilet behavior, thereby showing higher accuracy and robustness in the recognition of the user's toilet behavior pattern. During the encoding process, Bi-GRU generates a hidden state that integrates the contextual information by performing forward and backward bidirectional processing on the input of each time step, and then splices the hidden states of each time step into the pressure feature vector, thereby providing a rich and effective feature representation for subsequent toilet behavior analysis.

[0037] In the above-mentioned intelligent toilet water discharge intelligent control system 100, the sound feature extraction module 130 is used to extract the sound features of the user toilet sound signal to obtain the user toilet sound feature vector. Figure 3 FIG. 1 is a block diagram of a sound feature extraction module in an intelligent toilet water outlet intelligent control system according to an embodiment of the present application. Figure 3 As shown, the sound feature extraction module 130 includes: a mel-spectrogram calculation unit 131, used to calculate the mel-spectrogram of the toilet sound signal of the user; a sound feature encoding unit 132, used to pass the mel-spectrogram of the toilet sound signal through a sound feature extractor based on a convolutional neural network model to obtain the user toilet sound feature vector.

[0038] Specifically, the Mel-spectrogram calculation unit 131 is used to calculate the Mel-spectrogram of the toilet sound signal of the user's toilet sound signal. It should be understood that Mel-spectrogram (Mel-Spectrogram) is a feature representation method widely used in audio processing, which can effectively capture the frequency information in the audio signal and simulate the perception characteristics of the human ear to the sound. In the processing of the user's toilet sound signal, the collected original sound signal is first pre-processed, including pre-emphasis, framing and windowing operations to divide the continuous audio signal into a series of short-term stable signal frames. Subsequently, a fast Fourier transform (FFT) is performed on each signal frame to convert the time domain signal to the frequency domain to obtain the spectrum representation of each frame. Then, the spectrum representation of each frame is filtered separately using a Mel filter group, and the spectrum energy after passing through the Mel filter is logarithmically operated to convert the linear spectrum into a nonlinear Mel scale. Finally, the energy distribution on each Mel band is calculated to form a Mel-spectrogram of the toilet sound signal. This process not only retains the key frequency components in the sound signal, but also provides strong support for subsequent sound feature encoding by simulating the auditory characteristics of the human ear.

[0039] Specifically, the sound feature encoding unit 132 is used to pass the mel-spectrogram of the toilet sound signal through a sound feature extractor based on a convolutional neural network model to obtain the user's toilet sound feature vector. That is, in order to further extract and characterize the complex features in the user's toilet sound signal, in the technical solution of the present application, a convolutional neural network (CNN) model is used to perform deep sound feature extraction on the mel-spectrogram of the toilet sound signal. Those of ordinary skill in the art should know that the CNN model uses convolutional layers, pooling layers, and nonlinear activation functions, and can automatically learn and extract high-level sound features from the mel-spectrogram through the mechanism of local connection and weight sharing. In a specific example of the present application, the sound feature extractor based on the convolutional neural network model includes an input layer, a convolutional layer, an activation layer based on a ReLU function, a pooling layer, a batch normalization layer, and an output layer. Among them, the convolution layer is responsible for capturing local temporal dependencies and extracting local features from the audio signal through the operation of sliding windows and convolution kernels; the pooling layer is used to reduce the dimension of the features, reduce the amount of calculation, and enhance the robustness of the model to small changes in the input data; the activation layer introduces ReLU (Rectified Linear Unit, linear rectification function) to increase the nonlinear ability of the model; the batch normalization layer stabilizes the internal covariate offset of the convolutional neural network during the entire training process, accelerates the training process, and improves the generalization ability of the model. Through this network architecture, the dynamic changes of the sound signal during the user's toileting process can be fully extracted and represented, providing a more accurate and rich information basis for subsequent toileting behavior recognition.

