An artificial intelligence range hood and kitchen appliance linkage control system and method

Abstract

An AI-powered range hood and kitchen appliance linkage control system includes an AI range hood main body, linkage appliance modules, an environmental sensing module, an AI decision-making module, a communication module, and a human-computer interaction module for user operation. The linkage appliance modules communicate with the AI ​​decision-making module through the communication module, receiving control commands from the modules connected to the AI ​​decision-making module and controlling corresponding appliance units to link with the AI ​​range hood main body. The environmental sensing module collects environmental data in real time and feeds it back to the AI ​​decision-making module. The AI ​​decision-making module generates commands based on the environmental data and sends them to the linkage appliance modules and the AI ​​range hood main body. This system constructs a unified collaborative control system for kitchen equipment, achieving integrated intelligent management and panoramic visual monitoring.

Description

Technical Field

[0001] This invention relates to the field of range hoods and kitchen appliances, specifically to an artificial intelligence-based control system and method for linking range hoods and kitchen appliances. Background Technology

[0002] Traditional range hoods mostly use mechanical switches or simple infrared sensor controls, lacking the ability to dynamically respond to environmental changes. Users often need to manually adjust the airflow frequently during cooking, especially during stir-frying and deep-frying, which generate large amounts of smoke. The equipment's response is noticeably delayed, often exhibiting "reaction lag." The range hood often passively increases suction power only after the smoke has already dispersed, leading to smoke escape and increased kitchen cleaning burden. Furthermore, traditional range hoods typically only have basic overheat protection functions, lacking effective monitoring capabilities for potential risks such as minor gas leaks, abnormal oil temperature increases, and dry burning. As the kitchen is a high-risk area for home safety accidents, existing equipment systems struggle to establish a systematic risk warning mechanism, posing significant safety hazards, especially for elderly people living alone or those with limited cooking experience.

[0003] Furthermore, traditional range hoods and other kitchen appliances operate independently, lacking a unified data communication protocol. For example, changes in cooktop heat cannot be synchronized to the range hood in real time, and the high-temperature steam generated by the oven cannot automatically trigger the ventilation system to work in tandem. This causes a rapid increase in localized temperature and humidity in the kitchen, which not only affects the cooking experience but may also accelerate equipment aging and shorten its lifespan. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide an artificial intelligence-based range hood and kitchen appliance linkage control system and method, aiming to build a unified collaborative control system for kitchen equipment and realize integrated intelligent management and panoramic visual monitoring.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: An AI-powered range hood and kitchen appliance linkage control system includes an AI range hood body, a linkage appliance module, an environmental sensing module, an AI decision-making module, a communication module, and a human-computer interaction module for user operation. The linked appliance module communicates with the AI ​​decision module through the communication module, and is used to receive control commands from the AI ​​decision module connected module, and control the corresponding appliance unit to link with the AI ​​range hood body; The environmental perception module collects environmental data in real time and feeds the collected environmental data back to the AI ​​decision-making module; The AI ​​decision-making module generates instructions based on the environmental data and sends them to the linked appliance module and the AI ​​range hood body.

[0006] Compared to existing technologies, this system seamlessly integrates the range hood with kitchen appliances such as cooktops and ovens into a cohesive and intelligent whole through advanced communication protocols. Leveraging environmental sensing and artificial intelligence technologies, the system can accurately identify various cooking scenarios and dynamically adjust the range hood accordingly, significantly improving its ability to capture and efficiently purify cooking fumes. Simultaneously, the system possesses comprehensive safety monitoring capabilities, providing real-time warnings and automatic responses to risks such as gas leaks, abnormal high temperatures, and circuit failures, building a proactive intelligent safety barrier for the home kitchen. Regarding human-computer interaction, the system adaptively learns user habits, greatly simplifying operation processes and reducing manual intervention to achieve a smart experience of "devices proactively collaborating and users controlling seamlessly." Overall, by integrating artificial intelligence and IoT technologies, the system constructs a smart kitchen ecosystem characterized by interconnected devices, intelligent scenario collaboration, and green and efficient management, bringing a safer, more convenient, and healthier kitchen environment to modern home life.

