Multi-mode intelligent data acquisition system and method for monitoring smell of automotive trim

Abstract

The invention relates to a multi-mode intelligent data acquisition system and method for automobile interior odor monitoring. The multi-mode intelligent data acquisition system comprises a terminal shell; the multi-mode sensor array module is integrated in the terminal shell or externally connected through a standard interface and used for synchronously collecting various types of parameters, and an array comprises an environmental parameter sensor, a gas concentration sensor, a selectable visual auxiliary sensor and a technological process sensor interface; the core processing and edge intelligent module comprises a main controller and an edge intelligent unit which are arranged in the terminal shell; the hybrid communication module comprises a wired communication unit and a wireless communication unit which are arranged in the terminal shell; and a power management module. According to the odor consistency monitoring system, data acquisition is completed, real-time and efficient preliminary diagnosis of abnormities is realized through a built-in special algorithm, a hierarchical and collaborative intelligent monitoring network is formed with a cloud system, and the performance and efficiency of the whole odor consistency monitoring system are remarkably improved.

Description

Technical Field

[0001] This invention relates to the fields of Internet of Things (IoT) data acquisition and edge intelligence technology, specifically to a multimodal intelligent data acquisition system and method for monitoring odors in automotive interiors. Background Technology

[0002] In the existing "IoT-based mass production consistency monitoring system for automotive interior odors," the real-time performance, accuracy, and comprehensiveness of the data are fundamental to the effective operation of the backend intelligent analysis model. However, the following challenges are encountered when implementing the data acquisition layer: (1) Low sensor integration: Existing data acquisition schemes mostly use single-function sensors to be deployed in a decentralized manner, and data is collected through independent transmitters or PLCs. This results in complex wiring, high cost, and difficulty in ensuring time synchronization. (2) Poor environmental adaptability: The interior production site has a complex environment (such as high temperature and humidity in the drying room, chemical gases in the painting workshop, and electromagnetic interference in the final assembly line). Ordinary commercial data acquisition equipment is insufficient in terms of long-term stability, protection level and anti-interference ability. (3) Lack of data preprocessing and intelligent judgment: Raw sensor data is often accompanied by noise and outliers, and traditional acquisition equipment is only responsible for "transmitting data" and does not have the ability to perform real-time quality judgment and preliminary anomaly identification at the source of data generation. This results in a large amount of invalid or normal data occupying network bandwidth, while real abnormal signals may be overwhelmed or delayed in reporting, affecting the response speed of the backend system; (4) Inflexible deployment: For mobile workstations or temporary monitoring points, there is a lack of a data acquisition terminal that can be easily deployed, battery powered, and can adaptively transmit data back via wireless network; (5) Complex protocol conversion: The equipment on the production line is of varying age and uses a variety of communication protocols (such as 4-20mA, 0-10V, Modbus, CAN bus, etc.), requiring an additional gateway for protocol conversion, which increases the system complexity and potential for failure. Therefore, there is a need for a multimodal intelligent data acquisition terminal and method that integrates a lightweight real-time anomaly detection algorithm and is specifically designed for monitoring the consistency of odor in automotive interiors. Summary of the Invention

[0003] The multimodal intelligent data acquisition system and method for monitoring odors in automotive interiors proposed in this invention can at least solve one of the technical problems in the background art.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: A multimodal intelligent data acquisition system for monitoring odors in automotive interiors includes: Terminal casing; The multimodal sensor array module, integrated inside the terminal housing or connected externally via a standard interface, is used to simultaneously acquire multiple types of parameters. The array includes: environmental parameter sensors, gas concentration sensors, optional vision-assisted sensors, and process sensor interfaces. The core processing and edge intelligence module includes a main controller and an edge intelligence unit housed within the terminal casing; A hybrid communication module, comprising a wired communication unit and a wireless communication unit disposed within a terminal housing; The power management module includes a wide-voltage power input circuit and a high-capacity lithium battery and charge / discharge management circuit, both housed within the terminal housing.

