An automatic odor detection and air extraction device and a multi-modal intelligent ventilation control method

Abstract

The application discloses a kind of stench detection automatic exhaust device and multi-modal intelligent ventilation control method, including the sensing module and ventilation execution host of wireless connection;There is no electrical connection between sensing module and ventilation execution host;Sensing module includes built-in gas sensor, sensing end microprocessor, sensing end wireless communication module and power supply, for collecting environmental gas concentration data and sending;Ventilation execution host includes fan module, main control unit and mains input interface;Main control unit includes main controller, communication module and motor drive circuit;Communication module is used to receive the data sent by sensing module;Main controller is used to run artificial intelligence recognition algorithm, assess odor type and intensity, and generate control instructions according to the evaluation results;Motor drive circuit drives fan module to run according to control instructions.The present application solves the problem of low response lag and energy efficiency in odor control through wireless gas sensing and multi-mode switchable ventilation control.

Description

Technical Field

[0001] The present invention belongs to the technical field of indoor environmental monitoring and control, and particularly relates to an odor detection automatic exhaust device and a multi-modal intelligent ventilation control method. Background Technique

[0002] With the acceleration of the urbanization process and the increasingly strict environmental protection requirements, environmental odor pollution has become an important factor affecting the quality of life of residents. Odorous gases have complex components, mainly including ammonia, hydrogen sulfide, volatile organic compounds, and sulfur- and nitrogen-containing organic compounds, etc. Even at low concentrations, they can cause strong discomfort and pose potential hazards to the human respiratory and nervous systems.

[0003] Currently, for the treatment of odor pollution in places such as garbage stations, toilets, and kitchens, it mainly relies on manual inspection and then manually starting and stopping ventilation equipment, or adopting a crude mode of timed ventilation. Such methods lack real-time and accurate perception of pollution sources, resulting in two common problems: one is that the response is seriously lagged, and the treatment is carried out only after the odor has been generated and spread, affecting environmental comfort and health; the other is low energy efficiency, running regularly regardless of the actual air quality, or continuing to ventilate even after the pollution has been eliminated, causing significant energy waste. Although the online odor monitoring systems that have emerged in recent years can achieve concentration quantification and alarm, their monitoring units are usually installed independently and only provide data output. They cannot form an intelligent closed loop with the ventilation execution equipment and still require manual or upper computer intervention to be linked. The control link is long and the degree of automation is low, and the transformation from passive monitoring to active and on-demand treatment has not been fundamentally achieved. Summary of the Invention

[0004] Object of the Invention: In order to overcome the deficiencies in the prior art, the present invention provides an odor detection automatic exhaust device and a multi-modal intelligent ventilation control method, which solve the problems of lagged response and low energy efficiency in odor treatment through wireless gas sensing and multi-mode switchable ventilation control.

[0005] Technical Solution: To achieve the above object, an odor detection automatic exhaust device of the present invention includes a sensing module and a ventilation execution host connected wirelessly;

[0006] The sensing module is an independently arranged wireless sensing unit, and there is no electrical connection between it and the ventilation execution host; the sensing module includes an internal gas sensor, a sensing end microprocessor, a sensing end wireless communication module, and a power supply, and is used for collecting ambient gas concentration data and sending it;

[0007] The ventilation execution host includes a fan module, a main control unit, and a mains input interface;

[0008] The main control unit includes a main controller, a communication module, and a motor drive circuit; the communication module is used to receive data sent by the sensing module; the main controller is used to run an artificial intelligence recognition algorithm to evaluate the type and intensity of odors and generate control commands based on the evaluation results; the motor drive circuit drives the fan module to operate according to the control commands.

[0009] Furthermore, the artificial intelligence recognition algorithm running on the main controller is a lightweight model, which is used to infer the odor category and odor intensity level based on the feature vector extracted from the gas concentration time series.

[0010] Furthermore, the main controller is also configured to execute a multi-mode adaptive control strategy, which includes at least: a low-concentration steady-state maintenance mode, a sudden peak emergency ventilation mode, a combined pollution co-treatment mode, and an energy-saving standby mode, and can intelligently switch between different modes based on real-time sensor data and historical trends.

[0011] Furthermore, the execution of the multi-mode adaptive control strategy includes:

[0012] When a single gas component is identified as slightly exceeding the standard and the concentration change is gradual, the energy-saving standby mode is activated to control the fan to run intermittently at low speed.

[0013] When a sudden increase in gas concentration is detected in a very short period of time, switch to the emergency ventilation mode for sudden peak, control the fan to run at full speed and trigger an alarm.