[0007] Furthermore, the AI ​​decision-making module includes a data preprocessing unit, a scene recognition unit, and an instruction generation unit. The data preprocessing unit performs noise reduction and normalization on the environmental data. The scene recognition unit identifies the current cooking type based on the processed environmental data and preset cooking scene information. The instruction generation unit generates control instructions based on the current cooking type and sends them to the AI ​​range hood body and the linked appliance module. Through the noise reduction and normalization processing of the data preprocessing unit, measurement errors and noise interference from environmental sensors are effectively eliminated, providing a high-quality, standardized data foundation for subsequent decision-making.

[0008] This system possesses intelligent scene recognition and dynamic control capabilities. By analyzing environmental data in real time and comparing it with preset scene models, it can accurately determine various cooking types such as stir-frying and steaming, and then automatically generate dynamic control strategies to achieve precise adjustment of the range hood's airflow and intelligent linkage between multiple devices, significantly improving oil fume purification and overall smoke extraction efficiency. The system integrates a multi-dimensional proactive safety protection mechanism, continuously monitoring key parameters such as gas concentration and cookware temperature, conducting real-time multi-dimensional risk assessments, and initiating a tiered early warning mechanism based on risk levels, building a new safety protection system that shifts from reactive alarms to proactive prevention. Furthermore, the system continuously learns user habits and optimizes its control strategies, making the equipment increasingly tailored to individual needs and possessing long-term self-improvement capabilities.

[0009] Furthermore, the AI ​​decision-making module also includes a self-learning unit. This unit continuously collects and stores a dataset of the user's historical cooking operations and optimizes the instruction generation logic based on the dataset. This logic is stored within the self-learning unit to update the cooking scene information of the scene recognition unit. By continuously collecting historical user operation data through the self-learning unit, a personalized operation dataset is constructed, enabling the system to break free from general preset rules and deeply adapt to individual user cooking habits and operational preferences, achieving precise "one-household-one-policy" control. This mechanism establishes a data-driven error correction loop, effectively solving the problems caused by differences in cooking habits, regional dietary cultures, or environmental changes. Furthermore, the environmental data includes oil fume concentration, gas concentration, cooking area temperature, and cooking actions. The environmental sensing module collects the oil fume concentration every 50-120ms, the gas concentration every 50-80ms, and the cooking area temperature every 150-250ms, while capturing the cooking actions at a frame rate of 20-40fps. The gas concentration is collected at the highest frequency, enabling millisecond-level perception and instantaneous response to high-risk events such as gas leaks. This provides a crucial time window for the system to implement proactive safety interventions, significantly improving the timeliness and reliability of kitchen safety protection. Simultaneously, cooking actions are continuously captured at video frame rates and time-aligned with periodically collected environmental parameters such as temperature and oil fumes, forming a synchronously fused multimodal data source. This design not only fully recreates the correlation between dynamic behavior and static environmental states during cooking but also provides a highly consistent and interpretable data foundation for subsequent scene recognition and intelligent decision-making.

[0010] Furthermore, the data preprocessing unit uses a Kalman filter algorithm to remove fluctuations in smoke concentration, gas concentration, and cooking zone temperature. Then, a normalization algorithm maps the smoke concentration, cooking zone temperature, and gas concentration to set values. Finally, a CNN+LSTM hybrid neural network model identifies the cooking actions. By introducing the Kalman filter algorithm, sensor noise and instantaneous fluctuations in the complex kitchen environment are effectively eliminated, ensuring the continuity and accuracy of smoke concentration, gas concentration, and temperature data. Combined with normalization processing, the problem of fusing multi-source heterogeneous data is solved, providing high-quality, standardized input for deep learning models.

[0011] Furthermore, the scene recognition unit incorporates a pre-set cooking scene model. The scene recognition unit identifies the current cooking type and outputs it to the instruction generation unit. Simultaneously, it assesses the safety risk level based on gas concentration and cooking area temperature data collected by the environmental perception module and outputs the result to the human-computer interaction module. By integrating cooking scene recognition and safety risk level assessment functions in parallel within the scene recognition unit, this design achieves efficient reuse of multi-source sensor data, effectively reduces functional module redundancy, optimizes system computing resource allocation, and ensures real-time synchronization of intelligent control instruction generation and safety monitoring and early warning.