[0005] As a preferred embodiment of the multimodal intelligent data acquisition system for monitoring odors in automotive interiors as described in this invention, the environmental parameter sensors include a high-precision digital temperature and humidity sensor and an atmospheric pressure sensor. Gas concentration sensors are electrochemical or photoionization sensor modules targeting common odor substances in interior decoration, such as total volatile organic compounds, formaldehyde, and benzene series compounds. Optional visual auxiliary sensors integrate low-power miniature cameras or infrared thermal imaging modules to collect equipment status images or local temperature field distributions to assist in anomaly detection; The process sensor interface provides analog and digital input interfaces for connecting existing flow meters, pressure transmitters, and valve position feedback devices on the production line.

[0006] As a preferred embodiment of the multimodal intelligent data acquisition system for monitoring automotive interior odors as described in this invention, the main controller is responsible for sensor driving, data reading, protocol processing, and task scheduling. The edge intelligence unit has a built-in lightweight real-time time-series anomaly detection algorithm that performs online analysis on continuously collected sensor data, determines in real time whether the current data point is abnormal, and calculates an anomaly confidence score.

[0007] As a preferred embodiment of the multimodal intelligent data acquisition system for monitoring automotive interior odors according to the present invention, the real-time temporal anomaly detection algorithm includes: The edge intelligence unit processes each pre-processed sensor data stream. Perform the following calculations: 1) EWMA control chart calculation: Calculate the exponentially weighted moving average of the current data. Control upper / lower limits:

[0008]

[0009]

[0010] in, This represents the preprocessed sensor data value collected at the current moment, with the unit consistent with the measured physical quantity; It represents the exponentially weighted moving average at the current moment, reflecting the smoothed trend of recent changes in the data; This represents the weighting coefficient, with a value range of 0 < λ ≤ 1, controlling the weight of the current data point in the matrix. The closer it is to 1, the more attention it pays to the latest data; The closer it is to 0, the more it relies on historical data; This represents the sample mean of the data during the baseline learning phase. This represents the sample standard deviation of the data during the baseline learning phase; Indicates the upper limit of control; Indicates the lower limit of control; like Exceeding [ If this occurs, the "drift anomaly" flag will be triggered; This represents the control limit coefficient, a constant used to adjust the width of the control limit. It is usually related to the required confidence level and has a value range of 2 to 3. A higher value results in a wider control limit and a lower false alarm rate, but the false alarm rate may increase. 2) Fast-LOF estimation of local outliers: Maintain a fixed-length nearest neighbor data window For new points Calculate its position in the window before The average distance between the nearest neighbors Simultaneously calculate the average nearest neighbor distance of all points within the window. The estimated Quick LOF score is:

[0011] Set threshold ,like ( ) If this occurs, the "mutation anomaly" flag will be triggered.

[0012] As a preferred embodiment of the multimodal intelligent data acquisition system for monitoring automotive interior odors as described in this invention, the final abnormal state identifier is generated based on "drift anomaly" and "mutation anomaly" flags, combined with preset rules including: "an alarm is triggered by either of the two" or "an alarm is only triggered after N consecutive drift points". 0 is normal, 1 is abnormal; abnormal confidence score , ranging from 0 to 1.

[0013] As a preferred embodiment of the multimodal intelligent data acquisition system for monitoring automotive interior odors as described in this invention, the wired communication unit integrates RS-485 and CAN bus interfaces for accessing traditional industrial fieldbus networks; and integrates an Ethernet interface for high-speed and stable communication. The wireless communication unit integrates a 4G / 5G cellular communication module for wide-area wireless backhaul; integrates Wi-Fi and Bluetooth modules for flexible networking or near-field configuration and debugging within the workshop; and integrates a LoRa module for forming a star network with other low-power sensor nodes within the workshop.