[0040] In the above-mentioned intelligent toilet water discharge intelligent control system 100, the toilet behavior analysis module 140 is used to determine the user's toilet behavior pattern based on the multimodal fusion characteristics of the pressure feature vector and the user's toilet sound feature vector. Figure 4 FIG. 1 is a block diagram of a toilet behavior analysis module in an intelligent toilet water outlet intelligent control system according to an embodiment of the present application. Figure 4 As shown, the toilet behavior analysis module 140 includes: a multimodal feature fusion unit 141, used to fuse the pressure feature vector and the user toilet sound feature vector to obtain the user toilet behavior feature vector; a feature distribution optimization unit 142, used to optimize the feature distribution of the user toilet behavior feature vector to obtain an optimized user toilet behavior feature vector; a behavior pattern recognition unit 143, used to pass the optimized user toilet behavior feature vector through a classifier-based behavior pattern recognizer to obtain a recognition result, and the recognition result is used to represent the user's toilet behavior pattern label.

[0041] Specifically, the multimodal feature fusion unit 141 is used to fuse the pressure feature vector and the user toilet sound feature vector to obtain the user toilet behavior feature vector. It should be understood that the pressure feature vector and the user toilet sound feature vector respectively represent the information of different dimensions generated by the user in the toilet process. The former captures the pressure change of the user's body in contact with the toilet through the pressure sensor, and the latter records the sound characteristics of the user when using the toilet through the sound sensor. By combining the physical changes captured by the pressure sensor with the acoustic features captured by the sound sensor, the complementarity of the two in time and space can be used to more comprehensively characterize the user's toilet behavior. In a specific example of the present application, the processing method of fusing the pressure feature vector and the user toilet sound feature vector is to perform a cascade operation on the pressure feature vector and the user toilet sound feature vector to obtain the user toilet behavior feature vector. Here, the cascade operation is an intuitive and effective feature fusion method. By directly splicing the pressure feature vector and the user toilet sound feature vector into a longer vector, all the information of the original data can be retained, and a rich multimodal feature set can be provided for subsequent user toilet behavior pattern recognition.

[0042] Specifically, the feature distribution optimization unit 142 is used to optimize the feature distribution of the user's toilet behavior feature vector to obtain an optimized user toilet behavior feature vector. In the technical solution of the present application, it is considered that the data collected by the pressure sensor and the sound sensor may be affected by various factors, such as the accuracy of the sensor, the installation location, the environmental noise, etc. These factors may cause the collected data to deviate from the actual behavior. At the same time, the user pressure feature extractor based on Bi-GRU and the sound feature extractor based on CNN may not be able to fully capture all the features related to toilet behavior. For example, certain specific pressure changes or sound patterns may not be learned by the model. If the data set used to train the model is biased, such as certain types of toilet behavior are not common in the data set, then the model may not accurately identify these behaviors, resulting in a class harmony shift, so that the model may perform well on the training data, but when faced with new and unseen toilet behaviors, its generalization ability may be insufficient, resulting in a shift in the recognition of these behaviors. In other words, since there may be significant differences in the toilet behaviors of different users, the model may find it difficult to adapt to the individual characteristics of all users, resulting in a class harmony shift. Therefore, in order to reduce the category harmony deviation of the user's toilet behavior feature vector, in a specific example of the present application, the user's toilet behavior feature vector is input into a viewpoint metric correlation network of backward machine learning to obtain the optimized user's toilet behavior feature vector.

[0043] Figure 5FIG. 1 is a block diagram of a characteristic distribution optimization unit in an intelligent toilet water outlet intelligent control system according to an embodiment of the present application. Figure 5 As shown, the feature distribution optimization unit 142 includes: a probabilistic activation subunit 1421, which is used to perform probabilistic activation on the user toilet behavior feature vector based on the Sigmoid function to obtain a probabilistic user toilet behavior feature vector; a pre-classification subunit 1422, which is used to input the user toilet behavior feature vector into a classifier to obtain an initial class label probability value; an error measurement subunit 1423, which is used to calculate the absolute value of the difference between the feature value of each position in the probabilistic user toilet behavior feature vector and the initial class label probability value to obtain a backward machine learning fine-grained error measurement vector; an attention optimization subunit 1424, which is used to use the backward machine learning fine-grained error measurement vector as a key vector, the user toilet behavior feature vector as a query vector and a value vector, The key vector, the query vector and the value vector are input into the converter structure to obtain the backward learning type attention optimized user toilet behavior feature vector; the optimization scaling factor calculation subunit 1425 is used to calculate the ratio between the sum of the backward machine learning fine-grained error metric vector and the length of the backward machine learning fine-grained error metric vector, and calculate the natural exponential function value with the ratio as the power as the optimization scaling factor; the feature optimization scaling subunit 1426 is used to calculate the product between the optimization scaling factor and the backward learning type attention optimized user toilet behavior feature vector to obtain the optimized user toilet behavior feature vector.