[0012] Furthermore, the instruction generation unit generates control instructions based on the received current cooking type. These control instructions are output to the AI ​​range hood body and the linked appliance module via the communication module. The control instructions include equipment speed adjustment instructions and AI range hood body cleaning instructions. The instruction generation unit not only generates speed adjustment instructions based on the current cooking type to achieve collaborative operation between the AI ​​range hood body and the linked appliance module, but also innovatively integrates equipment cleaning instructions. Thus, the control scope of the system expands from simple environmental comfort adjustment to the management of the equipment's own health status, ultimately constructing a complete intelligent service closed loop covering environmental optimization and equipment maintenance.

[0013] This invention also provides a method for the linkage control of an artificial intelligence range hood and kitchen appliances. The method is applied to the aforementioned artificial intelligence range hood and kitchen appliance linkage control system, and includes: Step S1. Obtaining Instructions: The human-computer interaction module obtains the user's operation instructions; Step S2. System Startup: The human-computer interaction module converts the acquired operation instructions into system control signals and outputs the system control signals to the AI ​​decision module. The AI ​​decision module outputs start instructions to the AI ​​range hood body, the linkage appliance module, the environmental perception module, the communication module, and the human-computer interaction module. Step S3. Data Acquisition: The environmental perception module collects data in real time and outputs the data to the AI ​​decision-making module. The collected data includes oil fume concentration data M1, gas concentration data M2, cooking area temperature data M3, and captured cooking action data M4. Step S4. Data Processing: The AI ​​decision module uses the Kalman filter algorithm to remove the fluctuating data of M1-M3 from the collected data, and then uses the normalization algorithm to map the data of oil fume concentration, cooking zone temperature and gas concentration to the set values. Step S5. Scene Recognition: The scene recognition unit extracts the features of cooking actions through a CNN+LSTM hybrid neural network model and combines them with the processed cooking area temperature data M3 and oil fume concentration data M1 to obtain a data set. The data set is then compared and matched with a preset cooking scene model to obtain a matching result, and the matching result is output to the instruction generation unit. Step S6. Control Execution: The instruction generation unit generates targeted control instructions based on the matching results of the scene recognition unit, and outputs the control instructions to the AI ​​range hood body and the linked appliance module through the communication module. The AI ​​range hood body and the linked appliance module switch to the corresponding gear according to the control instructions, and the control instructions are stored in the historical data of the self-learning unit. Step S7. Cooking end determination: After the set waiting time, the scene recognition unit detects no cooking action through the CNN+LSTM hybrid neural network model. At the same time, the environmental perception module detects that the cooking area temperature M3 drops to the set value and the oil fume concentration M1 drops to the set value. The AI ​​range hood body determines that cooking is over and generates a cooking end command to output to the AI ​​decision module. Step S8. Post-cooking processing and purification: The environmental sensing module does not detect any gas concentration data M2 exceeding the set value throughout the cooking process and feeds the result back to the AI ​​decision module. After receiving the cooking end command, the AI ​​decision module generates a purification command to the AI ​​range hood body and the linked appliance module. The AI ​​range hood body runs at low speed for a set period of time according to the purification command and then automatically turns off the power. The linked appliance module turns on the corresponding air purification appliance and runs for a set period of time before automatically turning off the power.

[0014] Compared to existing technologies, this method achieves a comprehensive upgrade of kitchen systems from passive response to proactive perception, intelligent decision-making, precise execution, and continuous learning. Through the synergistic effect of multimodal perception fusion, hybrid AI model recognition, closed-loop control, and self-learning mechanisms, the system not only significantly improves the efficiency of fume purification and ease of operation, but also forms sustainable technological advantages in areas such as safety protection, energy efficiency optimization, equipment maintenance, and long-term system evolution. This provides a complete methodological framework to support the construction of truly intelligent, highly reliable, and customizable next-generation smart kitchens.

