Natural wake-up system based on offline speech recognition controlling acousto-optic time management

Abstract

This invention relates to the field of smart home control technology, and specifically discloses a natural wake-up system based on offline voice recognition control and audio-visual time management. The system includes: an offline voice recognition module that recognizes user voice commands in an offline state; a main control module that calculates a first personalized influence factor of smart home control parameters on the user's wake-up process and a second personalized influence factor of the user's specified wake-up time based on a home control parameter matrix, a wake-up efficiency vector, and a specified wake-up time vector, and then combines AI with user voice commands to generate an audio-visual time management strategy; a time management module that records real-time time; and an RF communication module and an infrared signal transmission module that send control signals to smart home devices capable of receiving RF and infrared control signals, respectively, according to the strategy and real-time time. An audio playback module that plays built-in audio files through a speaker according to the strategy creates a more comfortable and intelligent wake-up environment, improving the user's wake-up experience and system satisfaction.

Description

Technical Field

[0001] This invention relates to the field of smart home control technology, and in particular to a natural wake-up system based on offline voice recognition control and sound and light time management. Background Technology

[0002] In the smart home field, the natural wake-up system, as a key component for enabling natural interaction between people and smart home devices, is of great significance. As people's pursuit of quality of life increases, they desire more convenient and natural control of smart home devices, leading to the emergence of the natural wake-up system. It changes the traditional manual operation mode of smart homes, allowing users to easily control home devices through voice commands, greatly enhancing the user experience. The application of offline voice recognition technology adds new vitality to the natural wake-up system. Even in situations with unstable or no network, offline voice recognition ensures the natural wake-up system continues to function normally, guaranteeing the continuity and stability of user control over smart home devices. Simultaneously, the sound and light time management function further optimizes the usage scenarios of smart homes. By combining time factors to control devices such as lights and sounds, it can better meet users' needs at different times of the day, creating a comfortable and convenient home environment. This natural wake-up system, integrating offline voice recognition and sound and light time management, aligns with the intelligent and human-centered trends in smart home development, possesses broad market prospects, and will propel the smart home industry into a new stage of development.

[0003] However, existing natural wake-up systems have many shortcomings. Voice recognition lacks offline capabilities, failing to accurately recognize commands when the network is poor, thus limiting functionality. They lack comprehensive consideration for user personalization, failing to generate relevant data based on natural wake-up records, making it difficult to obtain personalized influencing factors and formulate precise sound and light timing management strategies. Furthermore, they have deficiencies in interacting with different smart home devices, unable to accurately control devices and create wake-up atmospheres based on strategies and timing through corresponding modules, resulting in low levels of intelligence and personalization, and failing to meet user needs.

[0004] Therefore, this invention proposes a natural wake-up system based on offline speech recognition control and audio-visual time management. Summary of the Invention

[0005] This invention provides a natural wake-up system based on offline speech recognition control and audio-visual time management, comprising:

[0006] First, the offline voice recognition module enables the recognition of user voice commands even in offline conditions. This means that even without a network connection, users can still easily interact with the system via voice, unrestricted by network conditions. This greatly improves the flexibility and stability of the system, ensuring users can issue commands at any time. Second, the main control module generates multiple key vectors and matrices based on the user's natural wake-up records and obtains personalized influencing factors. Combined with artificial intelligence and voice commands, it generates an audio-visual time management strategy, fully considering user habits and personalized needs. This allows for customized wake-up and smart home control solutions for different users, enhancing user experience while increasing the system's intelligence and targeting. Third, the time management module records real-time time, providing a precise time reference for the entire system. This ensures that the audio-visual time management strategy is executed based on accurate time information, guaranteeing timely and accurate control and ensuring that appropriate audio-visual and smart home control operations are performed at the appropriate time. Finally, the RF communication module and infrared signal transmission module send control signals to smart home devices that can receive different signals. This design is compatible with various types of smart home devices, expanding the system's applicability and allowing users to integrate various smart devices in their homes for unified and convenient control. Finally, the audio playback module plays built-in audio files according to the sound and light time management strategy and real-time time, providing users with a personalized wake-up audio experience. Together with other control functions, it creates a more comfortable and intelligent wake-up environment, improving users' wake-up experience and system satisfaction.

[0007] This invention provides a natural wake-up system based on offline speech recognition control and audio-visual time management, comprising:

[0008] The offline speech recognition module is used to recognize the user's voice commands received by the microphone in an offline state;

[0009] The main control module is used to obtain the first personalized influence factor of smart home control parameters on the user's wake-up process and the second personalized influence factor of the user's specified wake-up time on the user's wake-up process based on the home control parameter matrix, wake-up efficiency vector, and specified wake-up time vector generated by the user's natural wake-up records, and to generate an audio-visual time management strategy by combining artificial intelligence and the user's voice commands.

[0010] The time management module is used to record real-time time.

[0011] The RF communication module is used to send control signals to smart home devices that can receive RF signals based on the audio-visual time management strategy and real-time time.

[0012] The infrared signal transmitting module is used to send control signals to smart home devices that can receive infrared control signals, based on the audio-visual time management strategy, real-time time, and smart home remote control codes stored in the infrared remote control code library.

[0013] The audio playback module is used to play built-in audio files through a speaker based on the sound and light time management strategy and real-time time.

[0014] Preferred options also include:

[0015] The rechargeable battery provides power to the microphone, offline voice recognition module, main control module, time management module, RF communication module, infrared remote control code library, infrared signal transmission module, audio playback module, speaker, power management module, and LED display module.

[0016] The power management module is used to issue a reminder signal when the remaining capacity of the rechargeable battery is lower than a threshold, and at the same time, adjust the working mode of the main control module to power saving mode.

[0017] The LED display module is used to display real-time time, remaining battery power, time information in the latest voice command, and the working status of all communicable smart devices.

[0018] The storage module is used to store all received voice commands, the ongoing working status of all communicable smart devices, and all currently determined audio-visual time management strategies.

[0019] Preferred smart home devices that can receive RF signals include LED lights, smart curtains, and aroma diffusers;

[0020] LED lights gradually illuminate and change color temperature based on the received control signals;

[0021] The smart curtains gradually open based on the received control signals;

[0022] The aroma diffuser starts working based on the received control signal.

[0023] Preferably, the smart home device that can receive RF signals is an air conditioner, which gradually adjusts the temperature to the corresponding temperature based on the received control signal.

[0024] Preferably, the main control module includes:

[0025] The data processing submodule is used to generate a home control parameter matrix, a wake-up efficiency vector, and a specified wake-up time vector based on the user's natural wake-up records.

[0026] The correlation analysis submodule is used to generate a first personalized influence factor of smart home control parameters on the user's wake-up process and a second personalized influence factor of the user's specified wake-up time on the user's wake-up process based on the correlation analysis results between the home control parameter matrix and the wake-up efficiency vector and the correlation analysis results between the specified wake-up time vector and the wake-up efficiency vector.

[0027] The audio-visual time management strategy generation submodule is used to obtain the audio-visual time management strategy based on the first personalized influence factor of the user wake-up process, the second personalized influence factor of the user-specified wake-up time, the latest user-specified wake-up time represented by the latest received voice command, and the pre-built audio-visual time management strategy generation model.

