An off-flavor walking monitoring device based on an intelligent sensing chip

Abstract

The application provides a kind of based on smart sensing chip's smell walk monitoring equipment, install on mobile carrier, including: gas collection module, for real-time collection ambient air;Smart sensing chip array module, for real-time detection to ambient air obtains multidimensional electrical signal;Information acquisition and processing module, including: feature extraction unit, for extracting multidimensional feature vector from multidimensional electrical signal;Gas identification unit, for identifying the kind of odor gas in current environment according to multidimensional feature vector;Concentration inversion unit, for inverting the gas concentration corresponding to each odor gas species according to multidimensional feature vector;Odor concentration calculation unit, for calculating comprehensive odor concentration OU value based on gas concentration traversing olfactory threshold database;Display module, for real-time display comprehensive odor concentration OU value;Power module.The beneficial effect is that the application can meet the actual needs of regional foul odor smell rapid inspection, pollution degree classification and over-standard early warning.

Description

Technical Field

[0001] This invention relates to the technical field of environmental monitoring equipment, and more specifically, to a mobile odor monitoring device based on an intelligent sensor chip. Background Technology

[0002] Odor pollution is one of the main environmental complaints from urban residents, and it is widespread in industrial parks, landfills, sewage treatment plants, livestock farms, and other places. Odor gases are characterized by dispersed sources, complex composition, large concentration fluctuations, and sudden onset, which can irritate the human olfactory system and seriously affect the quality of life of residents and social stability.

[0003] Traditional odor monitoring mainly relies on manual sampling combined with laboratory analysis methods such as gas chromatography-mass spectrometry (GC-MS). Although these methods have high detection accuracy, they have drawbacks such as poor real-time performance, inability to cover a large area, long sampling cycle, and difficulty in rapid source tracing. They cannot meet the emergency monitoring needs of sudden odor pollution events and the workload of large-area patrols.

[0004] In recent years, with the development of sensor technology, portable gas detectors and sensor microstations have been applied to some extent in the field of environmental monitoring. However, existing equipment is mostly used for the quantitative detection of conventional air pollutants (such as CO, NO2, SO2) or single odor gases (such as H2S, NH3), and the output results are mostly raw sensor response values ​​or single component concentration data. For complex mixed odor gases, these devices have the following shortcomings: First, the cross-sensitivity effect between sensors leads to a decrease in detection accuracy when multiple components are present simultaneously; second, the output data of the equipment (such as resistance change value, ppm concentration) cannot intuitively reflect the actual intensity of odor stimulation to human olfactory senses, making it difficult for environmental managers to judge the severity of odor pollution and its actual impact on residents.

[0005] To address these issues, some studies have attempted to introduce electronic nose technology, using the combined response of multiple gas-sensitive units to simulate human olfaction and estimate odor intensity. However, existing electronic nose devices typically only collect DC resistance changes from sensors, providing limited information and difficulty in distinguishing complex gas mixtures. Furthermore, their outputs are mostly combinations of sensor response values, lacking a comprehensive quantitative indicator that reflects the overall intensity of the mixed odor and aligns with human olfactory perception. In addition, existing devices often employ a fragmented data processing design, with gas identification and concentration calculation occurring independently, hindering end-to-end automated processing from raw signals to odor intensity indicators. This fails to meet the real-time, intelligent, and comprehensive quantitative evaluation requirements of mobile monitoring.

[0006] Therefore, there is an urgent need for a mobile odor monitoring device that can adapt to complex and changing environments, accurately identify multi-component odor gases, convert the detection results into a comprehensive odor intensity index that conforms to human olfactory perception, and has automated processing capabilities, in order to meet the actual needs of rapid regional odor inspection, pollution level classification, and early warning of exceeding standards. Summary of the Invention

[0007] The technical problem to be solved by this invention is how to meet the actual needs of rapid inspection of regional malodorous odors, pollution level classification and early warning of exceeding standards. In order to overcome the defects of the above-mentioned existing technologies (or related technologies), this invention provides a mobile odor monitoring device based on an intelligent sensor chip.