[0014] When multiple gas components are identified as exceeding the standard simultaneously, the composite pollution collaborative treatment mode is activated, and the exhaust parameters are dynamically adjusted based on comprehensive environmental information.

[0015] Furthermore, the sensing module also includes a threaded protective shell, inside which the gas sensor is placed; the threaded protective shell is provided with a threaded mounting head and a micro vent hole; the threaded mounting head is used to fix the sensing module to the odor monitoring point.

[0016] Furthermore, the sensing module is powered by a button battery; the sensing terminal wireless communication module uses a Bluetooth Low Energy protocol chip.

[0017] Furthermore, the fan module includes a fan motor, an air inlet grille, an air inlet duct, a centrifugal impeller, a volute duct, and an air outlet. The fan motor drives the centrifugal impeller connected within the volute duct. The air inlet duct and the air outlet are respectively connected to the air inlet side and the air outlet side of the volute duct. The air inlet grille is installed on the air inlet duct. Gas enters through the air inlet grille, and the centrifugal impeller is driven to rotate by the fan motor to form an airflow, which is then discharged through the air outlet of the volute duct.

[0018] Furthermore, the fan motor is a single-phase AC capacitor-run motor or a DC brushless motor; the motor drive circuit is a power drive and speed control circuit corresponding to the type of fan motor, and receives PWM speed control signals or switching commands issued by the main controller.

[0019] Furthermore, the communication module includes a first communication unit and a second communication unit; the first communication unit is a low-power wireless receiving module that matches the sensing end wireless communication module; the second communication unit is a Wi-Fi module used for data interaction with a cloud platform or mobile terminal.

[0020] A multimodal intelligent ventilation control method for an automatic odor detection ventilation device, executed by the main controller, includes the following steps:

[0021] Step S1: Acquire and preprocess gas concentration data from the sensing module;

[0022] Step S2: Analyze the preprocessed data based on artificial intelligence recognition algorithms and output the evaluation results of odor category and odor intensity level;

[0023] Step S3: Based on the evaluation results and context information, query the preset strategy table to generate a control strategy that includes the target turbine speed and running time parameters;

[0024] Step S4: Drive the fan module to perform the corresponding ventilation action according to the control strategy, and record the event information.

[0025] In step S3, the strategy table defines the optimal ventilation strategy parameters corresponding to different combinations of odor categories, odor intensity levels, and time information.

[0026] It also includes step S5: uploading the recorded event information to the cloud server via the Wi-Fi module and receiving the model update file from the cloud server to complete the iterative optimization of the artificial intelligence recognition algorithm.

[0027] Beneficial Effects: This invention, through a separate architecture of a wireless independent sensing module and an intelligent ventilation host, and by incorporating a lightweight AI odor recognition model and multi-mode control strategies, achieves a millisecond-level rapid closed-loop response from odor perception to ventilation execution, significantly improving the timeliness of odor removal. Furthermore, the system can intelligently select the optimal ventilation mode and parameters based on the accurately identified odor type, intensity, and trend, truly achieving on-demand ventilation while minimizing ineffective energy consumption. The sensing module requires no wiring and can be flexibly deployed at the optimal detection point; the host installation is no different from that of a traditional exhaust fan, greatly simplifying the installation and modification process of the intelligent ventilation system and lowering the deployment threshold and cost. Attached Figure Description

[0028] Figure 1 Schematic diagram of the overall structure of the automatic exhaust fan for odor detection Figure 1 ;

[0029] Figure 2 Schematic diagram of the overall structure of the automatic exhaust fan for odor detection Figure 2 ;

[0030] Figure 3 Schematic diagram of the overall structure of the automatic exhaust fan for odor detection Figure 3 ;

[0031] Figure 4 This is a schematic diagram of a half-section of the sensing module;

[0032] Figure 5 This is a flowchart of the odor recognition and decision-making process of the present invention. Detailed Implementation

[0033] The invention will now be further described with reference to the accompanying drawings.