[0028] Preferably, the data processing submodule includes:

[0029] The wake-up efficiency value analysis unit is used to analyze the wake-up efficiency value of each natural wake-up process based on the continuous control execution time of each smart home in the sound and light time management strategy executed in each natural wake-up process in the user's natural wake-up record, and the corresponding time required for the user to be fully awakened.

[0030] The home control parameter matrix generation unit is used to generate a home control parameter matrix based on the continuous control parameter threads of all smart homes in the sound and light time management strategy executed in all natural wake-up processes in the natural wake-up record.

[0031] The wake-up efficiency vector generation unit is used to generate a wake-up efficiency vector based on the wake-up efficiency values ​​of all natural wake-up processes in the natural wake-up record.

[0032] The wake-up time vector generation unit is used to generate a specified wake-up time vector based on the user-specified wake-up time of all natural wake-up processes in the natural wake-up record.

[0033] Preferably, the wake-up efficiency value analysis unit includes:

[0034] The single wake-up efficiency value determination subunit is used to take the ratio of the continuous control execution time of each smart home in each natural wake-up process to the time required for the user to be fully woken up in the corresponding natural wake-up process as the single wake-up efficiency value of each smart home in the corresponding natural wake-up process.

[0035] The Single Wake-up Efficiency Value Summarization Subunit is used to sum the single wake-up efficiency values ​​of all smart home devices in a single natural wake-up process based on the wake-up weights of all smart home devices, so as to obtain the wake-up efficiency value for each natural wake-up process.

[0036] Preferably, the correlation analysis submodule includes:

[0037] The matrix dimensionality reduction and differentiation unit is used to reduce the dimensionality of the home control parameter matrix to obtain the home control parameter dimensionality reduction matrix. Based on the dimensionality reduction value sequence of the control parameters of each smart home in the home control parameter dimensionality reduction matrix, the dimensionality reduction function of the control parameters of each smart home is generated. Based on the first derivative values ​​of the dimensionality reduction functions of the control parameters of all smart homes at all times, the first derivative matrix of the home control parameter dimensionality reduction is generated.

[0038] The first influence factor calculation unit is used to analyze at least one common feature vector of the first-order derivative matrix of the home control parameters in the time dimension, and obtain the first influence factor of the smart home control parameters on the user wake-up process based on the cosine similarity between the common feature vector of the first-order derivative matrix of the home control parameters in the time dimension and the feature vector of the wake-up efficiency vector in the time dimension.

[0039] The second influence factor calculation unit is used to obtain the second influence factor of the user's specified wake-up time on the user's wake-up process based on the correlation analysis results between the specified wake-up time vector and the wake-up efficiency vector.

[0040] The wake-up difficulty assessment unit is used to assess the wake-up difficulty of a user based on their natural wake-up history.

[0041] The personalized influence factor calibration unit is used to determine the first personalized influence factor of smart home control parameters on the user wake-up process and the second personalized influence factor of the user-specified wake-up time on the user wake-up process based on the first influence factor of smart home control parameters on the user wake-up process, the second influence factor of the user-specified wake-up time on the user wake-up process, and the user's wake-up difficulty.

[0042] Preferably, the first impact factor calculation unit includes:

[0043] The first eigenvalue decomposition subunit is used to determine the covariance matrix of the first-order derivative matrix of the home control parameters. The eigenvalue decomposition is performed on the covariance matrix of the first-order derivative matrix of the home control parameters to obtain at least one eigenvalue and the corresponding eigenvector. Each eigenvector is regarded as the common eigenvector of the first-order derivative matrix of the home control parameters in the time dimension.

[0044] The vector time dimension extension sub-unit is used to construct the lag feature matrix of the wake-up efficiency vector based on the sliding window method. The mean of the lag feature matrix of the wake-up efficiency vector is normalized to zero and the variance is normalized to obtain the standardized lag feature matrix.

[0045] The second eigenvalue decomposition subunit is used to perform eigenvalue decomposition on the covariance matrix of the standardized lag feature matrix to obtain all eigenvalues ​​and corresponding eigenvectors of the covariance matrix of the standardized lag feature matrix. The eigenvector corresponding to the largest eigenvalue among all eigenvalues ​​and corresponding eigenvectors of the covariance matrix of the standardized lag feature matrix is ​​taken as the eigenvector of the wake-up efficiency vector in the time dimension.

[0046] The cosine similarity calculation subunit is used to calculate the cosine similarity between the feature vectors of each common feature vector with a corresponding feature value greater than the threshold and the wake-up efficiency vector in the time dimension.

[0047] The correlation synthesis subunit is used to calculate the weight value of each common feature vector whose corresponding feature value is greater than the threshold based on all feature values ​​greater than the threshold. Based on the weight values ​​of all common feature vectors whose corresponding feature values ​​are greater than the threshold, the cosine similarity of each common feature vector whose corresponding feature value is greater than the threshold and the wake-up efficiency vector in the time dimension is weighted and summed to obtain the first influencing factor of smart home control parameters on the user wake-up process.

[0048] Preferred, personalized impact factor calibration unit includes:

[0049] The adjustment coefficient determination subunit is used to determine the average wake-up time based on the user's natural wake-up record, determine the first adjustment coefficient based on the average wake-up time and the user's wake-up difficulty, and determine the second adjustment coefficient based on the user's wake-up difficulty and difficulty sensitivity coefficient.

[0050] The first influence factor personalized calibration subunit is used to take the product of the first influence factor of the smart home control parameters on the user wake-up process, the first adjustment coefficient, and the importance weight of the control parameters as the first personalized influence factor of the smart home control parameters on the user wake-up process.

[0051] The second influence factor personalized calibration subunit is used to take the product of the second personalized influence factor, the second adjustment coefficient, and the time window adjustment coefficient of the user-specified wake-up time on the user wake-up process as the second personalized influence factor of the user-specified wake-up time on the user wake-up process.

[0052] The beneficial effects of this invention compared to existing technologies are as follows: First, the offline voice recognition module enables the recognition of user voice commands in offline states. This means that even without a network connection, users can still conveniently interact with the system via voice, unrestricted by network conditions, greatly improving the flexibility and stability of the system and ensuring users can issue commands at any time. Second, the main control module generates multiple key vectors and matrices based on the user's natural wake-up records and obtains personalized influencing factors. Combined with artificial intelligence and voice commands, it generates an audio-visual time management strategy, fully considering user habits and personalized needs. This allows for customized wake-up and smart home control solutions for different users, enhancing user experience while increasing the system's intelligence and targeting. Third, the time management module records real-time time, providing a precise time reference for the entire system. This ensures that the audio-visual time management strategy is executed based on accurate time information, guaranteeing the timeliness and accuracy of control and ensuring that appropriate audio-visual and smart home control operations are performed at the appropriate time. Furthermore, the RF communication module and infrared signal transmission module transmit control signals to smart home devices that can receive different signals. This design is compatible with various types of smart home devices, expanding the system's applicability and facilitating the integration of various smart devices in the home for unified and convenient control. Finally, the audio playback module plays built-in audio files according to the sound and light time management strategy and real-time time, providing users with a personalized wake-up audio experience. Together with other control functions, it creates a more comfortable and intelligent wake-up environment, improving users' wake-up experience and system satisfaction.