[0015] Furthermore, in step S7, the cooking zone temperature M3 decreases to a set value of 90℃-95℃, and the oil fume concentration M1 decreases to a set value of 0.5mg / m³-0.9mg / m³, and the set values ​​are normalized. By normalizing heterogeneous physical quantities such as temperature and oil fume concentration, the numerical gap caused by different dimensions is effectively eliminated, providing standardized input features for the joint analysis of multimodal data. This not only reduces the computational complexity and resource consumption of subsequent AI algorithms and accelerates model inference speed, but also improves the accuracy and stability of multi-source data fusion, ensuring that the decision-making logic can operate reliably under different sensor configurations.

[0016] Furthermore, in step S6, the self-learning unit stores the gear control commands of the AI ​​range hood and linked appliance modules that the user has not manually adjusted and maps them to the cooking scenario model. This mechanism, through silent learning of user habits, construction of personalized cooking scenario response models, and continuous optimization of decision-making strategies, enables the system not only to adapt to the needs of different users but also to achieve a sustainable learning loop from data accumulation to strategy evolution, and from single-time response to long-term optimization. This marks a fundamental shift in intelligent kitchen systems from "operating according to fixed rules" to "evolving collaboratively with users." Attached Figure Description

[0017] Figure 1 This is a schematic diagram of an artificial intelligence-based range hood and kitchen appliance linkage control system according to the present invention. Figure 2 This is a flowchart of an artificial intelligence-based range hood and kitchen appliance linkage control method according to the present invention. Detailed Implementation

[0018] like Figure 1 As shown, an AI-powered range hood and kitchen appliance linkage control system includes an AI range hood body (1), a linkage appliance module (2), an environmental sensing module (3), an AI decision module (4), a communication module (5), and a human-computer interaction module (6) for user operation. The linkage appliance module (2) communicates with the AI ​​decision module (4) through the communication module (5) to receive control commands from the module connected to the AI ​​decision module (4) and control the corresponding appliance unit to link with the AI ​​range hood body (1). The environmental sensing module (3) collects environmental data in real time and feeds the collected environmental data back to the AI ​​decision module (4). The AI ​​decision module (4) generates commands based on the environmental data and sends them to the linkage appliance module (2) and the AI ​​range hood body (1).

[0019] The AI ​​decision module (4) includes a data preprocessing unit, a scene recognition unit, an instruction generation unit, and a self-learning unit. The data preprocessing unit is used to perform noise reduction and normalization processing on the environmental data. The scene recognition unit identifies the current cooking type based on the processed environmental data and the preset cooking scene information. The instruction generation unit generates control instructions based on the current cooking type and sends them to the AI ​​range hood body (1) and the linkage appliance module (2).

[0020] The data preprocessing unit performs noise reduction and normalization processing on the data collected by the environment perception module (3), and outputs the processed data to the scene recognition unit. By setting up a built-in data preprocessing unit, this invention not only ensures the consistency and purity of the input data, but also significantly improves the accuracy and real-time performance of subsequent scene recognition. At the same time, through modular decoupling, it greatly enhances the maintainability and generalization ability of the system, laying a high-quality data foundation for the continuous optimization of the self-learning unit.

[0021] The preset cooking scenario information is a preset cooking scenario model built into the scenario recognition unit. The scenario recognition unit identifies the current cooking type based on the data processed by the data preprocessing unit and the cooking scenario model, and outputs the result to the instruction generation unit. At the same time, it judges the safety risk level based on the data of gas concentration and cooking area temperature collected by the environmental perception module (3) and outputs the result to the human-computer interaction module (6). The cooking scenario model includes cooking types such as stir-frying, steaming and stewing.

[0022] The instruction generation unit generates control instructions based on the scene recognition results of the scene recognition unit and outputs them to the communication module (5). The control instructions include equipment gear adjustment instructions and cleaning instructions for the AI ​​range hood body (1). For example, the scene recognition unit identifies the current cooking type as stir-fry scene and outputs the result to the instruction generation unit. Based on the scene recognition results, the instruction generation unit generates targeted control instructions and sends stir-fry instructions to the AI ​​range hood body (1) and the gas stove, so that the AI ​​range hood body (1) and the gas stove switch to stir-fry gear. At the same time, after cooking is finished, the AI ​​range hood body (1) is controlled to start the cleaning function.