[0034] like Figure 1 , Figure 2 , Figure 3 as well as Figure 4 As shown, an automatic ventilation device for odor detection includes a wirelessly connected sensing module 1 and a ventilation actuator. The sensing module 1 is an independently deployable wireless sensing unit, and it is not electrically connected to the ventilation actuator. The sensing module 1 includes a built-in gas sensor 11, a sensing-end microprocessor 13, a sensing-end wireless communication module 14, and a power supply, used to collect and transmit ambient gas concentration data. The ventilation actuator includes a fan module 2, a main control unit 3, and a mains power input interface 4. The main control unit 3 includes a main controller 31, a communication module 32, and a motor drive circuit 33. The communication module 32 is used to receive data transmitted by the sensing module 1; the main controller 31 is used to run an artificial intelligence recognition algorithm to evaluate the odor type and intensity, and generate control commands based on the evaluation results; the motor drive circuit 33 drives the fan module 2 to operate according to the control commands. The odor monitoring sensing module 1 is an independent, arbitrarily deployable front-end sensing node. Leveraging its self-powered and wireless communication capabilities, this module can be installed at the source or key monitoring points along the odor diffusion path, such as directly above a toilet, near a stove, or at garbage dumps, to promptly capture signals of changes in original odor concentration. The ventilation actuator, through its communication module 32, aggregates wireless data streams from one or more sensing modules 1 in real time. The main controller 31 then centrally processes and analyzes the received data and directly drives the fan module 2, thus combining highly flexible distributed sensing with highly reliable centralized control to ensure rapid response and accurate control.

[0035] The main controller 31 runs a lightweight AI recognition algorithm, which is pre-trained and deployed within the controller. This model infers the odor category and intensity level based on feature vectors extracted from the gas concentration time series. Odor pattern recognition is performed by the main controller 31. Traditional solutions typically only perform simple threshold comparisons, resulting in limited recognition capabilities and high false alarm rates. Alternatively, raw data can be uploaded to the cloud for analysis, but this suffers from network latency and poor stability. In this invention, the main controller 31 first preprocesses the received raw concentration data and extracts multi-dimensional feature vectors from a data sequence within a time window. These features include not only instantaneous concentration values ​​but also dynamic information such as change rate, fluctuation pattern, and duration. Subsequently, this feature vector is input into the pre-set lightweight model for forward inference. The model directly outputs semantic recognition results, such as "kitchen waste odor, level 3," through its internally learned nonlinear mapping relationships, achieving millisecond-level real-time and rapid edge-side recognition.

[0036] More specifically, in this invention, the artificial intelligence recognition algorithm is constructed and deployed in the following ways:

[0037] I. Model Training: A large amount of time-series gas concentration data, labeled with corresponding odor categories and intensity levels, is collected from gas sensors 11 in different scenarios (such as restrooms, kitchens, and garbage stations) on the cloud or high-performance computing devices as a training set. Using this training set, a selected machine learning model (such as a Convolutional Neural Network (CNN), a Recurrent Neural Network (RNN), a Long Short-Term Memory Network (LSTM), or a Gradient Boosting Decision Tree) is trained under supervision to learn the mapping relationship between the dynamic features of sensor data and odor discrimination.

[0038] II. Model Lightweighting: The trained model is processed using one or more lightweighting techniques such as knowledge distillation, parameter quantization, network pruning, or structured pruning to generate a lightweight model suitable for the computing resources and storage space of the embedded main controller 31.

[0039] 3. Model Deployment: Convert the lightweight model into a format supported by the embedded platform and burn it into the memory of the main controller 31 to form a built-in artificial intelligence recognition algorithm.

[0040] IV. Online Inference: When the device is working, the main controller 31 calls the model, inputs the feature vector extracted in real time, the model performs forward propagation calculation, and finally outputs the odor category and odor intensity level.

[0041] The main controller 31 is also configured to execute a multi-mode adaptive control strategy, which includes at least: a low-concentration steady-state maintenance mode, a sudden peak emergency exhaust mode, a complex pollution co-processing mode, and an energy-saving standby mode. It can intelligently switch between different modes based on real-time sensor data and historical trends. The main controller 31 of this invention uses the odor category and intensity level identified by AI as the core input, and combines it with contextual information such as timestamps and historical equipment operation records to construct a dynamic environmental state profile, thereby selecting an appropriate mode. For example, the low-concentration steady-state maintenance mode is suitable for scenarios where background odors accumulate slowly, employing low-speed continuous or intermittent operation; the sudden peak emergency exhaust mode is for dangerous scenarios where concentrations suddenly surge, immediately initiating the highest-speed, powerful exhaust; the complex pollution co-processing mode is for complex scenarios with multiple mixed odors, potentially employing a combination strategy of variable wind speed and extended operating time; and the energy-saving standby mode shuts down the main power consumption units when there is no pollution for an extended period. This achieves dynamic switching between different modes based on real-time data streams and short-term historical trends (such as whether the concentration is continuously rising or falling). Therefore, the system no longer blindly responds with a fixed power, but selects ventilation strategies based on the nature and degree of pollution, thereby minimizing ineffective ventilation and energy waste while ensuring that air quality meets standards, and optimizing equipment operating noise and mechanical wear.

