A method for detecting and removing odors in a vehicle cabin

Abstract

The present application relates to the technical field of vehicles, in particular to a method for detecting and removing odors in a vehicle cabin, comprising a vehicle-mounted power supply module, a sampling pretreatment module, a multi-modal sensor array unit, a main control processing unit, a human-computer interaction module, a wireless communication module and a linkage control interface; the present application can identify complex odors of various sources in the vehicle cabin, distinguish the types and intensities of the odors, and can realize linkage with the air conditioner, ventilation system and the like of the vehicle, actively remove the odors in the vehicle cabin, take into account the driving comfort, and reduce the harm of odors to the health of the passengers.

Description

Technical Field

[0001] This invention relates to the field of vehicle technology, and more specifically to a method for detecting and removing odors inside automobiles. Background Technology

[0002] With the rapid development of the automotive industry and the increasing demands of consumers for driving and riding experience, in-car air quality has become one of the important indicators for measuring the quality of a car and one of the factors that consumers consider when purchasing a car. There are many sources of odors in a car, including volatile organic compounds (VOCs) released from interior materials, mold growing in the air conditioning system, body odor from passengers, food residue odors, and external pollutants. If no measures are taken to remove the odors in the car for a long time, it will not only affect the comfort of driving and riding, but also cause harm to the health of passengers if the odors are exposed for a long time.

[0003] In the existing technology, there are some in-vehicle odor detection solutions, such as: using a single gas sensor to detect the concentration of specific pollutants, but unable to distinguish the type and intensity of odors; some high-end models are equipped with air quality sensors, which can only detect conventional indicators such as PM2.5 or CO2, and lack the ability to identify complex odors; existing detection devices usually lack intelligent linkage with the vehicle's air conditioning and ventilation systems, and cannot actively eliminate odors.

[0004] Therefore, there is an urgent need for a method that can identify complex odors from multiple sources in a vehicle, distinguish the types and intensities of odors, and work in conjunction with the vehicle's air conditioning and ventilation systems to actively remove odors while ensuring driving comfort and reducing the harm to the health of occupants. Summary of the Invention

[0005] The present invention aims to provide a method for detecting and removing odors inside automobiles, which can identify complex odors from multiple sources, distinguish the types and intensities of odors, and can be linked with the vehicle's air conditioning and ventilation systems to actively remove odors inside the vehicle, while taking into account driving comfort and reducing the harm to the health of occupants caused by odors.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for detecting and removing odors inside a car, characterized in that it includes an on-board power supply module, a sampling preprocessing module, a multimodal sensor array unit, a main control processing unit, a human-machine interaction module, a wireless communication module, and a linkage control interface, and further includes the following steps: S1. Vehicle power-on self-test and baseline calibration: After power-on, the main controller completes the hardware connectivity self-test of the whole machine and performs zero-point baseline calibration on all sensors in the multi-modal sensor array unit. The vehicle power supply module includes the vehicle circuit and the built-in lithium battery. When the vehicle is ignited, it is powered by the vehicle circuit. After the vehicle is turned off, the vehicle power supply module automatically switches to the built-in lithium battery and supports offline continuous sampling. S2. Zonal and Time-Based Adaptive Gas Sampling: The sampling preprocessing module includes an electronically controlled solenoid valve group, a micro air pump, a flow regulating valve, a filter unit, a constant temperature air chamber, and four independent sampling branches. The four sampling branches are used to collect air samples from the driver's seat, passenger seat, rear seats, and trunk, respectively. The sampling branches are switched in stages by the electronically controlled solenoid valve group. The micro air pump is speed-regulated by the main control PWM pulse, and the sampling flow rate is constant from 0.5L / min to 1.0L / min. The sampling time for a single area is 5s, and the sampling cycle of the four sampling branches is fixed at 20s. The constant temperature air chamber maintains the internal temperature at 30±5℃. S3. Signal Conditioning and Temperature and Humidity Compensation Correction: The multimodal sensor array unit is set in a constant temperature chamber, including temperature and humidity sensors, as well as MOS metal oxide semiconductor sensors, QCM quartz crystal microbalance sensors and electrochemical sensors for detecting odor types. Different types of sensor signals output by the multimodal sensor array unit are sent to the main control processing unit. The main control processing unit performs gain amplification and filtering processing on different types of sensor signals, and then converts them into digital signals through an AD converter with 16-bit precision and 100Hz sampling frequency. The following formula is used for temperature and humidity compensation correction:

[0007] in, This is the corrected concentration value. K is the temperature correction coefficient of the corresponding sensor, H is the humidity correction coefficient of the corresponding sensor, T is the ambient temperature collected in real time by the temperature and humidity sensor, and RH is the ambient humidity collected in real time by the temperature and humidity sensor. S4. Feature Extraction and Odor Intelligent Recognition Steps: The main control processing unit has a built-in odor classification model based on the random forest algorithm. The main control processing unit extracts the 12-dimensional feature vector of the corrected concentration value in a 30-second complete detection cycle and inputs it into the odor classification model to identify and classify the odor. S5. Odor Classification Judgment: The main control unit pre-stores the national standard safety thresholds for odor components: formaldehyde ≤ 0.1 mg / m³, benzene ≤ 0.11 mg / m³, toluene ≤ 1.10 mg / m³, xylene ≤ 1.50 mg / m³, TVOCs ≤ 0.60 mg / m³, and ammonia ≤ 0.20 mg / m³. Based on the corrected concentration values ​​of each odor, the pollution level is classified as qualified, slightly exceeding the standard, or severely exceeding the standard. S6. Data display and graded sound and light warning: The human-machine interaction module includes a display screen, a three-color LED and a buzzer. The display screen shows the odor concentration, odor type and pollution level of each zone in real time. When the level is qualified, the three-color LED is always lit in green. When the level is slightly exceeded, the three-color LED is lit in yellow and flashes at a frequency of 1Hz. The buzzer alarms intermittently for 2 seconds. When the level is severely exceeded, the three-color LED is lit in red and flashes continuously at a frequency of 2Hz. The buzzer alarms continuously. S7. Vehicle-mounted equipment linkage for odor removal: The linkage control interface uses a relay and four-way opto-isolated circuits to output switch control signals. The linkage control interface is used to control the vehicle-mounted equipment for odor removal. The vehicle-mounted equipment includes a fragrance system, an air purifier, and a window lift module. The odor removal priority is from high to low as follows: formaldehyde, smoke, musty, food, leather, sweat, and compound odors. S8. Data storage and remote interaction: The main control processing unit has a built-in flash memory storage unit with a capacity of no less than 16GB, which can store at least 30 days of historical detection data. The wireless communication module integrates CAN bus, Bluetooth and Wi-Fi. The CAN bus synchronizes data to the vehicle's infotainment system, and Bluetooth and Wi-Fi connect to mobile terminal applications, supporting remote viewing of real-time data and retrieval of historical records.

[0008] Furthermore, in step S2, when the odor concentration in any area exceeds the standard and is 30% or more higher than that in other areas, the sampling time for that area is extended to 8 seconds, and the sampling time for other qualified areas is shortened to 3 seconds. When multiple areas exceed the standard at the same time, they are sorted by odor concentration from high to low, and the sampling time for the most polluted area can be adjusted up to 10 seconds.

[0009] Furthermore, the constant temperature chamber is equipped with a heating element, a semiconductor cooling element, and a cooling fan. When the temperature in the constant temperature chamber is below 25°C, the heating element is activated to raise the temperature; when the temperature is above 35°C, the semiconductor cooling element and the cooling fan are activated to lower the temperature.

[0010] Furthermore, in step S3, the MOS metal oxide semiconductor sensor detects VOCs, nitrogen-containing odor gases, and complex odors; QCM quartz crystal microbalance sensor detects high-boiling-point VOCs, aldehydes and ketones, benzene compounds, nitrogen- and sulfur-containing odor gases, and complex odors; Electrochemical sensors detect ppb-level irritating odors from aldehydes, nitrogen- and sulfur-containing odor gases, some VOCs, and complex odors.

[0011] Furthermore, in step S3, the temperature correction coefficient of the electrochemical sensor is 0.006 and the humidity correction coefficient is 0.002; the temperature correction coefficient of the QCM quartz crystal microbalance sensor is 0.004 and the humidity correction coefficient is 0.002; and the temperature correction coefficient of the MOS metal oxide semiconductor sensor is 0.005 and the humidity correction coefficient is 0.003.