[0014] As a preferred embodiment of the multimodal intelligent data acquisition system for monitoring automotive interior odors described in this invention, the wide-voltage power input circuit supports DC 9-36V industrial standard power input, and the high-capacity lithium battery and charge / discharge management circuit support continuous operation of the device terminal without external power supply and support solar panel charging.

[0015] A multimodal intelligent data acquisition method for monitoring odor in automotive interiors includes the following steps: S1. Configure the terminal device via Bluetooth or USB connection, set the data collection period and communication strategy; after the terminal is powered on, it automatically performs a self-test and enters a benchmark learning phase to initialize the parameters of the subsequent anomaly detection algorithm. S2. The main controller in the terminal device synchronously triggers the sensor array in the terminal device to collect data according to a preset cycle, and performs basic cleaning on the raw data. S3, For each preprocessed sensor data stream The edge intelligent units within the terminal device perform computations in parallel, including: 1) Calculation of EWMA control charts; 2) Fast-LOF estimation of local outlier factors; Based on the calculation results, the "drift anomaly" and "mutation anomaly" flags are determined, and then the final anomaly status identifier is generated by combining the results with preset rules. Anomaly confidence score ; S4, the core processing and edge intelligence module, will process data, and Encapsulated according to the protocol, the communication strategy is dynamically adjusted based on abnormal states: Normal mode Report aggregated feature values ​​such as mean, maximum and status heartbeat at low frequency, or use change reporting; Abnormal warning mode Immediately activate the high-speed reporting channel to upload alarm data packets containing raw high-density data points, anomaly confidence levels, and algorithm identifiers to the monitoring center via wired / high-speed wireless transmission. At the same time, detailed data for the abnormal time period can be stored locally. S5. The device terminal regularly reports its own status information to the monitoring center, including battery level, signal strength, health status of each sensor, and internal temperature; it receives remote instructions from the monitoring center to realize dynamic configuration updates, firmware over-the-air upgrades, and diagnostic test triggers; and it receives instructions to update algorithm parameters or fine-tune the benchmark model to adapt to production changes.

[0016] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0017] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0018] The beneficial effects of this invention are: 1. Dual-modal algorithm detection, covering all anomaly types: Combining EWMA and Fast-LOF algorithms, it can effectively detect both slow process drift and sudden spike anomalies simultaneously, with high detection coverage and low false alarm rate; 2. Reports are made only at low frequency or when there are changes under normal conditions, and at high speed under abnormal conditions, saving wireless network traffic and cloud storage space, and reducing cloud processing pressure. It is especially suitable for battery-powered and traffic-limited scenarios. Attached Figure Description

[0019] Figure 1 This is the overall system logic block diagram of the multimodal intelligent data acquisition system for monitoring odors in automotive interiors, as described in this invention.

[0020] Figure 2 This is a schematic block diagram of the multi-dimensional health data monitoring of the multimodal intelligent data acquisition system for monitoring odors in automotive interiors, as described in this invention.

[0021] Figure 3 This is a logic block diagram of the hierarchical response module of the multimodal intelligent data acquisition system for monitoring odors in automotive interiors, as described in this invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0023] Example 1 like Figure 1 As shown, the multimodal intelligent data acquisition system for monitoring automotive interior odors in this embodiment includes: Terminal housing; specifically, the terminal housing is made of metal or high-strength engineering plastic shell, with an IP65 or higher protection rating, dustproof and waterproof; the internal circuit board is treated with conformal coating to enhance moisture resistance and corrosion resistance; equipped with standard DIN rail mounting clips and magnetic base, supporting rail mounting, wall mounting and magnetic temporary installation, making deployment extremely flexible; The multimodal sensor array module, integrated inside the terminal housing or connected externally via a standard interface, is used to simultaneously acquire multiple types of parameters. The array includes: environmental parameter sensors, gas concentration sensors, optional vision-assisted sensors, and process sensor interfaces. The core processing and edge intelligence module includes a main controller and an edge intelligence unit housed within the terminal casing; A hybrid communication module, comprising a wired communication unit and a wireless communication unit disposed within a terminal housing; The power management module includes a wide-voltage power input circuit and a high-capacity lithium battery and charge / discharge management circuit, both housed within the terminal housing.