[0008] This invention provides a mobile odor monitoring device based on a smart sensor chip. The mobile odor monitoring device is installed on a mobile carrier and includes: A gas acquisition module is used to collect ambient air data in real time. A smart sensor chip array module, connected to the gas acquisition module, includes multiple odor gas sensitive units, used to detect the ambient air in real time and obtain corresponding multidimensional electrical signals. An information acquisition and processing module, connected to the intelligent sensor chip array module, includes: A feature extraction unit is used to extract a multidimensional feature vector from the multidimensional electrical signal; A gas identification unit, connected to the feature extraction unit, has a built-in machine learning-based gas classification model for identifying the types of odorous gases in the current environment based on the multi-dimensional feature vector. A concentration inversion unit is connected to the gas identification unit and has a built-in machine learning-based concentration inversion model for inverting the gas concentration corresponding to each type of odor gas based on the multi-dimensional feature vector. An odor concentration calculation unit is connected to the concentration inversion unit and has a built-in odor threshold database. It is used to calculate the comprehensive odor concentration OU value by traversing the odor threshold database based on the gas concentration. A display module, connected to the information acquisition and processing module, is used to display the comprehensive odor concentration (OU) value in real time; A power supply module is connected to the gas acquisition module, the intelligent sensor chip array module, the information acquisition and processing module, and the display module, respectively.

[0009] Compared with existing technologies, the odor mobile monitoring device based on an intelligent sensor chip of this invention has the following advantages: This invention integrates a gas acquisition module, an intelligent sensor chip array module, an information acquisition and processing module, a display module, and a power supply module onto a mobile carrier, constructing a complete mobile odor monitoring device. This achieves fully automated processing from ambient air acquisition, multi-dimensional electrical signal sensing, feature extraction, gas type identification, concentration inversion, to comprehensive odor intensity quantification. Through a built-in machine learning model and olfactory threshold database, the original multi-dimensional electrical signals from the odor gas sensing unit are converted into a comprehensive odor concentration (OU) value that conforms to human olfactory perception, enabling pollution level classification and exceeding warnings. This solves the technical problem of traditional electronic nose devices that only output sensor response values ​​and cannot intuitively reflect the actual impact of odor pollution on the human body. Furthermore, the mobile carrier configuration gives this mobile odor monitoring device mobile monitoring capabilities, enabling rapid inspection of large areas with foul odors.

[0010] In one possible implementation, each of the odor gas sensing units is integrated onto a single chip using MEMS technology, and the size of the chip is less than 5mm x 5mm.

[0011] Compared with existing technologies, the above-mentioned technical solution can achieve high integration and miniaturization of odor gas sensing units through microelectromechanical systems (MEMS) technology, reducing the overall size and power consumption of mobile odor monitoring equipment and facilitating integration and installation on various mobile carriers. At the same time, the small size design helps to improve the response speed and sensitivity of odor gas sensing units, meeting the rapid response requirements of mobile odor monitoring equipment.

[0012] In one possible implementation, the multidimensional feature vector extracted by the feature extraction unit includes steady-state features, dynamic features, and integral features. The steady-state features include the change in resistance value when the response reaches equilibrium, impedance modulus, real part of impedance, imaginary part of impedance, and phase angle. The dynamic features include the response rise slope, recovery slope, response time, and recovery time. The integral feature is the area under the response curve.

[0013] Compared with existing technologies, the above-mentioned technical solution can fully explore the rich information generated during the interaction between the odor gas sensitive unit and the gas by extracting multiple dimensions and types of features of resistance value and impedance signal. This provides more discriminative input data for subsequent machine learning models, which helps to improve the accuracy of odor gas type identification and gas concentration inversion. It also overcomes the problems of insufficient information utilization and poor anti-interference ability caused by traditional methods that rely only on a single steady-state response value.

[0014] In one possible implementation, the gas classification model is selected from one of the following: support vector machine, random forest, shallow neural network, and CNN classification model.

[0015] In one possible implementation, the concentration inversion model is selected from one of the following: BP neural network, support vector regression, extreme learning machine, and LSTM regression model.

[0016] Compared with existing technologies, the above-mentioned technical solution can identify mature machine learning algorithms that can be used, which not only ensures the feasibility of the solution, but also provides flexibility for selecting the optimal algorithm for different application scenarios. Algorithms such as support vector machines and neural networks have good nonlinear mapping capabilities and generalization performance, which can effectively handle the nonlinear relationship between the response of odor gas sensitive units and the gas type and concentration in complex mixed gas environments, and realize the accurate identification and quantitative analysis of multi-component odor gases.

[0017] In one possible implementation, the odor concentration calculation unit obtains the comprehensive odor concentration OU value using the following formula: in, This indicates the combined odor concentration (OU) value; Indicates the identified first The gas concentration corresponding to the odorous gas, ; This indicates the number of pre-stored olfaction thresholds in the olfaction threshold database. The olfaction threshold corresponding to each odorous gas .