[0042] More specifically, the execution of the multi-mode adaptive control strategy includes: when a single gas component is identified as slightly exceeding the standard and its concentration changes gradually, the energy-saving standby mode is activated, controlling the fan to operate intermittently at low speed; when a gas concentration is identified as rising sharply in a very short time, the system switches to the emergency exhaust mode for sudden peaks, controlling the fan to run at full speed and triggering an alarm; when multiple gas components are identified as exceeding the standard simultaneously, the composite pollution synergistic treatment mode is activated, dynamically adjusting exhaust parameters based on comprehensive environmental information. Wherein:

[0043] 1) For minor, slow-release pollution: The essence of activating the energy-saving standby mode is to use duty cycle control. The fan is in a stopped or very low-speed monitoring state most of the time, only starting to run at low speed for short periods of time periodically to maintain a slight negative pressure in the space and prevent odors from slowly accumulating and spreading. Furthermore, the start-stop cycle and low-speed setting are optimized to achieve a balance between preventing diffusion and saving energy.

[0044] 2) For sudden, high-risk pollution: The key to activating the emergency ventilation mode during peak periods lies in real-time monitoring and exceeding limits of the concentration change rate. Once a sharp increase in concentration is detected within a timescale of hundreds of milliseconds to seconds, exceeding the safety threshold, it is determined to be an emergency event such as a leak or fire. Immediately, any current mode is interrupted, and the motor drive circuit 33 sends a maximum duty cycle PWM signal or a direct high-voltage connection command to the fan motor 21, accelerating it to its rated maximum speed in the shortest possible time to achieve maximum ventilation volume, enabling rapid dilution and removal of hazardous substances, and simultaneously triggering an alarm.

[0045] 3) For complex and persistent pollution: The combined pollution co-treatment mode is relatively complex and is activated when multiple gaseous components in the environment simultaneously exceed standards and the pollution is expected to last for a long period. This mode integrates current environmental information from multiple sources, such as temperature, humidity, and wind speed, to dynamically calculate and adjust the required exhaust intensity and duration to avoid pollutant accumulation in local areas or secondary pollution. For example, a higher wind speed is used at the beginning of exhaust to quickly reduce the peak concentration, and then the wind speed is reduced to maintain the purification effect and reduce energy consumption, ensuring that pollutants with different volatility characteristics can be effectively removed, avoiding over-ventilation or incomplete purification.

[0046] like Figure 4 As shown, the sensing module 1 also includes a threaded protective shell 12, inside which the gas sensor 11 is housed. The threaded protective shell 12 is provided with a threaded mounting head 16 and a micro-vent hole 17. The threaded mounting head 16 is used to fix the sensing module 1 to the odor monitoring point. The cylindrical threaded protective shell 12 provides protection for the internal precision gas sensor 11, preventing damage caused by impacts, foreign object intrusion, or human contact. The threaded mounting head 16 allows the sensing module 1 to be directly screwed into a pre-drilled standard threaded hole or matching mounting base on a wall, ceiling, or equipment casing. The micro-vent hole 17 ensures that external air can diffuse to the sensor's sensing surface at a sufficient rate, ensuring real-time detection, while also being small enough to effectively block most dust, fibers, and other contaminants from intruding, providing excellent protection.

[0047] In this invention, preferably, the power source for the sensing module 1 is a button battery 15; and the sensing terminal wireless communication module 14 adopts a Bluetooth Low Energy protocol chip.

[0048] like Figure 1 , Figure 2 as well as Figure 3As shown, the fan module 2 includes a fan motor 21, an air inlet grille 22, an air inlet duct 23, a centrifugal impeller 24, a volute duct 25, and an air outlet 26. The fan motor 21 drives the centrifugal impeller 24 connected within the volute duct 25. The air inlet duct 23 and the air outlet 26 are respectively connected to the air inlet side and the air outlet side of the volute duct 25. The air inlet grille 22 is installed on the air inlet duct 23. Gas enters through the air inlet grille 22, and the centrifugal impeller 24 is driven by the fan motor 21 to rotate, forming an airflow, which is then discharged through the air outlet 26 of the volute duct 25. The gas passes through the air inlet duct 23, where the centrifugal impeller 24 performs work on it, giving it kinetic and pressure energy. Then, through the gradually expanding cross-section volute-shaped casing, i.e., the volute duct 25, the high-speed air thrown out by the impeller is collected in an orderly manner, converting its function into static pressure to form a stable airflow, which is then directionally discharged to the outside of the duct from the air outlet of the pre-installed exhaust duct in the building.