[0024] Specifically, environmental parameter sensors include high-precision digital temperature and humidity sensors and atmospheric pressure sensors; Gas concentration sensors are electrochemical or photoionization sensor modules targeting common odor substances in interior decoration, such as total volatile organic compounds, formaldehyde, and benzene series compounds. Optional visual auxiliary sensors integrate low-power miniature cameras or infrared thermal imaging modules to collect equipment status images or local temperature field distributions to assist in anomaly detection; The process sensor interface provides analog and digital input interfaces for connecting existing flow meters, pressure transmitters, and valve position feedback devices on the production line.

[0025] The main controller uses a high-performance, low-power ARM Cortex-M series or RISC-V microcontroller to handle sensor driving, data reading, protocol processing, and task scheduling. The edge intelligence unit has a built-in lightweight real-time anomaly detection algorithm that performs online analysis on continuously collected sensor data, determines in real time whether the current data point is abnormal, and calculates an anomaly confidence score.

[0026] Specifically, real-time time series anomaly detection algorithms include: The edge intelligence unit processes each pre-processed sensor data stream. Perform the following calculations: 1) EWMA control chart calculation: Calculate the exponentially weighted moving average of the current data. Control upper / lower limits:

[0027]

[0028]

[0029] in, This represents the preprocessed sensor data value collected at the current moment, with the unit consistent with the measured physical quantity; It represents the exponentially weighted moving average at the current moment, reflecting the smoothed trend of recent changes in the data; This represents the weighting coefficient, with a value range of 0 < λ ≤ 1, controlling the weight of the current data point in the matrix. The closer it is to 1, the more attention it pays to the latest data; The closer it is to 0, the more it relies on historical data; This represents the sample mean of the data during the baseline learning phase. This represents the sample standard deviation of the data during the baseline learning phase; Indicates the upper limit of control; Indicates the lower limit of control; like Exceeding [ If this occurs, the "drift anomaly" flag will be triggered; This represents the control limit coefficient, a constant used to adjust the width of the control limit. It is usually related to the required confidence level and has a value range of 2 to 3. The larger the value, the wider the control limit, the lower the false alarm rate, but the higher the probability of false alarm. 2) Fast-LOF estimation of local outliers: Maintain a fixed-length nearest neighbor data window For new points Calculate its position in the window before The average distance between the nearest neighbors Simultaneously calculate the average nearest neighbor distance of all points within the window. The estimated Quick LOF score is:

[0030] Set threshold ,like ( ) If this occurs, the "mutation anomaly" flag will be triggered.

[0031] Specifically, based on the "drift anomaly" and "mutation anomaly" flags, and combined with preset rules including "an alarm is triggered by either of the two" or "an alarm is triggered only after N consecutive drift points", the final anomaly status identifier is generated. 0 is normal, 1 is abnormal; abnormal confidence score , ranging from 0 to 1.

[0032] Among them, the hybrid communication module: Wired communication unit: integrates RS-485 and CAN bus interfaces for accessing traditional industrial fieldbus networks; integrates an Ethernet interface for high-speed and stable communication.

[0033] Wireless communication unit: integrates a 4G / 5G cellular communication module for wide-area wireless backhaul; integrates Wi-Fi and Bluetooth modules for flexible networking or near-field configuration and debugging within the workshop; integrates a LoRa module for forming a star network with other low-power sensor nodes within the workshop.

[0034] Specifically, the power management module: Wide voltage power input circuit: Supports DC9-36V industrial standard power input, with reverse connection protection and surge suppression; High-capacity lithium battery and charge / discharge management circuit: Supports continuous operation of the terminal without external power supply and supports charging via solar panel connection to meet the needs of mobile or temporary deployment.