[0012] Furthermore, in step S4, the 12-dimensional features include the peak response, response duration, attenuation slope, steady-state response value, and ambient temperature and humidity of each sensor. The 12-dimensional features are input into the odor classification model to identify and classify odors. The random forest algorithm contains 100 decision trees with a maximum depth of 15 layers per tree. Eight-dimensional features are randomly sampled in a single run. The Gini coefficient is used to complete node splitting. The classification results of all decision trees are fused through majority voting and the odor classification results are output. Simultaneously, the intensity level of the odor is divided according to step S5.

[0013] Furthermore, in step S7, the fragrance system adopts an intermittent release mode. When the level is qualified, the fragrance system is in standby mode; when the level is slightly exceeded, the fragrance system releases for 5 seconds at a time, with an interval of 15 minutes and a single dose of 0.1 ml, with the release rate controlled at 5 ml / h to 10 ml / h; when the level is severely exceeded, the fragrance system releases for 8 seconds at a time, with an interval of 8 minutes and a single dose of 0.2 ml, with the release rate controlled at 10 ml / h to 15 ml / h; when multiple odors are severely exceeded at the same time, the fragrance system releases for 10 seconds at a time, with an interval of 5 minutes and an hourly release rate of 15 ml / h to 20 ml / h. After the level is restored to qualified, the fragrance system continues to run for 10 minutes before going into standby mode.

[0014] Furthermore, when the level of odor exceeds the standard slightly, the corresponding purification mode of the air purifier will be activated. When the level of odor exceeds the standard severely, the purification mode of the air purifier will be activated and the window lift module will be controlled to open the window by 5cm to 10cm for auxiliary ventilation. When the user manually operates the in-vehicle equipment, the linkage control interface will suspend the control of the in-vehicle equipment until any of the following conditions occur: the pollution level of the odor in the vehicle changes, the odor in the vehicle continues to exceed the standard severely, or the user manually activates the automatic mode. In these three cases, the linkage control interface will resume the control of the in-vehicle equipment.

[0015] The beneficial effects of this solution are: the multi-sensor array unit simultaneously collects all types of odorous and harmful gases such as TVOCs, formaldehyde, benzene series compounds, ammonia, and hydrogen sulfide. Combined with gas filtration, constant temperature control, separate sensors, and a dedicated temperature and humidity algorithm for triple compensation, it reduces the detection errors caused by high and low temperatures, dust, and humidity fluctuations in the vehicle. The detection data is stable and reliable in long-term operation. The four-zone time-based sampling structure covers the entire interior space of the vehicle, including the driver's seat, passenger seat, rear seats, and trunk. It automatically identifies high-pollution areas and adaptively extends the sampling time to quickly locate the source of odor pollution, facilitating the inspection of the entire vehicle interior and the rectification of pollution sources. After the concentration detected by each sensor is corrected, it is input into the odor classification model based on the random forest algorithm. 12-dimensional features are extracted to identify and classify the odor, distinguish the intensity level of the odor, and classify it into three pollution levels of qualified, slightly exceeding the standard and severely exceeding the standard according to the national standard concentration threshold. The evaluation standard is unified, quantitative and comparable. Based on the type of odor, the level of exceeding the standard, and the priority of health hazards, the system automatically links the air conditioning internal and external circulation, window ventilation, dedicated air purification by the air purifier, and intermittent decomposition and release of in-vehicle fragrance, forming a closed-loop process of "detection-identification-early warning-active removal". There is no need to manually turn on the purification equipment, which actively improves the air quality inside the vehicle. The human-machine interaction module displays the zone detection data in real time, and the three-color LED and buzzer take into account both intuitiveness and noise reduction requirements; it is compatible with multiple communications such as CAN, Bluetooth and Wi-Fi, and is suitable for two types of application scenarios: OEM matching for car manufacturers and aftermarket vehicle modification. It supports remote viewing of real-time data and retrieval of at least 30 days of historical detection records via mobile APP. The hardware integrates overvoltage, reverse connection, and short circuit power supply protection. The linkage control end uses four opto-isolation circuits to isolate the detection device from the original vehicle electrical circuit interference. It automatically switches to the lithium battery when the engine is off to achieve offline sampling and monitoring, and is suitable for use in all working conditions of vehicle ignition and shutdown. The system prioritizes manual intervention, automatically pausing the linkage program after the user manually adjusts the air conditioning or windows. It also configures multiple automatic recovery conditions to balance the needs of autonomous control with the advantages of automated management. The fragrance system adopts a quantitative intermittent release mode to avoid excessive fragrance and secondary pollution. Attached Figure Description