[0044] That is, the feature distribution optimization unit 142 is used to process the user toilet behavior feature vector using the following feature distribution optimization formula to obtain the optimized user toilet behavior feature vector, wherein the feature distribution optimization formula is:

[0045]

[0046] Among them, V c represents the probabilistic user toileting behavior feature vector after the user toileting behavior feature vector is probabilistically activated, P represents the initial class label probability value, Indicates subtraction by position, V e represents the fine-grained error metric vector for backward machine learning, v i represents the eigenvalue of the ith position of the backward machine learning fine-grained error metric vector, L represents the length of the backward machine learning fine-grained error metric vector, (·) T represents the transpose of a vector, represents vector multiplication, Softmax represents the normalized exponential function, ⊙ represents point multiplication by position, V c ' represents the optimized user toilet behavior feature vector.

[0047] In the technical solution of the present application, the user's toilet behavior feature vector is input into the viewpoint metric correlation network of backward machine learning, which first implements the probabilistic activation based on the Sigmoid function on the user's toilet behavior feature vector to obtain a probabilistic user's toilet behavior feature vector, and inputs the user's toilet behavior feature vector into the classifier to obtain the initial class label probability value. Then, the error between the probabilistic user's toilet behavior feature vector and the initial class label probability value is calculated through backward error analysis to generate a fine-grained error metric vector, which quantifies the viewpoint metric contribution of each feature position to the classification error.

[0048] Furthermore, the backward machine learning fine-grained error metric vector is used as a key vector, the user toilet behavior feature vector is used as a query vector and a value vector, and the key vector, the query vector and the value vector are input into a transformer structure to perform error metric-based attention-based backward machine learning on the user toilet behavior feature vector through a Transformer-like architecture. In this way, the user toilet behavior feature vector is subjected to a backward interpretability analysis so that the obtained backward learning-like attention-optimized user toilet behavior feature vector has a backward learning sensitive preference relative to the predetermined class label of the classifier. The ratio of the sum of the error metric vector and the vector length is further calculated, and the optimized scaling factor obtained by calculating the natural exponential function value with this ratio as an exponent is used to perform backward scaling optimization (BackwardOptimization) on the backward learning-like attention-optimized user toilet behavior feature vector. In this way, the sensitive preference and robustness of the user toilet behavior feature vector relative to the predetermined class label of the classifier are improved, thereby improving the accuracy of the classification result.

[0049] Specifically, the behavior pattern recognition unit 143 is used to pass the optimized user toilet behavior feature vector through a classifier-based behavior pattern recognizer to obtain a recognition result, and the recognition result is used to represent the user's toilet behavior pattern label. That is, a classification algorithm is used to perform pattern recognition on the optimized user toilet behavior feature vector to obtain the user's toilet behavior pattern label. During the recognition process, the optimized user toilet behavior feature vector is fed into the behavior pattern recognizer, which contains the weights and parameters of the pre-trained classification model, and can capture the subtle differences and pattern characteristics of the user's toilet behavior. In other words, the behavior pattern recognizer can perform feature analysis on the optimized user toilet behavior feature vector based on the classification logic and feature mapping rules learned during the training process through the internal network weight parameters and map it to the predefined user toilet behavior pattern space, thereby outputting an accurate recognition result, that is, the user's toilet behavior pattern label.