[0035] Example 2 like Figures 2-3 As shown, the multimodal intelligent data acquisition method for monitoring odors in automotive interiors includes the following steps: S1. Configure the terminal device via Bluetooth or USB connection, set the data collection period and communication strategy; after the terminal is powered on, it automatically performs a self-test and enters a benchmark learning phase to initialize the parameters of the subsequent anomaly detection algorithm. S2. The main controller in the terminal device synchronously triggers the sensor array in the terminal device to collect data according to a preset cycle, and performs basic cleaning on the raw data. S3, For each preprocessed sensor data stream The edge intelligent units within the terminal device perform computations in parallel, including: 1) Calculation of EWMA control charts; 2) Fast-LOF estimation of local outlier factors; Based on the calculation results, the "drift anomaly" and "mutation anomaly" flags are determined, and then the final anomaly status identifier is generated by combining the results with preset rules. Anomaly confidence score ; S4, the core processing and edge intelligence module, will process data, and Encapsulated according to the protocol, the communication strategy is dynamically adjusted based on abnormal states: Normal mode Report aggregated feature values ​​such as mean, maximum and status heartbeat at low frequency, or use change reporting; Abnormal warning mode Immediately activate the high-speed reporting channel to upload alarm data packets containing raw high-density data points, anomaly confidence levels, and algorithm identifiers to the monitoring center via wired / high-speed wireless transmission. At the same time, detailed data for the abnormal time period can be stored locally. S5. The device terminal regularly reports its own status information to the monitoring center, including battery level, signal strength, health status of each sensor, and internal temperature; it receives remote instructions from the monitoring center to realize dynamic configuration updates, firmware over-the-air upgrades, and diagnostic test triggers; and it receives instructions to update algorithm parameters or fine-tune the benchmark model to adapt to production changes.

[0036] Specific examples: Taking the formaldehyde removal process monitoring in automotive dashboard painting workshops as an example: (1) Terminal configuration: An intelligent data acquisition terminal integrating a TVOC sensor, a high-precision flow meter interface (4-20mA), and a pressure sensor interface is selected. The spraying robot is configured with a data acquisition cycle of approximately 60 seconds per work cycle, and workshop Wi-Fi is the preferred communication method. (2) Benchmark learning: During a full day when the spraying process parameters are normal and qualified products are produced, the terminal automatically learns and obtains the average background concentration of TVOC. =50ppm, standard deviation =5; Average spray flow rate =1.2L / min, standard deviation =0.05, The algorithm parameters are initialized as follows: =0.2, =2.5, =2.0; (3) Normal operation: During subsequent production, the terminal collects data once per minute. Due to the stable process, the FastLOF score and EWMA score are... All are within the normal range. The terminal is in "normal mode" and reports the average and maximum values ​​of traffic and VOC to the cloud every 5 minutes. Network communication volume is extremely low. (4) Anomaly detection and response: Scenario A (Sudden Anomaly): After a spraying operation begins, a brief fluctuation in the air supply causes a momentary spike in the flow sensor reading (1.8 L / min). The Fast-LOF estimation module immediately calculates the nearest neighbor ratio relative to this point within the nearest window, and the score instantly rises to 3.5. This triggers a "mutation anomaly," and the terminal will [respond within 100 milliseconds]. =1, =0.9 and the original abnormal data points were urgently reported via Wi-Fi. The cloud system received the alarm within seconds and, combined with other information, determined it to be a transient disturbance. The system recorded the data but did not trigger a shutdown, while observing subsequent data. Scenario B (Slow Drift): A filter in the spraying equipment gradually becomes clogged, causing the flow rate to decrease slowly, linearly decreasing from 1.2 L / min to 1.0 L / min over 30 cycles; the EWMA module's... The value smoothly tracked this downward trend; at the 25th period, The value is lower than the calculated value. The system continuously triggers "drift anomaly"; after the terminal confirms the anomaly continues, it immediately switches to "anomaly warning mode" and begins to report high-density data at a frequency of once per second; the cloud system combines this warning and the traffic reduction mode, and issues a "filter maintenance warning" 30 minutes in advance before the traffic drops to the process lower limit, successfully avoiding the generation of batch defective products; (5) Algorithm update: One month later, due to a minor adjustment in the paint formula, the normal TVOC background concentration rose to 60ppm; the cloud system discovered through statistical analysis that the distribution of the "normal value" reported by the terminal had shifted, and therefore used downlink instructions to adjust the terminal's baseline average value. Updated to 60 ppm, standard deviation Updated to 6ppm to ensure the edge algorithm remains accurate.