[0053] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0054] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0055] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0056] Figure 1 This is a schematic diagram of a natural wake-up system based on offline speech recognition control and sound-light time management in an embodiment of the present invention;

[0057] Figure 2 This is a schematic diagram illustrating the control of various smart home devices based on offline voice recognition in an embodiment of the present invention;

[0058] Figure 3 This is a schematic diagram of the execution logic of the main control module in an embodiment of the present invention. Detailed Implementation

[0059] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0060] Example 1:

[0061] This invention provides a natural wake-up system based on offline speech recognition control and audio-visual time management. The various modules work collaboratively to achieve the natural wake-up function. (See reference...) Figure 1 Offline voice recognition controlled alarm clock, including:

[0062] Offline voice recognition module: This module enables the system to recognize user voice commands collected through the microphone even when offline without a network connection. This feature eliminates reliance on the network, allowing users to easily transmit commands to the system regardless of their network environment. For example, in an environment without Wi-Fi or mobile data, a user can say a command such as "Set wake-up mode for 7 AM tomorrow," and the offline voice recognition module can accurately recognize it, initiating the wake-up setup process and meeting the user's need to control the natural wake-up system in different scenarios.

[0063] The main control module utilizes the user's natural wake-up records to generate a home control parameter matrix, a wake-up efficiency vector, and a specified wake-up time vector. Through analysis and processing of this data, it identifies the first personalized influencing factor of smart home control parameters (such as light brightness and color temperature, sound volume and type) on the user's wake-up process, and the second personalized influencing factor of the user's specified wake-up time. For example, by analyzing multiple natural wake-up records, it was found that users have higher wake-up efficiency under warm light of a specific color temperature and with soft birdsong, thus determining the influencing factors of these parameters on the wake-up process. Then, the main control module combines artificial intelligence technology with the user's voice commands to generate a suitable sound and light time management strategy. This strategy fully considers the user's personalized needs and habits, ensuring the creation of the most suitable natural wake-up environment for the user.

[0064] Time Management Module: This module is responsible for accurately recording real-time time, providing an accurate time reference for the entire natural wake-up system. All system operations, such as executing wake-up tasks according to set times and adjusting the rhythm of environmental elements like lights and sounds based on real-time time, rely on the time information recorded by the time management module. For example, if a user sets 7:00 AM as the wake-up time, the time management module accurately keeps track of the time and triggers a series of wake-up actions at that moment, ensuring the accuracy and timeliness of the wake-up process.

[0065] RF Communication Module: Based on the audio-visual time management strategy generated by the main control module and the real-time time recorded by the time management module, the RF communication module sends control signals to smart home devices capable of receiving RF signals. This enables the system to interact with smart devices that support RF communication, allowing for remote control of them. For example, before the set wake-up time, the RF communication module sends a signal to the smart curtains according to the strategy, controlling them to slowly open, simulating the gradual increase of natural light and creating an atmosphere close to natural wake-up.

[0066] Infrared signal transmitting module: Based on the audio-visual time management strategy, real-time time, and smart home remote control codes stored in the infrared remote control code library, this module sends control signals to smart home devices that can receive infrared control signals. For some traditional smart home devices that rely on infrared signals for control, such as certain models of air conditioners, the infrared signal transmitting module plays a crucial role. For example, during the wake-up process, it sends infrared signals to the air conditioner at appropriate times according to the strategy to adjust the indoor temperature and create a comfortable wake-up environment for the user.

[0067] Audio playback module: Based on the sound and light time management strategy and real-time time, the built-in audio files are played through the speaker. During the natural wake-up process, the audio playback module plays an important role in creating the right atmosphere. For example, as the wake-up time approaches, the volume of natural birdsong or other natural sounds is gradually increased from a low level, simulating the sound changes in the natural environment in the early morning. This helps the user gradually transition from deep sleep to light sleep, ultimately waking up naturally and achieving the goal of a healthy wake-up.

[0068] Example 2:

[0069] Building upon Example 1, this paper further refines the natural wake-up system based on offline speech recognition control and audio-visual time management. Each newly added module plays a crucial role in the overall system. Considering the overall technical solution: (Refer to...) Figure 1 It also includes:

[0070] Rechargeable battery: It acts as the energy hub of the entire natural wake-up system, powering numerous key components such as the microphone, offline voice recognition module, and main control module. This design is closely aligned with the system's offline usage requirements. Even without network access and in the event of an external power outage, the system can still operate normally thanks to the rechargeable battery, continuing its simulated natural wake-up function. For example, in the event of a sudden power outage, the system can still create a natural wake-up environment for the user by controlling devices such as LED lights and smart curtains according to preset programs, ensuring the system's reliability and stability, meeting the user's need to "wake up naturally," and also aligning with the system's characteristics of operating without a network connection and protecting user privacy and security.

[0071] The power management module features dual functions: power monitoring and mode adjustment. When the remaining capacity of the rechargeable battery falls below a threshold, it issues a reminder signal. This function aims to promptly inform the user of the battery status so they can charge the battery in time, preventing the system from failing to perform wake-up tasks due to depleted power. Simultaneously, it switches the main control module to power-saving mode. This operation, when power is limited, reduces the main control module's power consumption to ensure the continuous operation of critical system functions. For example, even when the battery is low and the main control module enters power-saving mode, it can still control various smart devices based on the user's previously set natural wake-up strategy, maintaining the system's basic wake-up function. This maximizes the use of limited power while meeting user needs, demonstrating the system's human-centered and intelligent design.

[0072] The LED display module, serving as a window for information interaction, is closely related to the system's time management and device status monitoring. It displays the real-time time, corresponding to the time recorded by the time management module, providing users with an accurate time reference to help them understand the current moment and plan subsequent activities accordingly. It displays the remaining battery power, allowing users to know the rechargeable battery's reserve and, combined with reminders from the power management module, better plan charging time. It displays the time information from the latest voice commands, helping users confirm key commands such as wake-up time and ensuring accurate natural wake-up settings. Furthermore, it displays the working status of all communicable smart devices, allowing users to intuitively understand whether devices such as LED lights, smart curtains, aromatherapy diffusers, and air conditioners are operating according to preset strategies, such as the brightness and color temperature of LED lights and the opening / closing degree of smart curtains, thereby effectively monitoring and managing the entire natural wake-up environment.

[0073] The storage module bears the crucial responsibility of recording and analyzing system data. It stores all received voice commands, enabling the system to deeply analyze user wake-up needs and operating habits. For example, by analyzing user voice commands in different time periods and scenarios, it can further optimize offline voice recognition, improve recognition accuracy and natural language understanding, allowing users to express their needs more easily and naturally. It stores the continuous operating status of all communicable smart devices, providing data support for the system to study device usage patterns and optimize control strategies. For instance, it can understand the opening speed and travel requirements of smart curtains in different seasons and wake-up times, thus more accurately simulating natural wake-up scenarios. It stores all currently determined sound and light time management strategies, facilitating the system's reuse and adjustment of these strategies in subsequent operations. Based on long-term user habits and environmental changes, it continuously optimizes the natural wake-up mode to achieve better wake-up effects and realize the healthy wake-up goal of "waking up naturally." It also provides data support for continuous system improvement and functional expansion.