[0049] The fan motor 21 is a single-phase AC capacitor-run motor or a DC brushless motor; the motor drive circuit 33 is a power drive and speed control circuit corresponding to the type of the fan motor 21, and receives PWM speed control signals or switching commands issued by the main controller 31.

[0050] The communication module 32 includes a first communication unit and a second communication unit; the first communication unit is a low-power wireless receiving module that matches the sensing end wireless communication module 14; the second communication unit is a Wi-Fi module used for data interaction with a cloud platform or mobile terminal.

[0051] like Figure 5 As shown, a multimodal intelligent ventilation control method for an automatic odor detection ventilation device, executed by the main controller 31, includes the following steps:

[0052] Step S1: Acquire and preprocess gas concentration data from sensing module 1.

[0053] Step S2: Analyze the preprocessed data based on artificial intelligence recognition algorithms and output the evaluation results of odor category and odor intensity level.

[0054] Step S3: Based on the evaluation results and context information, query the preset strategy table to generate a control strategy that includes the target turbine speed and running time parameters.

[0055] Furthermore, in step S3, the strategy table defines the optimal ventilation strategy parameters corresponding to different combinations of odor categories, odor intensity levels, and time information. For ease of understanding, the following table lists strategy examples for several typical scenarios:

[0056] Table 1: Examples of Strategy Representations

[0057]

[0058] As shown in Table 1:

[0059] When the AI ​​model of the main controller 31 identifies the current situation as "cooking fumes, intensity level 3" and the time as afternoon, the system will query the strategy table and match Example 3. Subsequently, the main controller 31 will generate control instructions containing parameters such as "target gear: medium-low speed" and "operating mode: continuous operation until conditions are met", and send them to the motor drive circuit 33 to drive the fan into a stable ventilation state.

[0060] If a burning event subsequently occurs, and the concentration spikes to level 5 within seconds, the model identification result changes to "burnt odor, intensity level 5". The system will immediately interrupt the current strategy, seamlessly switch to and match the strategy in Example 4, control the fan to run at full speed and trigger an alarm, thus achieving an intelligent and rapid switch from normal maintenance to emergency handling.

[0061] Furthermore, the initial strategy parameters of the strategy table can be preset based on a large amount of historical operating data and expert experience, and the strategy parameters can be optimized and remotely updated after data analysis and model iteration through the cloud platform.

[0062] Step S4: Drive the fan module 2 to perform the corresponding ventilation action according to the control strategy, and record the event information.

[0063] The multimodal intelligent ventilation control method of the present invention further includes step S5: uploading the recorded event information to the cloud server via the Wi-Fi module, and receiving the model update file from the cloud server to complete the iterative optimization of the artificial intelligence recognition algorithm.

[0064] This invention, by employing a separate architecture of a wireless independent sensing module and an intelligent ventilation host, and incorporating a lightweight AI odor recognition model and multi-mode control strategies, offers the following advantages:

[0065] First, it achieves a millisecond-level rapid closed-loop response from odor detection to ventilation execution, significantly improving the timeliness of odor removal. Second, the system can intelligently select the optimal ventilation mode and parameters based on the accurately identified odor type, intensity, and trend, truly achieving on-demand ventilation while minimizing unnecessary energy consumption and ensuring purification effectiveness. Finally, the sensing module requires no wiring and can be flexibly deployed at the optimal detection point. The main unit installation is no different from that of a traditional exhaust fan, greatly simplifying the installation and modification process of the intelligent ventilation system and lowering the deployment threshold and cost.

[0066] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An automatic exhaust device for odor detection, characterized in that: Includes a wirelessly connected sensing module (1) and a ventilation actuator; The sensing module (1) is an independently deployable wireless sensing unit, and it has no electrical connection with the ventilation actuator host; the sensing module (1) includes a built-in gas sensor (11), a sensing end microprocessor (13), a sensing end wireless communication module (14) and a power supply, used to collect and transmit ambient gas concentration data. The ventilation actuator includes a fan module (2), a main control unit (3), and a mains power input interface (4). The main control unit (3) includes a main controller (31), a communication module (32), and a motor drive circuit (33); the communication module (32) is used to receive data sent by the sensing module (1); the main controller (31) is used to run an artificial intelligence recognition algorithm to evaluate the type and intensity of odors and generate control commands based on the evaluation results; the motor drive circuit (33) drives the fan module (2) to run according to the control commands.