[0023] The self-learning unit continuously collects and stores a dataset of the user's historical cooking operations, and optimizes the instruction generation logic based on the dataset. The logic is stored in the self-learning unit. The self-learning unit records user cooking habits such as the fan speed and cooking time of the commonly used AI range hood body (1), and continuously optimizes the instruction generation logic based on the recorded data.

[0024] The principle for determining the safety risk level is to quantify and classify the safety risk by using a two-dimensional threshold matrix of gas concentration and temperature and the dynamic trend of real-time data changes; the safety risk levels include safe level, early warning level, emergency level and dangerous level.

[0025] The scene recognition unit presets a static threshold for gas concentration, a static threshold for temperature, and a data change rate threshold. The static threshold for gas concentration includes a safety threshold C0, a warning threshold C1, and a danger threshold C2 for gas concentration, with the current gas concentration value C in %LEL. The static threshold for temperature includes a normal threshold T0 for cooking area temperature, a warning threshold T1 for cooking area temperature, and a danger threshold T2 for cooking area temperature, with the current cooking area temperature value T in °C. The data change rate threshold includes a gas concentration rise rate threshold α and a temperature rise rate threshold β. For example, if the gas concentration rises by ≥3%LET within 1 minute or the temperature rises by ≥20 °C within 1 minute, a trend warning is triggered.

[0026] The safety level is determined by the following conditions: C≤C0 and T≤T0, the gas concentration rises less than α in the last minute and the temperature rises less than β in the last minute; all parameters are within the safe range and there is no accelerating upward trend.

[0027] The conditions for determining the warning level are: C0 < C ≤ C1 and T ≤ T0, the gas concentration rises by less than α in the last minute; or C ≤ C0 and T0 ≤ T ≤ T1, the temperature rises by less than β in the last minute; a single parameter shows a slight abnormality, but the trend is stable and there is no accelerating upward trend.

[0028] The emergency level is determined by the following conditions: C1 < C ≤ C2 and T ≤ T1 or C ≤ C1 and T1 ≤ T ≤ T2, the gas concentration increases by ≥ α or the temperature increases by ≥ β in the most recent minute, the parameter reaches a high risk level, or any one of the parameters shows a rapid upward trend.

[0029] The criteria for determining the hazard level are C > C2 or T > T2 or C1 < C ≤ C2 and T1 < T ≤ T2, where the parameters reach the hazard level, or the gas and temperature are simultaneously in their respective high-risk ranges, indicating a compound risk.

[0030] The communication module (5) enables real-time data transmission and interaction between the AI ​​decision module (4), the AI ​​range hood body (1), and the linked appliance module (2) via Wi-Fi, Bluetooth, or ZigBee communication protocols. This system integrates multiple communication protocols through a unified gateway to achieve data interoperability and intelligent linkage between different devices. The system has broad compatibility, real-time response capability, and multiple security protections to ensure real-time data interaction between the AI ​​decision module (4), the AI ​​range hood body (1), and the linked appliances, reducing transmission delay issues.

[0031] The human-computer interaction module (6) includes a touch display unit and a voice interaction unit. The touch display unit collects the control commands input by the user and outputs them to the AI ​​decision module (4) to control the AI ​​range hood body (1) and the linkage electrical module (2). At the same time, the touch display unit displays the operating status of the equipment. The voice interaction unit converts the user's voice commands into system control signals through real-time sound pickup and recognition. The AI ​​decision module (4) monitors the operation of each system in real time and outputs the results to the touch display unit in real time. Users can manually input control commands and view the operating status of the equipment through the touch screen, or control the system operation through voice commands. When abnormal conditions such as gas leakage, abnormal temperature, or short circuit are detected, the system will alert the user through sound and light alarms. Users can disconnect the circuit on the APP.

[0032] The linked appliance module (2) includes various kitchen appliances. The linked appliance module (2) operates by receiving control commands from the AI ​​decision module (4) in real time. At the same time, the linked appliance module (2) outputs real-time operating status signals to the AI ​​decision module (4) through the communication module (5). The linked appliance module (2) includes kitchen appliances such as gas stoves, air purifiers, ovens, and disinfection cabinets. Each appliance is equipped with an intelligent control interface, which supports receiving external control commands and feeding back operating status signals.