[0018] In one possible implementation, a positioning and navigation module is further included, connected to the information acquisition and processing module, for acquiring the geographic location information of the mobile carrier in real time. The information acquisition and processing module also includes a data merging unit, connected to the odor concentration calculation unit, for time alignment and delay compensation of the comprehensive odor concentration OU value and the geographic location information, generating odor monitoring data with geographic tags and sending it to the display module for display.

[0019] Compared with existing technologies, the above-mentioned technical solution can achieve accurate integration of comprehensive odor concentration (OU) value and geographical location information, overcome the positioning deviation caused by the response delay of odor gas sensitive units, and ensure that each odor monitoring data has accurate geographical coordinates, laying the foundation for subsequent spatial distribution heat map mapping and source tracing analysis of odor pollution.

[0020] In one possible implementation, the information acquisition and processing module further includes a spatial interpolation unit connected to the data merging unit, used to perform gridded interpolation on the discrete odor monitoring data using an inverse distance weighted interpolation algorithm, generate a spatial distribution heat map of odor pollution, and send it to the display module for display.

[0021] Compared with existing technologies, the above-mentioned technical solution can transform discrete point data collected along the mobile route into a continuous spatial distribution heat map of odor pollution, which intuitively shows the diffusion range, concentration gradient and high value areas of odor pollution. This helps environmental managers quickly identify pollution hotspots and assess the scope of pollution impact, providing intuitive visualization basis for accurate source tracing and governance decisions. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the structure of the present invention; Figure 2 This is a schematic diagram of the structure of the mobile carrier of the present invention; Figure 3 This is a schematic diagram of the top structure of the odor mobile monitoring device of the present invention; Figure 4 This is a schematic diagram of the bottom structure of the odor mobile monitoring device of the present invention; Figure 5 This is a schematic diagram of the structure of the intelligent sensor chip array module of the present invention; Explanation of the numerical markings in the attached diagram: 1. Mobile carrier; 2. Gas acquisition module; 21. Air inlet; 22. Air pump; 23. Filter unit; 24. Gas path switching valve; 3. Intelligent sensor chip array module; 4. Information acquisition and processing module; 41. Feature extraction unit; 42. Gas identification unit; 43. Concentration inversion unit; 44. Odor concentration calculation unit; 45. Data merging unit; 46. Spatial interpolation unit; 5. Display module; 6. Power supply module; 7. Positioning and navigation module; 8. Data storage and transmission module. Detailed Implementation

[0023] First, those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention. Those skilled in the art can make adjustments as needed to adapt to specific application scenarios.

[0024] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0025] See Figure 1 and Figure 2This invention discloses a mobile odor monitoring device based on an intelligent sensor chip. The mobile odor monitoring device is installed on a mobile carrier 1, which can be an electric vehicle, a drone, or a robot, to achieve mobile monitoring on the ground or in the air. The mobile odor monitoring device includes a gas collection module 2, an intelligent sensor chip array module 3, an information collection and processing module 4, a display module 5, a power supply module 6, and optional positioning and navigation modules 7 and data storage and transmission modules 8.

[0026] See Figure 3 and Figure 4 In this embodiment of the invention, the gas acquisition module 2 is used to collect ambient air in real time and deliver it to the intelligent sensor chip array module 3. The gas acquisition module 2 specifically includes an air inlet 21, an air pump 22, a filter unit 23, and a gas path switching valve 24. The air inlet 21 is exposed to the environment, the air pump 22 provides power to draw in ambient air, the filter unit 23 is used to remove particulate matter (such as dust and aerosol) in the ambient air to avoid contaminating the odor gas sensitive unit, and the gas path switching valve 24 can be controlled to directly introduce ambient air into the detection chamber or switch to a bypass for cleaning, calibration, and other operations. During the navigation process, the gas acquisition module 2 continuously samples at a constant flow rate (e.g., 0.5 L / min) to ensure real-time response.

[0027] See Figure 5 In this embodiment of the invention, the intelligent sensing chip array module 3 is connected to the gas acquisition module 2 and includes multiple integrated odor gas sensing units. In this embodiment, each odor gas sensing unit is integrated on a single chip using MEMS (Micro-Electro-Mechanical Systems) technology. The size of a single chip is less than 5mm × 5mm. Each chip contains no less than 8 independent odor gas sensing units. Each odor gas sensing unit can be optimized for specific odor gases or gas families (such as H2S, NH3, methanethiol, trimethylamine, VOCs, etc.) by material optimization (such as metal oxide semiconductors or nanomaterials doped with different catalysts). Each odor gas sensing unit is integrated with a micro heater below it, and the operating temperature can be independently adjusted by the information acquisition and processing module 4 (e.g., from room temperature to 400℃) to enhance the selective response to different odor gases.