[0016] Figure 1 This is a system structure block diagram of the present invention; Figure 2 This is a schematic diagram of the sampling preprocessing and zoned gas path structure in this invention; Figure 3 This is a connection block diagram of the peripheral circuit of the main control processing unit in this invention; Figure 4 This is a flowchart of the odor detection method in this invention. Detailed Implementation

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0018] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0019] The basic implementation examples are as follows: Figure 1-4 As shown, the hardware of this method consists of seven major modules: an on-board power supply module, a sampling preprocessing module, a multimodal sensor array unit, a main control processing unit, a human-machine interaction module, a wireless communication module, and a linkage control interface. All modules are electrically connected through onboard cables and wiring harnesses and have a sealed dustproof structure design. Only the gas sampling port and external wiring terminals are reserved. The reason for the sealed design is to prevent dust and moisture inside the vehicle from entering the circuit board and sensor, causing short circuits and gas blockages, and ensuring long-term stable operation in the complex on-board environment. The effect of the sealed structure is to significantly reduce the equipment failure rate and extend the service life of the sensor and main control circuit.

[0020] The vehicle power supply module adopts a standard 12V DC input for automobiles and is connected to the vehicle's ACC gear circuit. This enables the entire unit to automatically power on when the vehicle is started and switch to offline detection standby mode after the vehicle is turned off, eliminating the need for manual power on / off. As a preferred embodiment, the module incorporates a DC-DC step-down circuit to convert the input 12V voltage into 5V and 3.3V low-voltage power supplies compatible with the main control, sensors, air pump, and communication modules. It also integrates overvoltage, reverse connection, and short-circuit protection circuits to prevent surges and reverse currents from burning out components, making it suitable for various vehicle electrical environments. The lithium battery is designed in parallel with the vehicle's ACC circuit, allowing it to independently power the flow control valve, micro air pump, and sensor array after the vehicle is turned off, enabling offline odor sampling when the vehicle is stationary.

[0021] The sampling pretreatment module is the core of the entire device's air circuit, integrating a miniature diaphragm air pump, flow regulating valve, filter unit, constant temperature air chamber, electronically controlled solenoid valve group, and four independent sampling branches. The ends of the four sampling branches are respectively located in four key areas inside the vehicle: the driver's seat, the front passenger seat, the rear seats, and the trunk, forming four sampling ports. This regional sampling structure is designed because the concentration of odors generated by interior decoration and occupant activities varies significantly in different areas of the vehicle. Single-point sampling cannot reflect the overall air quality of the vehicle. Simultaneous, time-sharing sampling in four areas can accurately capture localized high-pollution points, enabling odor source tracing and rapid location of odor-generating areas for targeted cleaning. The electronically controlled solenoid valve group switches between the four air circuits at different times. The basic control logic is that the four circuits sample sequentially in a loop, with a basic sampling time of 5 seconds per circuit and a fixed complete cycle of 20 seconds. The miniature air pump is controlled via the main control output. The PWM pulse signal adjusts the rotation speed to stably control the air sampling flow rate between 0.5L / min and 1.0L / min. Under normal operating conditions, a standard flow rate of 0.8L / min is selected. The constant flow rate design aims to ensure that the volume of air samples collected each time is uniform, avoiding deviations in sensor detection data caused by airflow fluctuations. The flow regulating valve is equipped with an electromagnetic drive component, which is connected to the power management unit of the vehicle power supply module to synchronously switch between vehicle power supply and lithium battery offline power supply modes. The regulating valve's function is to evenly distribute the sampling flow rate of the four branch pipes, preventing excessive or insufficient airflow in some pipes from causing insufficient sample representativeness. The filter unit can use two stages of high-density non-woven fabric connected in series in the middle of the air path to intercept hair, dust, lint, oil, and water vapor in the air inside the vehicle, preventing small particles from entering the constant temperature air chamber and clogging the sensor's sensitive elements, thus avoiding sensor sensitivity attenuation.