[0033] like Figure 2 As shown, an AI-powered range hood and kitchen appliance linkage control method includes the aforementioned AI-powered range hood and kitchen appliance linkage control system, the method comprising: Step S1. Obtaining instructions: The voice interaction unit obtains the user's operation instructions; Step S2. System Start-up: The voice interaction unit converts the acquired operation instructions into system control signals and outputs the system control signals to the AI ​​decision module (4). The AI ​​decision module (4) outputs electrical signals to the AI ​​range hood body (1), the linkage appliance module (2), the environmental perception module (3), the communication module (5), and the human-computer interaction module (6) to realize the start-up. Step S3. Data Acquisition: The environmental perception module (3) collects data in real time and outputs the data to the AI ​​decision module (4). The collected data includes oil fume concentration data M1, gas concentration data M2, cooking area temperature data M3, and captured cooking action data M4. The environmental perception module (3) collects oil fume concentration M1 every 50-120ms, gas concentration M2 every 50-80ms, cooking area temperature M3 every 150-250ms, and captures cooking action M4 at a frame rate of 20-40fps. Step S4. Data processing: The AI ​​decision module (4) uses the Kalman filter algorithm to remove the fluctuating data of M1-M3 from the collected data, and then uses the normalization algorithm to map the data of oil fume concentration, cooking area temperature and gas concentration to the set values; for example, the data preprocessing unit of the AI ​​decision module (4) uses the Kalman filter algorithm (filter coefficient is 0.85) to remove the fluctuating data of M1-M3, and then uses the Min-MAX mathematical operation to perform normalization processing, mapping the oil fume concentration (initial value 2.3mg / m³) to 0.46, the temperature (initial value 210℃) to 0.95, and the gas concentration (initial value 80ppm) to 0.05; Step S5. Scene recognition: The scene recognition unit extracts the features of cooking actions in the CAM of the environment perception module (3) through the CNN+LSTM hybrid neural network model and combines them with the processed cooking area temperature data M3 and oil fume concentration data M1 to obtain a data set. The data set is compared and matched with the preset cooking scene model to obtain the matching result, and the matching result is output to the instruction generation unit. Step S6. Control Execution: The instruction generation unit generates targeted control instructions based on the matching results of the scene recognition unit, and outputs the control instructions to the AI ​​range hood body (1) and the linkage appliance module (2) through the communication module (5). At the same time, the control instructions are stored in the historical data of the self-learning unit. The AI ​​range hood body (1) and the linkage appliance module (2) switch the corresponding gear according to the control instructions. Step S7. Cooking end determination: After the set waiting time, the scene recognition unit detects no cooking action through CAM, and at the same time, the temperature M3 of the cooking area collected by the environmental perception module (3) drops to the set value and the oil fume concentration M1 drops to the set value. The AI ​​range hood body (1) determines that cooking is over and generates a cooking end command to output to the AI ​​decision module (4). For example, after 15 minutes, CAM detects "no stir-frying action", the temperature collected by M3 drops to 90℃-95℃ (normalized to 0.36-0.38), and the oil fume concentration detected by M1 drops to 0.5mg / m³-0.9mg / m³ (normalized to 0.10-0.18). The range hood determines that cooking is over and turns off the power. Step S8. Post-cooking processing and purification: The environmental sensing module (3) does not detect any gas concentration data M2 exceeding the set value during the entire cooking process and feeds the result back to the AI ​​decision module (4). After receiving the cooking end command, the AI ​​decision module (4) generates a purification command to the AI ​​range hood body (1) and the linkage appliance module (2). The AI ​​range hood body (1) runs at low speed for a set time according to the purification command and then automatically shuts off the power. The linkage appliance module (2) turns on the corresponding air purification appliance and runs for a set time before automatically shutting off the power. No delayed concentration exceeding the standard was detected during the entire cooking process. After cooking, a command will be generated, and the AI ​​range hood body (1) will maintain low speed operation for 5 minutes to cool down and purify. After 5 minutes, the AI ​​range hood will automatically shut off the power and start the air purifier to run for 10 minutes.