[0074] Example 3:

[0075] Based on Example 1, and referring to Figure 2 Smart home devices that can receive RF signals include LED lights, smart curtains, and aroma diffusers;

[0076] Smart home devices capable of receiving RF signals: This category includes LED lighting, smart curtains, and aroma diffusers. These devices are an important component of natural wake-up systems that simulate environmental conditions.

[0077] LED lights: When an LED light receives a control signal from the RF communication module, it will gradually light up according to the settings, and gradually change its color temperature during the lighting process. For example, at the beginning of the wake-up process, the LED light may slowly light up with a lower brightness and a warmer color temperature, simulating the process of natural light in the morning changing from weak to strong and from warm to bright, helping users gradually adapt to the changes in light from deep sleep to light sleep, which meets the needs of natural wake-up systems to simulate changes in natural ambient light.

[0078] Smart curtains: Based on control signals transmitted through an RF communication module, smart curtains gradually open. For example, as the wake-up time approaches, the smart curtains slowly open, allowing more natural light to enter the room, further enhancing the light change effect. Combined with LED lights, this creates a more realistic natural wake-up light environment, making the wake-up process closer to a natural state.

[0079] Aroma diffuser: Once it receives a control signal, the aroma diffuser will start working. In a natural wake-up system, the specific fragrances released by the aroma diffuser help create a comfortable wake-up atmosphere, such as releasing pleasant scents that help people wake up, thus enhancing the natural wake-up experience from an olfactory perspective.

[0080] Example 4:

[0081] Based on Example 1, and referring to Figure 2 In this context, a smart home device capable of receiving RF signals is an air conditioner. The air conditioner gradually adjusts the temperature to the corresponding level based on the received control signal. This section specifically addresses smart home devices capable of receiving RF signals, focusing on air conditioners and their response to control signals. Specifically, it states that a smart home device capable of receiving RF signals is an air conditioner. When the air conditioner receives a control signal from the RF communication module based on an audio-visual time management strategy and real-time timing, it will gradually adjust the temperature to the corresponding level.

[0082] For example, in a natural wake-up scenario, to create a comfortable wake-up environment for the user, the system sends a control signal to the air conditioner via the RF communication module before the wake-up time approaches, based on a preset sound and light time management strategy. The air conditioner then gradually adjusts the temperature. Assuming the set suitable wake-up temperature is 25℃, it will adjust from the current temperature to 25℃ at a relatively gentle pace, avoiding sudden temperature changes that may cause discomfort to the user. This aligns with the natural wake-up system's goal of comprehensively simulating a natural and comfortable environment, allowing the user to smoothly transition from deep sleep to light sleep until naturally waking up.

[0083] Example 5:

[0084] Based on Example 1, and referring to Figure 3 The main control module includes:

[0085] The data processing submodule generates a home control parameter matrix, a wake-up efficiency vector, and a specified wake-up time vector based on the user's natural wake-up records. This submodule operates on the user's natural wake-up records. Natural wake-up records refer to detailed information collected by the system regarding each natural wake-up by the user, such as the brightness and color temperature settings of LED lights, the degree to which smart curtains are opened, the type of scent released by the aroma diffuser, the content and volume of audio playback, and the corresponding wake-up time. The data processing submodule organizes and analyzes these records to generate the home control parameter matrix, the wake-up efficiency vector, and the specified wake-up time vector.

[0086] The home control parameter matrix is ​​a data structure that arranges various home control parameters in a matrix format. It clearly displays the combination of different home control parameters under different wake-up scenarios. For example, the rows of the matrix can represent different wake-up events, and the columns can represent different home control parameters, such as the first column being LED light brightness, the second column being color temperature, and the third column being the opening ratio of smart curtains, etc. In this way, the matrix allows you to intuitively see the values ​​of each home control parameter at each wake-up.

[0087] The wake-up efficiency vector is a data representation used to measure the efficiency of a user's transition from deep sleep to light sleep and finally to natural wake-up during each wake-up process. Simply put, each element in the vector corresponds to a wake-up event, and the value of each element can be a numerical value to quantify wake-up efficiency, such as a range of 0-100. A higher value indicates higher wake-up efficiency, and this value can be derived based on factors such as the user's state after wake-up.

[0088] The specified wake-up time vector records the wake-up time points set by the user each time. Each element in the vector is a specific time value, such as "7:00 AM" or "8:30 AM", which facilitates the system's analysis of the relationship between the user's set wake-up time and wake-up efficiency.

[0089] The correlation analysis submodule is used to generate a first personalized influence factor of smart home control parameters on the user's wake-up process and a second personalized influence factor of the user's specified wake-up time on the user's wake-up process based on the correlation analysis results between the home control parameter matrix and the wake-up efficiency vector, and the correlation analysis results between the specified wake-up time vector and the wake-up efficiency vector. Specifically, it includes:

[0090] Analysis based on the home control parameter matrix and wake-up efficiency vector: The home control parameter matrix contains the combination of various smart home control parameters during each natural wake-up process, such as the brightness and color temperature of LED lights, and the degree to which smart curtains are opened, recording the values ​​of these parameters under different wake-up scenarios. The wake-up efficiency vector quantifies the efficiency with which a user transitions from deep sleep to light sleep and ultimately wakes up naturally during each wake-up process. By analyzing the relationship between these two, such as studying the correlation between specific brightness and color temperature lighting settings and wake-up efficiency, combinations of home control parameters that have a significant impact on wake-up efficiency are identified, thus determining the first personalized influencing factor of smart home control parameters on the user's wake-up process. Based on these correlation analysis results, the generated first personalized influencing factor reflects the unique degree of influence of smart home control parameters on the user's wake-up process. For example, if specific combinations of parameters such as light brightness, color temperature, and the degree to which smart curtains are opened are found to be closely related to higher wake-up efficiency, then these parameters will have a larger weight in the first personalized influencing factor, indicating that their impact on the user's wake-up process is more critical.

[0091] Based on the analysis of the specified wake-up time vector and the wake-up efficiency vector: The specified wake-up time vector records the wake-up time point set by the user each time. Correlation analysis between this vector and the wake-up efficiency vector reveals the impact of different wake-up times on wake-up efficiency. For example, the analysis found that wake-up efficiency is higher when users set their wake-up time between 7:00 and 7:30 AM. Therefore, this time period will have a higher weight when generating the second personalized influencing factor, reflecting the degree to which the user's specified wake-up time affects the user's wake-up process. Through the correlation analysis of these two aspects, the generated first and second personalized influencing factors can accurately reflect the role of smart home control parameters and specified wake-up time in the unique wake-up process of each user.