[0050] In the above-mentioned intelligent toilet water discharge intelligent control system 100, the water discharge control module 150 is used to execute the corresponding water discharge control instruction based on the user's toilet behavior pattern, wherein the water discharge control instruction includes flushing water volume and flushing water pressure. That is, when the classifier-based behavior pattern recognizer outputs the recognition result, the recognition result is quickly transmitted to the control system of the intelligent toilet to guide the water discharge control of the toilet. Specifically, first, different flushing parameter configurations are set for different user toilet behavior patterns. Then, when the control system receives the user's toilet behavior pattern label, it immediately searches for the flushing water volume and flushing water pressure that match the current recognition result according to the preset flushing parameter configuration table, thereby providing the user with the best flushing experience, while also taking into account the environmental protection concept of water saving and energy saving. For example, for users identified as "light use" mode, a low water volume and low water pressure flushing mode is started, and only a small amount of water resources is used for quick and effective cleaning, which not only meets the basic needs of users, but also effectively saves water resources. For users in the "normal use" mode, a flushing solution with medium water volume and water pressure will be selected. For users in "heavy use" or "special needs" mode, such as constipation, diarrhea and other special situations, it will automatically adjust to high water volume, high water pressure flushing mode, by enhancing the flushing force and extending the flushing time to ensure thorough removal of dirt while providing a comfortable flushing experience.

[0051] In summary, the intelligent control system for the water discharge of the smart toilet according to the embodiment of the present application is explained, which uses artificial intelligence technology based on deep learning to perform time series analysis on the pressure data collected by the pressure sensor deployed on the toilet seat to capture the weight change when the user uses the toilet, and at the same time combines the sound signal of the user using the toilet collected by the sound sensor to extract the sound characteristics of the user using the toilet, and then through the fusion of multimodal data, intelligently identify the user's toilet behavior pattern, so as to execute the corresponding flushing water volume and flushing water pressure based on the user's toilet behavior pattern. In this way, it can not only significantly improve the user's comfort and satisfaction, but also greatly promote the rational use of water resources and realize the green transformation of home life.

[0052] Figure 6 FIG. 1 is a flow chart of a method for intelligently controlling water output of an intelligent toilet according to an embodiment of the present application. Figure 6 As shown, the intelligent control method for water discharge of an intelligent toilet according to an embodiment of the present application includes the following steps: S110, obtaining a time series of pressure values ​​collected by a pressure sensor deployed on a toilet seat, and a user's toilet sound signal collected by a sound sensor deployed on the toilet; S120, performing a timing analysis on the time series of the pressure values ​​to obtain a pressure feature vector; S130, extracting the sound features of the user's toilet sound signal to obtain a user's toilet sound feature vector; S140, determining the user's toilet behavior pattern based on the multimodal fusion features of the pressure feature vector and the user's toilet sound feature vector; S150, executing corresponding water discharge control instructions based on the user's toilet behavior pattern, wherein the water discharge control instructions include flushing water volume and flushing water pressure.

[0053] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned intelligent toilet water discharge intelligent control method have been referred to above. Figures 1 to 5 The intelligent toilet water outlet intelligent control system is described in detail, and therefore, its repeated description will be omitted.

[0054] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0055] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiment described above is only schematic. For example, the module division is only a logical function division, and there may be other division methods in actual implementation. The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0056] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0057] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0058] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0059] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0060] Finally, it should be noted that the above description has been given for the purpose of illustration and description. In addition, the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. An intelligent toilet water outlet intelligent control system, characterized in that: include: An information acquisition module, used to acquire a time series of pressure values ​​collected by a pressure sensor deployed on the toilet seat, and a sound signal of a user using the toilet collected by a sound sensor deployed on the toilet; A pressure feature extraction module, used for performing time series analysis on the time series of the pressure values ​​to obtain a pressure feature vector; A sound feature extraction module, used to extract the sound features of the sound signal of the user going to the toilet to obtain a sound feature vector of the user going to the toilet; A toilet behavior analysis module, used to determine the toilet behavior pattern of the user based on the multimodal fusion features of the pressure feature vector and the user toilet sound feature vector; The water output control module is used to execute corresponding water output control instructions based on the toilet behavior pattern of the user, wherein the water output control instructions include flushing water volume and flushing water pressure.

2. The intelligent toilet water outlet intelligent control system according to claim 1, characterized in that: The pressure feature extraction module is used to: The time series of the pressure values ​​is passed through a sequence encoder based on a bidirectional gated recurrent unit to obtain the pressure feature vector.