[0034] According to steps S5 and S6, the scene recognition unit extracts the "high-frequency stir-frying action" feature in CAM through the CNN+LSTM hybrid neural network model. The feature is that the stir-frying frequency is 4 times / second. Combined with the normalized temperature of 0.98≥0.8 and the oil fume concentration of 0.46≥0.4, the data group is compared and matched with the preset cooking scene model to obtain the matching result as a stir-frying scene. The instruction generation unit generates an instruction, and the wind power of the AI ​​range hood body (1) is switched from the first level to the maximum level, and the gas stove maintains the high firepower level.

[0035] In step S6, the self-learning unit stores the gear control commands of the AI ​​range hood body (1) and the linked appliance module (2) that the user has not manually adjusted and maps them to the cooking scene model. The self-learning unit records that the user has not manually adjusted the gear this time, marks the gear mapping of the cooking scene, and automatically updates the mapping mode after accumulating multiple similar scenes. The next time a similar scene occurs, there is no need to identify it, and it can be directly adjusted.

[0036] Based on the disclosure and teachings of the foregoing specification, those skilled in the art can make changes and modifications to the above embodiments. Therefore, the present invention is not limited to the specific embodiments disclosed and described above, and some modifications and changes to the present invention should also fall within the protection scope of the claims of the present invention. Furthermore, although some specific terms are used in this specification, these terms are only for convenience of explanation and do not constitute any limitation on the present invention.

Claims

1. An AI-powered range hood and kitchen appliance linkage control system, comprising an AI range hood body (1), a linkage appliance module (2), an environmental sensing module (3), an AI decision-making module (4), a communication module (5), and a human-computer interaction module (6) for user operation, characterized in that, The linkage electrical module (2) communicates with the AI ​​decision module (4) through the communication module (5) to receive control instructions from the AI ​​decision module (4) connected to the module and control the corresponding electrical unit to link with the AI ​​range hood body (1); The environmental perception module (3) collects environmental data in real time and feeds the collected environmental data back to the AI ​​decision-making module (4); The AI ​​decision module (4) generates instructions based on the environmental data and sends them to the linkage appliance module (2) and the AI ​​range hood body (1).

2. The artificial intelligence-based range hood and kitchen appliance linkage control system according to claim 1, characterized in that, The AI ​​decision module (4) includes a data preprocessing unit, a scene recognition unit, and an instruction generation unit. The data preprocessing unit is used to perform noise reduction and normalization processing on the environmental data. The scene recognition unit identifies the current cooking type based on the processed environmental data and the preset cooking scene information. The instruction generation unit generates control instructions based on the current cooking type and sends them to the AI ​​range hood body (1) and the linkage appliance module (2).

3. The artificial intelligence-based range hood and kitchen appliance linkage control system according to claim 2, characterized in that, The AI ​​decision module (4) also includes a self-learning unit. The self-learning unit continuously collects and stores a dataset of the user's historical cooking operations, and optimizes the instruction generation logic based on the dataset. The logic is stored in the self-learning unit to update the cooking scene information of the scene recognition unit.

4. The artificial intelligence-based range hood and kitchen appliance linkage control system according to claim 2, characterized in that, The environmental data includes oil fume concentration, gas concentration, cooking area temperature and cooking actions. The environmental sensing module (3) collects the oil fume concentration every 50-120ms, the gas concentration every 50-80ms, and the cooking area temperature every 150-250ms, and captures the cooking actions at a frame rate of 20-40fps.

5. The artificial intelligence-based range hood and kitchen appliance linkage control system according to claim 4, characterized in that, The data preprocessing unit uses a Kalman filter algorithm to remove fluctuations in smoke concentration, gas concentration, and cooking zone temperature. Then, a normalization algorithm maps the smoke concentration, cooking zone temperature, and gas concentration to set values. Finally, a CNN+LSTM hybrid neural network model is used to identify the cooking action.