[0092] The second personalized influencing factor reflects the unique impact of the user-defined wake-up time on the wake-up process. For example, if the analysis shows that users have higher wake-up efficiency when setting their wake-up time between 6:30 and 7:30 in the morning, then this time period will have a greater weight in the second personalized influencing factor, indicating that this time period has a more significant impact on the user's wake-up process.

[0093] The audio-visual time management strategy generation submodule is used to obtain the audio-visual time management strategy based on the first personalized influence factor of the user wake-up process, the second personalized influence factor of the user-specified wake-up time, the latest user-specified wake-up time represented by the latest received voice command, and the pre-built audio-visual time management strategy generation model.

[0094] This submodule takes as important input the first personalized influence factor of the smart home control parameters generated by the relevance analysis submodule on the user wake-up process, and the second personalized influence factor of the user-defined wake-up time on the user wake-up process. These two factors reflect the user's personalized wake-up characteristics. Simultaneously, the latest user-defined wake-up time, represented by the most recently received voice command, is also important information, clarifying the user's desired wake-up time. For example, if the user says, "Set wake-up time for tomorrow morning at 7:15," then "7:15 AM" is the latest user-defined wake-up time.

[0095] Furthermore, the pre-built audio-visual time management strategy generation model is a model constructed based on a large amount of data and algorithms. It can output audio-visual time management strategies suitable for users based on various input parameters. This model is obtained through deep learning of numerous excellent wake-up examples from users (including the first personalized influence factor of smart home control parameters on the user wake-up process, the second personalized influence factor of the user-specified wake-up time on the user wake-up process, the latest user-specified wake-up time represented by the latest received voice command, and the audio-visual time management strategy adopted when the user is woken up under this premise and gives a good wake-up evaluation).

[0096] When the first personalized influencing factor of smart home control parameters on the user wake-up process, the second personalized influencing factor of the user-defined wake-up time, and the latest user-defined wake-up time represented by the latest received voice command are input into a pre-built audio-visual time management strategy generation model, a suitable audio-visual time management strategy can be generated. If the model detects that a user prefers a softer gradual change in light and a gentle flowing water sound to wake up, and sets the wake-up time to be earlier, then it will generate an audio-visual time management strategy that starts at an earlier time and gradually increases the brightness and volume of the lights at a slower pace. For example, at 6:45 am, the LED lights slowly turn on with a lower brightness and warm color tone, while a gentle flowing water sound is played. As the time approaches 7:15 am, the brightness and volume gradually increase to create a comfortable wake-up environment.

[0097] Example 6:

[0098] Based on Example 5, the data processing submodule includes:

[0099] The wake-up efficiency value analysis unit is used to analyze the wake-up efficiency value of each natural wake-up process based on the continuous control execution time of each smart home in the sound and light time management strategy executed in each natural wake-up process in the user's natural wake-up record, and the corresponding time required for the user to be fully awakened.

[0100] The wake-up efficiency analysis unit operates based on the audio-visual time management strategy executed during each natural wake-up process, as recorded in the user's natural wake-up log. This audio-visual time management strategy involves the continuous control execution time of each smart home device. For example, during a natural wake-up process, the duration it takes for an LED light to gradually brighten until it reaches a certain brightness, or the time it takes for a smart curtain to fully open from the start, constitutes the continuous control execution time of the smart home. The corresponding time required for the user to fully wake up refers to the time elapsed from the start of the natural wake-up process until the user is truly fully awake. This time can be obtained by inputting the information into the wake-up device through rotating a knob or other input methods (such as button input) when the user feels fully awake during the wake-up process, and then statistically analyzing the data. This unit derives the wake-up efficiency value for each natural wake-up process by analyzing the relationship between these two times.

[0101] The home control parameter matrix generation unit generates a home control parameter matrix based on the continuous control parameter threads of all smart home devices within the sound and light time management strategy executed during all natural wake-up processes recorded in the natural wake-up log. Each smart home device has a series of related control parameters, such as the brightness and color temperature of LED lights, and the opening speed and travel distance of smart curtains. The sequence formed by the changes of these parameters over time constitutes the continuous control parameter thread. For example, during multiple natural wake-up processes, the sequence of values ​​for the brightness and color temperature of LED lights at different time points, and the sequence of values ​​for the opening speed and travel distance of smart curtains at each stage, are all integrated to form the home control parameter matrix. This matrix comprehensively displays the changes of various smart home control parameters during different natural wake-up processes.

[0102] The wake-up efficiency vector generation unit generates a wake-up efficiency vector based on the wake-up efficiency values ​​of all natural wake-up processes recorded in the natural wake-up log. As the wake-up efficiency value analysis unit calculates the wake-up efficiency value for each natural wake-up process, these values ​​are arranged in the order of the natural wake-up processes to form the wake-up efficiency vector. For example, the wake-up efficiency value for the first natural wake-up process is 70, the second is 80, the third is 75, and so on. This vector of values ​​visually reflects the changes in wake-up efficiency across different natural wake-up processes.

[0103] The wake-up time vector generation unit generates a specified wake-up time vector based on the user-defined wake-up times of all natural wake-up processes recorded in the natural wake-up log. Each user-defined wake-up time, such as "7:00 AM" or "9:30 AM," arranged chronologically according to the natural wake-up events, constitutes the specified wake-up time vector. This vector allows analysis of patterns in user-defined wake-up times and their relationships with other data (such as wake-up efficiency).

[0104] Example 7:

[0105] Based on Example 6, the wake-up efficiency value analysis unit includes:

[0106] The single wake-up efficiency value determination subunit is used to take the ratio of the continuous control execution time of each smart home in each natural wake-up process to the time required for the user to be fully woken up in the corresponding natural wake-up process as the single wake-up efficiency value of each smart home in the corresponding natural wake-up process.

[0107] This sub-unit is responsible for calculating the single wake-up efficiency value of each smart home device in the corresponding natural wake-up process. Specifically, it compares the total time required for the user to fully wake up in each natural wake-up process with the individual continuous control execution time of each smart home device during that process. For example, if in a natural wake-up process, it takes 20 minutes for the user to fully wake up, while the LED lights take 15 minutes to reach their set brightness after initial adjustment, then the single wake-up efficiency value of the LED lights in this natural wake-up process is 15 ÷ 20 = 0.75 (this assumes the ratio is the calculation method for the single wake-up efficiency value). Similarly, for other smart home devices such as smart curtains and aroma diffusers, their single wake-up efficiency values ​​in the same way are calculated for each device in the natural wake-up process. This allows us to understand the individual contribution of each smart home device to the user's wake-up process.

[0108] The Single Wake-up Efficiency Value Summarization Subunit is used to sum the single wake-up efficiency values ​​of all smart home devices in a single natural wake-up process based on the wake-up weights of all smart home devices, so as to obtain the wake-up efficiency value for each natural wake-up process.