2. The automatic exhaust device for odor detection according to claim 1, characterized in that: The main controller (31) runs an artificial intelligence recognition algorithm, which is a lightweight model, used to infer the odor category and odor intensity level based on the feature vector extracted from the gas concentration time series.

3. The automatic exhaust device for odor detection according to claim 2, characterized in that: The main controller (31) is also configured to execute a multi-mode adaptive control strategy, which includes at least: a low-concentration steady-state maintenance mode, a sudden peak emergency ventilation mode, a composite pollution synergistic treatment mode, and an energy-saving standby mode, and can intelligently switch between different modes based on real-time sensor data and historical trends.

4. The automatic exhaust device for odor detection according to claim 3, characterized in that: The execution of the multi-mode adaptive control strategy includes: When a single gas component is identified as slightly exceeding the standard and the concentration change is gradual, the energy-saving standby mode is activated to control the fan to run intermittently at low speed. When a sudden increase in gas concentration is detected in a very short period of time, switch to the emergency ventilation mode for sudden peak, control the fan to run at full speed and trigger an alarm. When multiple gas components are identified as exceeding the standard simultaneously, the composite pollution collaborative treatment mode is activated, and the exhaust parameters are dynamically adjusted based on comprehensive environmental information.

5. The automatic exhaust device for odor detection according to claim 1, characterized in that: The sensing module (1) also includes a threaded protective shell (12) in which the gas sensor (11) is placed; the threaded protective shell (12) is provided with a threaded mounting head (16) and a micro vent hole (17); the threaded mounting head (16) is used to fix the sensing module (1) to the odor monitoring point.

6. The automatic exhaust device for odor detection according to claim 1, characterized in that: The power source for the sensing module (1) is a button battery (15); the sensing terminal wireless communication module (14) uses a Bluetooth Low Energy protocol chip.

7. The automatic exhaust device for odor detection according to claim 1, characterized in that: The fan module (2) includes a fan motor (21), an air inlet grille (22), an air inlet duct (23), a centrifugal impeller (24), a volute duct (25), and an air outlet (26). The fan motor (21) drives the centrifugal impeller (24) connected to the volute duct (25). The air inlet duct (23) and the air outlet (26) are respectively connected to the air inlet side and the air outlet side of the volute duct (25). The air inlet grille (22) is installed on the air inlet duct (23). Gas enters through the air inlet grille (22), and the centrifugal impeller (24) is driven by the fan motor (21) to rotate to form an airflow, which is then discharged through the air outlet (26) of the volute duct (25).

8. An automatic exhaust device for odor detection according to claim 7, characterized in that: The fan motor (21) is a single-phase AC capacitor-run motor or a DC brushless motor; the motor drive circuit (33) is a power drive and speed regulation circuit corresponding to the type of the fan motor (21), and receives PWM speed regulation signals or switching commands issued by the main controller (31).

9. An automatic exhaust device for odor detection according to claim 1, characterized in that: The communication module (32) includes a first communication unit and a second communication unit; the first communication unit is a low-power wireless receiving module that matches the sensing end wireless communication module (14); the second communication unit is a Wi-Fi module used for data interaction with a cloud platform or mobile terminal.

10. A multimodal intelligent ventilation control method for an automatic odor detection ventilation device according to any one of claims 1-9, characterized in that: Executed by the main controller (31), the following steps are included: Step S1: Acquire and preprocess gas concentration data from sensing module (1); Step S2: Analyze the preprocessed data based on artificial intelligence recognition algorithms and output the evaluation results of odor category and odor intensity level; Step S3: Based on the evaluation results and context information, query the preset strategy table to generate a control strategy that includes the target turbine speed and running time parameters; Step S4: Drive the fan module (2) to perform the corresponding ventilation action according to the control strategy and record the event information.

11. The multimodal intelligent ventilation control method for an automatic odor detection exhaust device according to claim 10, characterized in that: In step S3, the strategy table defines the optimal ventilation strategy parameters corresponding to different combinations of odor categories, odor intensity levels, and time information.

12. The multimodal intelligent ventilation control method for an automatic odor detection exhaust device according to claim 10, characterized in that: It also includes step S5: uploading the recorded event information to the cloud server via the Wi-Fi module and receiving the model update file from the cloud server to complete the iterative optimization of the artificial intelligence recognition algorithm.

Images