[0028] In this embodiment of the invention, each odor gas sensing unit includes a substrate and a central electrode, an output voltage electrode, a heating voltage electrode, a ground electrode, and a source voltage electrode disposed on the top of the substrate. The output voltage electrode, heating voltage electrode, ground electrode, and source voltage electrode are distributed around the central electrode. The central electrode includes a gas-sensitive thin film, interdigitated electrodes, an insulating layer, a heating resistor, and a support layer stacked sequentially from top to bottom.

[0029] In this embodiment of the invention, the information acquisition and processing module 4 is the core of the odor mobile monitoring device. It is electrically connected to the intelligent sensor chip array module 3 and is used to acquire and process the multi-dimensional electrical signals of the odor gas sensing units. This information acquisition and processing module 4 includes a constant temperature control circuit, a multi-channel signal acquisition circuit, and a microprocessor. The constant temperature control circuit stabilizes the operating temperature of the odor gas sensing units, reducing the impact of ambient temperature changes on the response. The multi-channel signal acquisition circuit has a built-in AC / DC impedance multi-parameter analysis model, which can simultaneously obtain the DC resistance of each odor gas sensing unit. R) and wideband AC impedance parameters; specifically, the multi-channel signal acquisition circuit includes an impedance signal acquisition chip, which can be selected from models such as AD5933, AD5934, AD5940, AD5941, ADuCM355, etc. It can apply a frequency-adjustable sine wave excitation and measure the real and imaginary parts of the response signal, thereby calculating the impedance modulus and phase angle. The scanning frequency range can be set to any sub-range between 0.1Hz and 1MHz, such as 0.1–10kHz or 1–100kHz, to cover the response characteristic frequency bands of different odor gases.

[0030] In this embodiment of the invention, the microprocessor has multiple built-in functional units, including a feature extraction unit 41, a gas identification unit 42, a concentration inversion unit 43, an odor concentration calculation unit 44, etc. These functional units can be implemented by software modules. The microprocessor can be a high-performance embedded platform, such as an STM32 series main controller (e.g., STM32F / H series) with an Intel Movidius Myriad X coprocessor for neural network acceleration, or an Intel Cyclone V SoC FPGA for hardware acceleration.

[0031] In this embodiment of the invention, the feature extraction unit 41 is used to extract multi-dimensional feature vectors from the multi-dimensional electrical signals of each odor gas sensing unit. The specific process is as follows: the DC resistance and the real part, imaginary part, and phase angle of the AC impedance at multiple frequency points (e.g., 5 characteristic frequency points) of each odor gas sensing unit are continuously acquired at a fixed sampling rate (e.g., 10 Hz). In one complete exposure cycle (including three stages: baseline stabilization, gas response, and recovery), the following three types of features are extracted for each odor gas sensing unit: Steady-state characteristics: the change in resistance (ΔR = R_gas - R_air), the change in impedance modulus, the change in the real part of impedance, the change in the imaginary part of impedance, and the change in phase angle when the response reaches equilibrium. Relative changes such as (R_gas - R_air) / R_air are also usually calculated. Dynamic characteristics: response rise slope (e.g., the average slope from gas injection to the response reaching 90% of the change), recovery slope, response time (t90), and recovery time (t10). Integral characteristic: Area under the response curve (AUC), which is the integral of the response value with respect to the baseline from the beginning to the end of the response; If the intelligent sensor chip array module 3 contains n odor gas sensing units, and each odor gas sensing unit extracts m features, then the total feature vector dimension is n×m. To reduce computational complexity, principal component analysis (PCA) can be further used for dimensionality reduction, retaining the top k principal components with a contribution rate of over 95% as the final multidimensional feature vector.