[0109] Different smart home devices may have varying degrees of importance in the overall wake-up process, so a wake-up weight is assigned to each smart home device. For example, suppose the wake-up weight of an LED light fixture is 0.4, that of a smart curtain is 0.3, that of an aroma diffuser is 0.2, and the total weight of other devices is 0.1 (these weights are just examples). In a particular natural wake-up process, the individual wake-up efficiency value of the LED light fixture is 0.2, that of the smart curtain is 0.5, that of the aroma diffuser is 0.8, and the combined individual wake-up efficiency value of the other devices is 0.3. Therefore, the total wake-up efficiency value for this natural wake-up process is 0.2 × 0.4 + 0.5 × 0.3 + 0.8 × 0.2 + 0.3 × 0.1 = 0.42. This weighted summation method comprehensively considers the different roles of each smart home device in the wake-up process, resulting in a wake-up efficiency value that more accurately reflects the overall wake-up effect.

[0110] Example 8:

[0111] Based on Example 5, the correlation analysis submodule includes:

[0112] The matrix dimensionality reduction and differentiation unit is used to reduce the dimensionality of the home control parameter matrix to obtain the home control parameter dimensionality reduction matrix. Based on the dimensionality reduction value sequence of the control parameters of each smart home in the home control parameter dimensionality reduction matrix, the dimensionality reduction function of the control parameters of each smart home is generated. Based on the first derivative values ​​of the dimensionality reduction functions of the control parameters of all smart homes at all times, the first derivative matrix of the home control parameter dimensionality reduction is generated.

[0113] Since the original home control parameter matrix may contain data with multiple dimensions, dimensionality reduction is needed to facilitate the analysis of its relationship with the wake-up efficiency vector. This can be achieved by merging control parameter values ​​of the same type at adjacent time points, thus reducing the amount of data in the time dimension of the matrix. This results in a dimensionality-reduced home control parameter matrix. For example, the original matrix may contain multiple parameter dimensions from various smart home devices. After dimensionality reduction, some dimensions with less impact on the analysis are removed, making the data simpler while retaining key information.

[0114] Based on the dimensionality reduction matrix of home control parameters, the dimensionality reduction values ​​of the control parameters of each smart home device at different times will form a sequence, which can be used to generate the dimensionality reduction function of the control parameters of each smart home. For example, the sequence of the brightness dimensionality reduction values ​​of LED lights changing over time can be fitted with a function to describe the changing pattern of the brightness dimensionality reduction values ​​over time.

[0115] Then, the first derivative values ​​of the dimensionality reduction function of the control parameters of all smart home devices are calculated at all times. The first derivative reflects the rate of change of the function, and these first derivative values ​​are used to generate the first derivative matrix of the dimensionality reduction of the home control parameters. This matrix can show the rate of change of the control parameters of each smart home device at different times, providing an important basis for subsequent analysis.

[0116] The first influence factor calculation unit is used to analyze at least one common feature vector of the first-order derivative matrix of the home control parameters in the time dimension, and obtain the first influence factor of the smart home control parameters on the user wake-up process based on the cosine similarity between the common feature vector of the first-order derivative matrix of the home control parameters in the time dimension and the feature vector of the wake-up efficiency vector in the time dimension.

[0117] These shared feature vectors can reflect the common changing characteristics of home control parameters over time. For example, there may be a feature vector that reflects similar trends in the changes of control parameters of multiple smart home devices over certain time periods.

[0118] Simultaneously, the eigenvectors of the wake-up efficiency vector in the time dimension are determined. The first influencing factor of smart home control parameters on the user's wake-up process is obtained by calculating the cosine similarity between the common eigenvectors of the reduced-dimensional first derivative matrix of home control parameters in the time dimension and the eigenvectors of the wake-up efficiency vector in the time dimension. Cosine similarity measures the degree of similarity between two vectors; the higher the similarity, the stronger the correlation between changes in home control parameters and wake-up efficiency, and the larger the first influencing factor, indicating a more significant impact of smart home control parameters on the user's wake-up process.

[0119] The second influencing factor calculation unit is used to obtain the second influencing factor of the user's specified wake-up time on the user's wake-up process based on the correlation analysis results between the specified wake-up time vector and the wake-up efficiency vector. It determines the second influencing factor of the user's specified wake-up time on the user's wake-up process through specific analytical methods, such as calculating the correlation coefficient (or cosine similarity) between the two. If there is a strong correlation between the specified wake-up time and wake-up efficiency, for example, if the wake-up efficiency is high at certain specific times, then the corresponding second influencing factor will be larger, reflecting that the specified wake-up time has a significant impact on the wake-up process.

[0120] The wake-up difficulty assessment unit is used to evaluate the wake-up difficulty of a user based on their natural wake-up history; for example, it determines the ease or difficulty of waking up a user by analyzing the wake-up time required during multiple natural wake-ups. If a user frequently requires a long time to be fully awakened, then that user can be considered to have a high wake-up difficulty.

[0121] The personalized influence factor calibration unit is used to determine the first personalized influence factor of smart home control parameters on the user wake-up process, the second personalized influence factor of the user's specified wake-up time on the user wake-up process, and the user's wake-up difficulty, based on the first influence factor of smart home control parameters on the user wake-up process and the second personalized influence factor of the user's specified wake-up time on the user wake-up process. For example, if the user's wake-up difficulty is high, the first and second influence factors may be appropriately adjusted so that the influence of smart home control parameters and the specified wake-up time on the wake-up process is more in line with the user's actual situation. This determines the first personalized influence factor of smart home control parameters on the user wake-up process and the second personalized influence factor of the user's specified wake-up time on the user wake-up process, providing more accurate data support for subsequently generating sound and light time management strategies that better meet user needs.

[0122] Example 9:

[0123] Based on Example 8, the first impact factor calculation unit includes:

[0124] The first eigenvalue decomposition subunit is used to determine the covariance matrix of the first-order derivative matrix of the home control parameters. Eigenvalue decomposition is performed on the covariance matrix to obtain at least one eigenvalue and a corresponding eigenvector. Each eigenvector is considered a common eigenvector of the first-order derivative matrix of the home control parameters in the time dimension. For example, if the changes in LED lighting control parameters and smart curtain control parameters are correlated in some way, the covariance matrix can reflect this relationship. Then, eigenvalue decomposition is performed on the covariance matrix to obtain at least one eigenvalue and a corresponding eigenvector. These eigenvectors can reflect the common characteristics of the first-order derivative matrix of the home control parameters in the time dimension, so each eigenvector is considered a common eigenvector of the first-order derivative matrix of the home control parameters in the time dimension. For example, a certain eigenvector may reflect the changing trend of multiple smart home control parameters within a specific time period.

[0125] The vector time dimension extension subunit is used to construct the lag feature matrix of the wake-up efficiency vector based on the sliding window method. The lag feature matrix of the wake-up efficiency vector is mean-zeroed and variance-normalized to obtain a standardized lag feature matrix. The sliding window acts like a moving box on the wake-up efficiency vector, capturing a certain number of elements each time and constructing a new matrix, i.e., the lag feature matrix, based on these elements. For example, with a window size of 3, it moves sequentially on the wake-up efficiency vector, taking 3 consecutive elements each time to construct a new row vector. All these row vectors form the lag feature matrix. Then, the lag feature matrix is ​​mean-zeroed and variance-normalized to ensure the data has a uniform scale, facilitating subsequent analysis. The resulting standardized lag feature matrix more accurately reflects the characteristics of the wake-up efficiency vector in the time dimension.