6. The artificial intelligence-based range hood and kitchen appliance linkage control system according to claim 2, characterized in that, The scene recognition unit has a built-in preset cooking scene model. The scene recognition unit identifies the current cooking type and outputs it to the instruction generation unit. At the same time, it judges the safety risk level based on the gas concentration and cooking area temperature data collected by the environmental perception module (3) and outputs the result to the human-computer interaction module (6).

7. The artificial intelligence-based range hood and kitchen appliance linkage control system according to claim 2, characterized in that, The instruction generation unit will receive the current cooking type and generate control instructions. The control instructions will be output to the AI ​​range hood body (1) and the linkage appliance module (2) through the communication module (5). The control instructions include equipment gear adjustment instructions and AI range hood body (1) cleaning instructions.

8. A method for linking and controlling an artificial intelligence range hood with kitchen appliances, comprising the artificial intelligence range hood and kitchen appliance linkage control system as described in any one of claims 1-7, characterized in that, The method includes: Step S1. Obtaining instructions: The human-computer interaction module (6) obtains the user's operation instructions; Step S2. System Start-up: The human-computer interaction module (6) converts the acquired operation instructions into system control signals and outputs the system control signals to the AI ​​decision module (4). The AI ​​decision module (4) outputs start-up instructions to the AI ​​range hood body (1), the linkage appliance module (2), the environmental perception module (3), the communication module (5), and the human-computer interaction module (6). Step S3. Data acquisition: The environmental perception module (3) collects data in real time and outputs the data to the AI ​​decision module (4). The collected data includes oil fume concentration data M1, gas concentration data M2, cooking area temperature data M3, and captured cooking action data M4. Step S4. Data processing: The AI ​​decision module (4) uses the Kalman filter algorithm to remove the fluctuating data of M1-M3 from the collected data, and then uses the normalization algorithm to map the data of oil fume concentration, cooking area temperature and gas concentration to the set value. Step S5. Scene Recognition: The scene recognition unit extracts the features of cooking actions through a CNN+LSTM hybrid neural network model and combines them with the processed cooking area temperature data M3 and oil fume concentration data M1 to obtain a data set. The data set is then compared and matched with a preset cooking scene model to obtain a matching result, and the matching result is output to the instruction generation unit. Step S6. Control Execution: The instruction generation unit generates targeted control instructions based on the matching results of the scene recognition unit, and outputs the control instructions to the AI ​​range hood body (1) and the linkage appliance module (2) through the communication module (5). The AI ​​range hood body (1) and the linkage appliance module (2) switch the corresponding gear according to the control instructions, and the control instructions are stored in the historical data of the self-learning unit. Step S7. Cooking end determination: After the set waiting time, the scene recognition unit detects no cooking action through the CNN+LSTM hybrid neural network model. At the same time, the environmental perception module (3) collects the temperature M3 of the cooking area and detects that the oil fume concentration M1 drops to the set value. The AI ​​range hood body (1) determines that cooking is over and generates a cooking end command to output to the AI ​​decision module (4). Step S8. Post-cooking processing and purification: The environmental sensing module (3) does not detect that the gas concentration data M2 exceeds the set value during the entire cooking process, and feeds the result back to the AI ​​decision module (4). After receiving the cooking end instruction, the AI ​​decision module (4) generates a purification instruction to the AI ​​range hood body (1) and the linkage appliance module (2). The AI ​​range hood body (1) runs at low speed for a set time according to the purification instruction and then automatically turns off the power. The linkage appliance module (2) turns on the corresponding air purification appliance and runs for a set time and then automatically turns off the power.

9. The method for linkage control of an artificial intelligence range hood and kitchen appliances according to claim 8, characterized in that, In step S7, the cooking zone temperature M3 is reduced to a set value of 90℃-95℃, the oil fume concentration M1 is reduced to a set value of 0.5mg / m³-0.9mg / m³, and the set values ​​are normalized.

10. The method for linkage control of an artificial intelligence range hood and kitchen appliances according to claim 8, characterized in that, In step S6, the self-learning unit stores the gear control commands of the AI ​​range hood body (1) and the linkage appliance module (2) that the current user has not manually adjusted and maps them to the cooking scene model.

Images