[0037] As can be seen from this embodiment, the intelligent data acquisition terminal not only completes data acquisition, but also achieves real-time and efficient preliminary diagnosis of anomalies through built-in dedicated algorithms. It forms a hierarchical and collaborative intelligent monitoring network with the cloud system, which significantly improves the performance and efficiency of the entire odor consistency monitoring system.

[0038] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0039] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0040] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the multimodal intelligent data acquisition methods for monitoring the odor of automotive interiors described in the above embodiments.

[0041] It is understood that the systems, devices, and storage media provided in the embodiments of the present invention correspond to the methods provided in the embodiments of the present invention, and the explanations, examples, and beneficial effects of the relevant content can be referred to the corresponding parts of the above methods.

[0042] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., a solid-state disk (SSD)).

[0043] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0044] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0045] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multimodal intelligent data acquisition system for monitoring odor in automotive interiors, characterized in that, include: Terminal casing; The multimodal sensor array module, integrated inside the terminal housing or connected externally via a standard interface, is used to simultaneously acquire multiple types of parameters. The array includes: environmental parameter sensors, gas concentration sensors, optional vision-assisted sensors, and process sensor interfaces. The core processing and edge intelligence module includes a main controller and an edge intelligence unit housed within the terminal casing; A hybrid communication module, comprising a wired communication unit and a wireless communication unit disposed within a terminal housing; The power management module includes a wide-voltage power input circuit and a high-capacity lithium battery and charge / discharge management circuit, both housed within the terminal housing.

2. The multimodal intelligent data acquisition system for monitoring automotive interior odors according to claim 1, characterized in that: The environmental parameter sensors include a high-precision digital temperature and humidity sensor and an atmospheric pressure sensor; Gas concentration sensors are electrochemical or photoionization sensor modules targeting common odor substances in interior decoration, such as total volatile organic compounds, formaldehyde, and benzene series compounds. Optional visual auxiliary sensors integrate low-power miniature cameras or infrared thermal imaging modules to collect equipment status images or local temperature field distributions to assist in anomaly detection; The process sensor interface provides analog and digital input interfaces for connecting existing flow meters, pressure transmitters, and valve position feedback devices on the production line.

3. The multimodal intelligent data acquisition system for monitoring automotive interior odors according to claim 1, characterized in that: The main controller is responsible for sensor driving, data reading, protocol processing, and task scheduling; The edge intelligence unit has a built-in lightweight real-time time-series anomaly detection algorithm that performs online analysis on continuously collected sensor data, determines in real time whether the current data point is abnormal, and calculates an anomaly confidence score.

4. The multimodal intelligent data acquisition system for monitoring automotive interior odors according to claim 3, characterized in that: The real-time time series anomaly detection algorithm includes: The edge intelligence unit processes each pre-processed sensor data stream. Perform the following calculations: 1) EWMA control chart calculation: Calculate the exponentially weighted moving average of the current data. Control upper / lower limits: in, This represents the preprocessed sensor data value collected at the current moment, with the unit consistent with the measured physical quantity; It represents the exponentially weighted moving average at the current moment, reflecting the smoothed trend of recent changes in the data; This represents the weighting coefficient, with a value range of 0 < λ ≤ 1, controlling the weight of the current data point in the matrix. The closer it is to 1, the more attention it pays to the latest data; The closer it is to 0, the more it relies on historical data; This represents the sample mean of the data during the baseline learning phase. This represents the sample standard deviation of the data during the baseline learning phase; Indicates the upper limit of control; Indicates the lower limit of control; like Exceeding [ If this occurs, the "drift anomaly" flag will be triggered; This represents the control limit coefficient, a constant used to adjust the width of the control limit. It is related to the required confidence level and has a value range of 2 to 3. The larger the value, the wider the control limit, the lower the false alarm rate, but the higher the probability of false alarm. 2) Fast-LOF estimation of local outliers: Maintain a fixed-length nearest neighbor data window For new points Calculate its position in the window before The average distance between the nearest neighbors Simultaneously calculate the average nearest neighbor distance of all points within the window. The estimated Quick LOF score is: Set threshold ,like ( ) If so, the "mutation anomaly" flag will be triggered.