[0126] The second eigenvalue decomposition subunit is used to perform eigenvalue decomposition on the covariance matrix of the standardized lag feature matrix to obtain all eigenvalues ​​and corresponding eigenvectors of the covariance matrix of the standardized lag feature matrix. The eigenvector corresponding to the largest eigenvalue among all eigenvalues ​​and corresponding eigenvectors of the covariance matrix of the standardized lag feature matrix is ​​taken as the eigenvector of the wake-up efficiency vector in the time dimension. This is because the eigenvector corresponding to the largest eigenvalue contains the most important change information in the standardized lag feature matrix and can represent the key features of the wake-up efficiency vector in the time dimension.

[0127] The cosine similarity calculation subunit is used to calculate the cosine similarity between each shared feature vector with a corresponding eigenvalue greater than a threshold and the wake-up efficiency vector in the time dimension. The threshold is set to filter out shared feature vectors that have a significant impact on the results. Cosine similarity measures the similarity between two vectors in a direction, with a value ranging from -1 to 1; the closer to 1, the more similar the two vectors are. By calculating cosine similarity, we can understand the similarity between the shared features of the first-order derivative matrix of the home control parameters and the features of the wake-up efficiency vector.

[0128] The correlation synthesis subunit is used to calculate the weight value of each common feature vector whose corresponding feature value is greater than the threshold based on all feature values ​​greater than the threshold. That is, the ratio of a single feature value greater than the threshold to the sum of all feature values ​​greater than the threshold is used as the weight value of the single common feature vector whose corresponding feature value is greater than the threshold. The larger the feature value, the higher the weight of the corresponding common feature vector, because it contains more important information.

[0129] Based on the weights of all shared feature vectors with corresponding eigenvalues ​​greater than a threshold, the cosine similarity of each shared feature vector with a corresponding eigenvalue greater than the threshold and the wake-up efficiency vector in the time dimension is weighted and summed to obtain the first influencing factor of smart home control parameters on the user wake-up process. This first influencing factor comprehensively considers the similarity between home control parameters and the wake-up efficiency vector in the time dimension, as well as the importance of each shared feature vector, and can more accurately reflect the impact of smart home control parameters on the user wake-up process.

[0130] Example 10:

[0131] Based on Example 8, the personalized impact factor calibration unit includes:

[0132] The adjustment coefficient determination sub-unit is used to determine the average wake-up time based on the user's natural wake-up records. This requires statistically analyzing the time spent by the user from the start to full wake-up in each natural wake-up process, and then calculating the average of these times. For example, after statistically analyzing multiple natural wake-up processes, the average wake-up time is found to be 15 minutes.

[0133] The first adjustment coefficient is determined based on the average wake-up time and the user's wake-up difficulty, which is:

[0134] First adjustment factor = 1 + (user's wake-up difficulty - average wake-up time) × positive calibration factor;

[0135] The positive calibration factor has a value range of [0,2], and a value of 0.5 is recommended here. The unit for the user's wake-up difficulty is minutes.

[0136] The second adjustment coefficient is determined based on the user's wake-up difficulty and difficulty sensitivity coefficient, which is:

[0137] Second adjustment coefficient = 1 - Difficulty sensitivity coefficient (1 - User's wake-up difficulty).

[0138] The difficulty sensitivity coefficient is a preset value, ranging from [0,1]. A value of 0.3 is recommended here to measure the impact of wake-up difficulty on the adjustment coefficient. For example, if the user wakes up with high difficulty and has a high difficulty sensitivity coefficient, the second adjustment coefficient will also increase accordingly.

[0139] The first influence factor personalized calibration subunit is used to take the product of the first influence factor, the first adjustment coefficient, and the importance weight of the control parameter on the user's wake-up process of the smart home control parameter as the first personalized influence factor of the smart home control parameter on the user's wake-up process. By multiplying these three factors, the first influence factor can be calibrated according to the user's average wake-up time, wake-up difficulty, and the importance of the control parameter itself, so that it is more in line with the user's personalized situation. For example, if the first influence factor is 0.6, the first adjustment coefficient is 1.2, and the importance weight of a certain control parameter is 0.8, then the first personalized influence factor is 0.6 × 1.2 × 0.8 = 0.576.

[0140] The second influence factor personalized calibration subunit is used to take the product of the second personalized influence factor, the second adjustment coefficient, and the time window adjustment coefficient of the user-specified wake-up time on the user wake-up process as the second personalized influence factor of the user-specified wake-up time on the user wake-up process.

[0141] The time window adjustment factor is related to the time range of the wake-up time set by the user. For example, within the time period of 6-8 am, the impact of different times on the wake-up effect may be different. The time window adjustment factor is used to reflect this difference.

[0142] In this way, the second personalized influence factor is calibrated by comprehensively considering user wake-up difficulty, difficulty sensitivity coefficient, and time window factors. For example, if the second personalized influence factor is 0.7, the second adjustment coefficient is 1.3, and the time window adjustment coefficient is 0.9, then the calibrated value of the second personalized influence factor is 0.7 × 1.3 × 0.9 = 0.819. The two personalized influence factors obtained in this way can more accurately reflect the user's personalized characteristics in smart home control parameters and specified wake-up times, providing a basis for generating more optimized sound and light time management strategies.