[0032] In this embodiment of the invention, the gas identification unit 42 incorporates a machine learning-based gas classification model to identify the types of odorous gases in the current environment based on multi-dimensional feature vectors. The gas classification model can be selected from Support Vector Machine (SVM), Random Forest, Shallow Neural Network, or Convolutional Neural Network (CNN). Taking SVM as an example, a one-to-many strategy is adopted, and the kernel function is Radial Basis Function (RBF). The penalty parameter C and kernel parameter γ are optimized through grid search. The model needs to be pre-trained: different types and concentrations of standard gas samples are prepared in the laboratory, the responses of the odor gas sensitive unit are collected and multi-dimensional feature vectors are extracted, and a feature-label dataset is constructed for training. In actual monitoring, the real-time multi-dimensional feature vectors are input into the model, and the probability of each category is output. The category corresponding to the maximum probability is taken as the identified odorous gas type. If the probability of all categories is lower than a set threshold (e.g., 0.6), it is determined to be an unknown gas.

[0033] In this embodiment of the invention, the concentration inversion unit 43 has a built-in concentration inversion model for different odor gases, used to invert the gas concentration corresponding to each odor gas based on multi-dimensional feature vectors. The concentration inversion model can be selected from BP neural network, support vector regression (SVR), extreme learning machine (ELM), or long short-term memory network (LSTM). Taking BP neural network as an example, the number of nodes in the input layer is the feature dimension, there are 2 hidden layers (64 and 32 nodes respectively), and 1 output layer (corresponding to the concentration of a certain odor gas). The activation function is ReLU in the hidden layer and a linear function in the output layer. The loss function is mean squared error, and the optimizer is Adam. Each odor gas is trained with an independent concentration inversion model. During inference, for each gas output by the odor gas sensitive unit, the corresponding concentration inversion model is called to obtain its gas concentration. (Unit: ppm or ppb)

[0034] In this embodiment of the invention, the odor concentration calculation unit 44 has a built-in odor threshold database, which pre-stores the odor thresholds of various common odor gases. (Units are consistent with concentrations), the olfaction thresholds in the olfaction threshold database are shown in Table 1 below: Table 1. Example table of olfaction thresholds in the olfaction threshold database. The odor concentration calculation unit 44 calculates the concentrations of each gas output by the concentration inversion unit 43. Calculate the odor contribution value of each odor gas. The maximum value among them is taken as the comprehensive odor concentration (OU) value of the current monitoring point. The calculation formula is as follows: in, This indicates the overall odor concentration (OU) value. Indicates the identified first The gas concentration corresponding to the odorous gas, ; This indicates the number of snoring thresholds pre-stored in the database. The olfaction threshold corresponding to each odorous gas ; This dimensionless value directly reflects the degree of influence of the strongest component of current environmental odors on human olfactory stimulation, serving as a key evaluation indicator of the degree of odor pollution.

[0035] In this embodiment of the invention, the display module 5 is connected to the information acquisition and processing module 4 and is used to display the comprehensive odor concentration OU value in real time, as well as the selectable gas type, concentration, navigation trajectory, alarm information, etc. The display module 5 can be a touch screen and supports human-computer interaction, such as setting alarm thresholds and manually controlling sampling.

[0036] In this embodiment of the invention, the power module 6 is connected to each of the above modules to provide stable power. In this embodiment, a high-energy-density lithium battery pack is used, which can support continuous mobile monitoring for more than 8 hours.

[0037] In this embodiment of the invention, to further enhance the spatial perception capability of the device, the embodiment may also include a positioning and navigation module 7 and a data storage and transmission module 8. The positioning and navigation module 7 adopts a GPS / BeiDou dual-mode positioning system to acquire the geographical location information (latitude and longitude, timestamp) of the mobile carrier 1 in real time and send the data to the information acquisition and processing module 4. The information acquisition and processing module 4 also includes a data merging unit 45, which is connected to the odor concentration calculation unit 44 and the positioning and navigation module 7. Since there is a response delay in gas detection (including gas path transmission delay and odor gas sensitive unit response time), the data merging unit 45 needs to perform time alignment and delay compensation. The specific method is as follows: the total delay time Td required from gas acquisition to signal stabilization is pre-determined (e.g., 5 seconds), the timestamp of each comprehensive odor concentration OU value is shifted backward by Td, and then matched with the GPS timestamp to generate odor monitoring data (lon, lat, OU, time, component concentration, etc.) with accurate geographic labels. This data can be sent to the display module 5 for real-time location display, or stored in a local storage card or uploaded to the cloud platform via a 4G / 5G module.