3. The intelligent toilet water outlet intelligent control system according to claim 2, characterized in that: The sound feature extraction module comprises: A mel-spectrogram calculation unit, used to calculate a mel-spectrogram of the toilet sound signal of the user's toilet sound signal; The sound feature encoding unit is used to obtain the toilet sound feature vector of the user using the toilet by passing the mel-spectrogram of the toilet sound signal through a sound feature extractor based on a convolutional neural network model.

4. The intelligent toilet water outlet intelligent control system according to claim 3, characterized in that: The sound feature extractor based on the convolutional neural network model includes an input layer, a convolution layer, an activation layer based on a ReLU function, a pooling layer, a batch normalization layer and an output layer.

5. The intelligent toilet water outlet intelligent control system according to claim 4, characterized in that: The toilet behavior analysis module comprises: A multimodal feature fusion unit, used for fusing the pressure feature vector and the user toileting sound feature vector to obtain a user toileting behavior feature vector; A feature distribution optimization unit, used for performing feature distribution optimization on the user toilet behavior feature vector to obtain an optimized user toilet behavior feature vector; The behavior pattern recognition unit is used to pass the optimized user toilet behavior feature vector through a classifier-based behavior pattern recognizer to obtain a recognition result, and the recognition result is used to represent the user's toilet behavior pattern label.

6. The intelligent toilet water outlet intelligent control system according to claim 5, characterized in that: The multimodal feature fusion unit is used to: The pressure feature vector and the user toileting sound feature vector are cascaded to obtain the user toileting behavior feature vector.

7. The intelligent toilet water discharge intelligent control system according to claim 5, characterized in that: The feature distribution optimization unit is used to: The user's toilet behavior feature vector is input into a viewpoint metric correlation network for backward machine learning to obtain the optimized user's toilet behavior feature vector.

8. The intelligent toilet water outlet intelligent control system according to claim 7, characterized in that: The feature distribution optimization unit comprises: A probabilistic activation subunit, used for performing probabilistic activation on the user's toilet behavior feature vector based on a Sign function to obtain a probabilistic user toilet behavior feature vector; A pre-classification subunit, used for inputting the user's toilet behavior feature vector into a classifier to obtain an initial class label probability value; An error metric subunit, used to calculate the absolute value of the difference between the feature value of each position in the probabilistic user toilet behavior feature vector and the initial class label probability value to obtain a backward machine learning fine-grained error metric vector; an attention optimization subunit, configured to use the backward machine learning fine-grained error metric vector as a key vector, the user toilet behavior feature vector as a query vector and a value vector, and input the key vector, the query vector and the value vector into a converter structure to obtain a backward learning class attention optimized user toilet behavior feature vector; an optimization scaling factor calculation subunit, configured to calculate a ratio between a sum of the backward machine learning fine-grained error metric vector and a length of the backward machine learning fine-grained error metric vector, and to calculate a natural exponential function value raised to a power of the ratio as an optimization scaling factor; The feature optimization scaling subunit is used to calculate the product between the optimization scaling factor and the backward learning type attention optimized user toilet behavior feature vector to obtain the optimized user toilet behavior feature vector.

9. An intelligent toilet water outlet intelligent control method, characterized in that: include: Obtain a time series of pressure values ​​collected by a pressure sensor deployed on a toilet seat, and a sound signal of a user using the toilet collected by a sound sensor deployed on the toilet; Performing time series analysis on the time series of the pressure values ​​to obtain a pressure feature vector; Extracting the sound features of the user toileting sound signal to obtain a user toileting sound feature vector; Determining the user's toilet behavior pattern based on the multimodal fusion features of the pressure feature vector and the user's toilet sound feature vector; Based on the toilet behavior pattern of the user, a corresponding water output control instruction is executed, wherein the water output control instruction includes flushing water volume and flushing water pressure.

10. The intelligent toilet water outlet intelligent control method according to claim 9, characterized in that: Performing a time series analysis on the time series of the pressure values ​​to obtain a pressure feature vector includes: The time series of the pressure values ​​is passed through a sequence encoder based on a bidirectional gated recurrent unit to obtain the pressure feature vector.

Images