[0143] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A natural wake-up system based on offline voice recognition control and sound-light time management, characterized in that, include: The offline speech recognition module is used to recognize the user's voice commands received by the microphone in an offline state; The main control module is used to obtain the first personalized influence factor of smart home control parameters on the user's wake-up process and the second personalized influence factor of the user's specified wake-up time on the user's wake-up process based on the home control parameter matrix, wake-up efficiency vector, and specified wake-up time vector generated by the user's natural wake-up records, and to generate an audio-visual time management strategy by combining artificial intelligence and the user's voice commands. The time management module is used to record real-time time. The RF communication module is used to send control signals to smart home devices that can receive RF signals based on the audio-visual time management strategy and real-time time. The infrared signal transmitting module is used to send control signals to smart home devices that can receive infrared control signals, based on the audio-visual time management strategy, real-time time, and smart home remote control codes stored in the infrared remote control code library. The audio playback module is used to play built-in audio files through a speaker based on sound and light time management strategies and real-time time. The main control module includes: The data processing submodule is used to generate a home control parameter matrix, a wake-up efficiency vector, and a specified wake-up time vector based on the user's natural wake-up records. The correlation analysis submodule is used to generate a first personalized influence factor of smart home control parameters on the user's wake-up process and a second personalized influence factor of the user's specified wake-up time on the user's wake-up process based on the correlation analysis results between the home control parameter matrix and the wake-up efficiency vector and the correlation analysis results between the specified wake-up time vector and the wake-up efficiency vector. The audio-visual time management strategy generation submodule is used to obtain the audio-visual time management strategy based on the first personalized influence factor of the user wake-up process, the second personalized influence factor of the user-specified wake-up time on the user wake-up process, the latest user-specified wake-up time represented by the latest received voice command, and the pre-built audio-visual time management strategy generation model. The data processing submodule includes: The wake-up efficiency value analysis unit is used to analyze the wake-up efficiency value of each natural wake-up process based on the continuous control execution time of each smart home in the sound and light time management strategy executed in each natural wake-up process in the user's natural wake-up record, and the corresponding time required for the user to be fully awakened. The home control parameter matrix generation unit is used to generate a home control parameter matrix based on the continuous control parameter threads of all smart homes in the sound and light time management strategy executed in all natural wake-up processes in the natural wake-up record. The wake-up efficiency vector generation unit is used to generate a wake-up efficiency vector based on the wake-up efficiency values ​​of all natural wake-up processes in the natural wake-up record. The wake-up time vector generation unit is used to generate a specified wake-up time vector based on the user-specified wake-up time of all natural wake-up processes in the natural wake-up record. The correlation analysis submodule includes: The matrix dimensionality reduction and differentiation unit is used to reduce the dimensionality of the home control parameter matrix to obtain the home control parameter dimensionality reduction matrix. Based on the dimensionality reduction value sequence of the control parameters of each smart home in the home control parameter dimensionality reduction matrix, the dimensionality reduction function of the control parameters of each smart home is generated. Based on the first derivative values ​​of the dimensionality reduction functions of the control parameters of all smart homes at all times, the first derivative matrix of the home control parameter dimensionality reduction is generated. The first influence factor calculation unit is used to analyze at least one common feature vector of the first-order derivative matrix of the home control parameters in the time dimension, and obtain the first influence factor of the smart home control parameters on the user wake-up process based on the cosine similarity between the common feature vector of the first-order derivative matrix of the home control parameters in the time dimension and the feature vector of the wake-up efficiency vector in the time dimension. The second influence factor calculation unit is used to obtain the second influence factor of the user's specified wake-up time on the user's wake-up process based on the correlation analysis results between the specified wake-up time vector and the wake-up efficiency vector. The wake-up difficulty assessment unit is used to assess the wake-up difficulty of a user based on their natural wake-up history. The personalized influence factor calibration unit is used to determine the first personalized influence factor of smart home control parameters on the user wake-up process and the second personalized influence factor of the user-specified wake-up time on the user wake-up process based on the first influence factor of smart home control parameters on the user wake-up process, the second influence factor of the user-specified wake-up time on the user wake-up process, and the user's wake-up difficulty.

2. The natural wake-up system based on offline speech recognition control and audio-visual time management according to claim 1, characterized in that, Also includes: The rechargeable battery provides power to the microphone, offline voice recognition module, main control module, time management module, RF communication module, infrared remote control code library, infrared signal transmission module, audio playback module, speaker, power management module, and LED display module. The power management module is used to issue a reminder signal when the remaining capacity of the rechargeable battery is lower than a threshold, and at the same time, adjust the working mode of the main control module to power saving mode. The LED display module is used to display real-time time, remaining battery power, time information in the latest voice command, and the working status of all communicable smart devices. The storage module is used to store all received voice commands, the ongoing working status of all communicable smart devices, and all currently determined audio-visual time management strategies.

3. The natural wake-up system based on offline speech recognition control and audio-visual time management according to claim 1, characterized in that, Smart home devices that can receive RF signals include LED lights, smart curtains, and aroma diffusers; LED lights gradually illuminate and change color temperature based on the received control signals; The smart curtains gradually open based on the received control signals; The aroma diffuser starts working based on the received control signal.

4. The natural wake-up system based on offline speech recognition control and audio-visual time management according to claim 1, characterized in that, Smart home devices that can receive RF signals include air conditioners, which gradually adjust the temperature to the corresponding level based on the received control signals.

5. The natural wake-up system based on offline speech recognition control and audio-visual time management according to claim 1, characterized in that, The wake-up efficiency value analysis unit includes: The Single Wake-up Efficiency Value Determination Subunit is used to take the ratio of the continuous control execution time of each smart home in each natural wake-up process to the time required for the user to be fully awakened in the corresponding natural wake-up process as the Single Wake-up Efficiency Value of each smart home in the corresponding natural wake-up process. The Single Wake-up Efficiency Value Summarization Subunit is used to sum the single wake-up efficiency values ​​of all smart home devices in a single natural wake-up process based on the wake-up weights of all smart home devices, so as to obtain the wake-up efficiency value for each natural wake-up process.

6. The natural wake-up system based on offline speech recognition control and audio-visual time management according to claim 1, characterized in that, The first impact factor calculation unit includes: The first eigenvalue decomposition subunit is used to determine the covariance matrix of the first-order derivative matrix of the home control parameters. The eigenvalue decomposition is performed on the covariance matrix of the first-order derivative matrix of the home control parameters to obtain at least one eigenvalue and the corresponding eigenvector. Each eigenvector is regarded as the common eigenvector of the first-order derivative matrix of the home control parameters in the time dimension. The vector time dimension extension sub-unit is used to construct the lag feature matrix of the wake-up efficiency vector based on the sliding window method. The mean and variance of the lag feature matrix of the wake-up efficiency vector are normalized to obtain the standardized lag feature matrix. The second eigenvalue decomposition subunit is used to perform eigenvalue decomposition on the covariance matrix of the standardized lag feature matrix to obtain all eigenvalues ​​and corresponding eigenvectors of the covariance matrix of the standardized lag feature matrix. The eigenvector corresponding to the largest eigenvalue among all eigenvalues ​​and corresponding eigenvectors of the covariance matrix of the standardized lag feature matrix is ​​taken as the eigenvector of the wake-up efficiency vector in the time dimension. The cosine similarity calculation subunit is used to calculate the cosine similarity between the feature vectors of each common feature vector with a corresponding feature value greater than the threshold and the wake-up efficiency vector in the time dimension. The correlation synthesis subunit is used to calculate the weight value of each common feature vector whose corresponding feature value is greater than the threshold based on all feature values ​​greater than the threshold. Based on the weight values ​​of all common feature vectors whose corresponding feature values ​​are greater than the threshold, the cosine similarity of each common feature vector whose corresponding feature value is greater than the threshold and the wake-up efficiency vector in the time dimension is weighted and summed to obtain the first influencing factor of smart home control parameters on the user wake-up process.

7. The natural wake-up system based on offline speech recognition control and audio-visual time management according to claim 1, characterized in that, Personalized impact factor calibration unit, including: The adjustment coefficient determination subunit is used to determine the average wake-up time based on the user's natural wake-up record, determine the first adjustment coefficient based on the average wake-up time and the user's wake-up difficulty, and determine the second adjustment coefficient based on the user's wake-up difficulty and difficulty sensitivity coefficient. The first influence factor personalized calibration subunit is used to take the product of the first influence factor of the smart home control parameters on the user wake-up process, the first adjustment coefficient, and the importance weight of the control parameters as the first personalized influence factor of the smart home control parameters on the user wake-up process. The second influence factor personalized calibration subunit is used to take the product of the second personalized influence factor, the second adjustment coefficient, and the time window adjustment coefficient of the user-specified wake-up time on the user wake-up process as the second personalized influence factor of the user-specified wake-up time on the user wake-up process.

Images