5. The multimodal intelligent data acquisition system for monitoring automotive interior odors according to claim 4, characterized in that: Based on the "drift anomaly" and "mutation anomaly" flags, and combined with preset rules including "an alarm is triggered by either of the two" or "an alarm is triggered only after N consecutive drift points", the final anomaly status identifier is generated. 0 is normal, 1 is abnormal; abnormal confidence score , ranging from 0 to 1.

6. The multimodal intelligent data acquisition system for monitoring automotive interior odors according to claim 1, characterized in that: The wired communication unit integrates RS-485 and CAN bus interfaces for accessing traditional industrial fieldbus networks; it also integrates an Ethernet interface for high-speed and stable communication. The wireless communication unit integrates a 4G / 5G cellular communication module for wide-area wireless backhaul; integrates Wi-Fi and Bluetooth modules for flexible networking or near-field configuration and debugging within the workshop; and integrates a LoRa module for forming a star network with other low-power sensor nodes within the workshop.

7. The multimodal intelligent data acquisition system for monitoring automotive interior odors according to claim 1, characterized in that: The wide-voltage power input circuit supports DC 9-36V industrial standard power input, and the high-capacity lithium battery and charge / discharge management circuit support the device terminal to work continuously without external power supply and support solar panel charging.

8. A data acquisition method for a multimodal intelligent data acquisition system for monitoring automotive interior odors, based on any one of claims 1-7, characterized in that, Includes the following steps: S1. Configure the terminal device via Bluetooth or USB connection, and set the data collection period and communication strategy; After the terminal is powered on, it automatically performs a self-test and enters a benchmark learning phase to initialize the parameters of the subsequent anomaly detection algorithm. S2. The main controller in the terminal device synchronously triggers the sensor array in the terminal device to collect data according to a preset cycle, and performs basic cleaning on the raw data. S3, For each preprocessed sensor data stream The edge intelligent units within the terminal device perform computations in parallel, including: 1) Calculation of EWMA control charts; 2) Fast-LOF estimation of local outlier factors; Based on the calculation results, the "drift anomaly" and "mutation anomaly" flags are determined, and then the final anomaly status identifier is generated by combining the results with preset rules. Anomaly confidence score ; S4, the core processing and edge intelligence module, will process data, and Encapsulated according to the protocol, the communication strategy is dynamically adjusted based on abnormal states: Normal mode Report aggregated feature values ​​such as mean, maximum and status heartbeat at low frequency, or use change reporting; Abnormal warning mode Immediately activate the high-speed reporting channel to upload alarm data packets containing raw high-density data points, anomaly confidence levels, and algorithm identifiers to the monitoring center via wired / high-speed wireless transmission. At the same time, detailed data for the abnormal time period can be stored locally. S5. The device terminal regularly reports its own status information to the monitoring center, including battery level, signal strength, health status of each sensor, and internal temperature; it receives remote instructions from the monitoring center to realize dynamic configuration updates, firmware over-the-air upgrades, and diagnostic test triggers; and it receives instructions to update algorithm parameters or fine-tune the benchmark model to adapt to production changes.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, the processor performs the steps of the method as described in claim 8.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the computer program is executed by the processor, it causes the processor to perform the steps of the method as described in claim 8.

Images