[0038] In this embodiment of the invention, to further visualize the pollution distribution, the information acquisition and processing module 4 may further include a spatial interpolation unit 46 connected to the data merging unit 45. The spatial interpolation unit 46 uses the inverse distance weighted interpolation (IDW) algorithm to perform gridded interpolation on the discrete odor monitoring data to generate a spatial distribution heat map of odor pollution. The interpolation formula is as follows: in, Indicates the point to be interpolated and the first... The distance between monitoring points This represents the power exponent (usually 2). This indicates the number of neighboring points participating in the interpolation. The interpolation grid resolution can be set to 10 m × 10 m. Based on the calculated comprehensive odor concentration (OU) value, a heat map is generated and overlaid on the map according to the preset color scale (e.g., 0-1 green, 1-5 yellow, 5-20 orange, >20 red). The heat map is then updated in real time on the display module 5 to intuitively show the pollution diffusion trend and high-value areas. In addition, the display module 5 can combine geographical location information to draw the mobile trajectory in real time and mark the comprehensive odor concentration OU value level of each monitoring point with different colors. When the comprehensive odor concentration OU value exceeds the preset threshold (such as 5), the odor mobile monitoring equipment will automatically trigger an audible and visual alarm to remind the operator to pay attention.

[0039] In the description of this invention, the references to "one embodiment," "some embodiments," "in this embodiment," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0040] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An odor mobile monitoring device based on an intelligent sensor chip, characterized in that, The odor mobile monitoring equipment is installed on a mobile carrier and includes: A gas acquisition module is used to collect ambient air data in real time. A smart sensor chip array module, connected to the gas acquisition module, includes multiple odor gas sensitive units, used to detect the ambient air in real time and obtain corresponding multidimensional electrical signals. An information acquisition and processing module, connected to the intelligent sensor chip array module, includes: A feature extraction unit is used to extract a multidimensional feature vector from the multidimensional electrical signal; A gas identification unit, connected to the feature extraction unit, has a built-in machine learning-based gas classification model for identifying the types of odorous gases in the current environment based on the multi-dimensional feature vector. A concentration inversion unit is connected to the gas identification unit and has a built-in machine learning-based concentration inversion model for inverting the gas concentration corresponding to each type of odor gas based on the multi-dimensional feature vector. An odor concentration calculation unit is connected to the concentration inversion unit and has a built-in odor threshold database. It is used to calculate the comprehensive odor concentration OU value by traversing the odor threshold database based on the gas concentration. A display module, connected to the information acquisition and processing module, is used to display the comprehensive odor concentration (OU) value in real time; A power supply module is connected to the gas acquisition module, the intelligent sensor chip array module, the information acquisition and processing module, and the display module, respectively.

2. The odor mobile monitoring device according to claim 1, characterized in that, Each of the odor gas sensing units is integrated onto a single chip using MEMS technology, and the chip size is less than 5mm x 5mm.

3. The odor mobile monitoring device according to claim 1, characterized in that, The multidimensional feature vector extracted by the feature extraction unit includes steady-state features, dynamic features, and integral features. The steady-state features include the change in resistance value when the response reaches equilibrium, impedance modulus, real part of impedance, imaginary part of impedance, and phase angle. The dynamic features include the response rise slope, recovery slope, response time, and recovery time. The integral feature is the area under the response curve.

4. The odor mobile monitoring device according to claim 1, characterized in that, The gas classification model is selected from one of the following: support vector machine, random forest, shallow neural network, and CNN classification model.

5. The odor mobile monitoring device according to claim 1, characterized in that, The concentration inversion model is selected from one of the following: BP neural network, support vector regression, extreme learning machine, and LSTM regression model.

6. The odor mobile monitoring device according to claim 1, characterized in that, The odor concentration calculation unit obtains the comprehensive odor concentration OU value using the following formula: in, This indicates the combined odor concentration (OU) value; Indicates the identified first The gas concentration corresponding to the odorous gas, ; This indicates the number of pre-stored olfaction thresholds in the olfaction threshold database. The olfaction threshold corresponding to each odorous gas .

7. The odor mobile monitoring device according to claim 1, characterized in that, It also includes a positioning and navigation module, connected to the information acquisition and processing module, for acquiring the geographical location information of the mobile carrier in real time. The information acquisition and processing module also includes a data merging unit, connected to the odor concentration calculation unit, for aligning the comprehensive odor concentration OU value with the geographical location information in time and compensating for delay, generating odor monitoring data with geographical tags and sending it to the display module for display.

8. The odor mobile monitoring device according to claim 7, characterized in that, The information acquisition and processing module further includes a spatial interpolation unit connected to the data merging unit. This unit is used to perform gridded interpolation on the discrete odor monitoring data using an inverse distance weighted interpolation algorithm, generating a spatial distribution heat map of odor pollution and sending it to the display module for display.

Images