[0022] The constant temperature chamber integrates a small heating element, a semiconductor cooling element, and a miniature cooling fan. The constant temperature control target is to maintain the gas in the chamber at 30±5℃. When the temperature is below 25℃, the heating element is activated to raise the temperature. When the temperature is above 35℃, the cooling element and the cooling fan are activated to lower the temperature. The constant temperature structure is designed because the gas sensor detection values ​​are greatly affected by the ambient temperature. High and low temperatures can cause positive and negative deviations in concentration detection. The constant temperature preprocessing can unify the detection temperature of all air samples, eliminate the detection error caused by temperature difference, and achieve dual error correction in conjunction with subsequent temperature and humidity algorithm compensation. The adaptive sampling adjustment logic is synchronously integrated into the gas path control program. When the main controller detects that the odor concentration in a certain sampling area exceeds the standard, and the concentration in that area is 30% or more higher than that in the other three areas, it automatically extends the sampling time of the high-pollution area to 8 seconds, and shortens the sampling time of the other qualified areas to 3 seconds. When multiple areas exceed the standard at the same time, they are sorted according to the concentration from high to low. The sampling time of the most polluted area can be adjusted up to 10 seconds. The total cycle of the four independent sampling branches always remains unchanged at 20 seconds. The adaptive adjustment design effect is to increase the amount of air sample collected in high-pollution areas, amplify the odor characteristic signal, and further improve the detection accuracy of pollution source areas.

[0023] The multimodal sensor array unit is fully integrated inside the constant-temperature chamber, simultaneously collecting parameters of in-vehicle gaseous pollutants, ambient temperature and humidity, and particulate matter. The array includes electrochemical sensors, MOS metal oxide semiconductor sensors, QCM quartz crystal microbalance sensors, and temperature and humidity sensors. The reason for the multi-type sensor fusion design is that a single sensor can only detect a single type of gas and cannot cover volatile organic compounds (VOCs) emitted from the interior. The three types of sensors (abbreviation for Compounds) cover all pollution sources, including irritating gases from human metabolism, musty odors, and combustion odors. Their detection ranges complement each other, achieving full coverage of all odor components. Among them, the MOS metal oxide semiconductor sensor is used to detect typical VOCs such as formaldehyde, benzene, toluene, xylene, and acetaldehyde, total volatile organic compounds (TVOCs) released from vehicle interior materials (plastics, leather, adhesives, etc.), ammonia (mainly from interior materials, sweat and respiratory metabolic products of drivers and passengers), and compound odors (gases such as carbon monoxide and hydrogen sulfide, as well as smoke, food residue odors, and musty interior odors generated during driving). The term "compound odor" in the following text refers to the same content and will not be repeated hereafter.

[0024] QCM quartz crystal microbalance sensors are used to detect high-boiling-point volatile organic compounds (phthalates, fatty acids, alcohols, ketones, esters, etc.), aldehydes and ketones and benzene compounds (formaldehyde, acetaldehyde, benzene, toluene, xylene, etc.), nitrogen- and sulfur-containing odor gases (ammonia, hydrogen sulfide and other irritating odor gases, hydrogen sulfide mainly comes from mold growth in air conditioning systems, mold on seat fabrics, etc.) and complex odors, making up for the lack of sensitivity in low-concentration detection of semiconductor sensors; Electrochemical sensors are used to detect ppb-level irritating odors from aldehydes (formaldehyde, acetaldehyde, acrolein, etc.), nitrogen- and sulfur-containing odor gases (same as described above), some VOCs (carbon monoxide, benzene compounds, etc.), and complex odors. The combination of these three types of sensors covers the toxic odor substances in vehicles as defined by national standards. Temperature and humidity sensors simultaneously collect real-time temperature and humidity data from the air chamber, providing raw environmental parameters for subsequent algorithm compensation. The output signal types of each sensor are different: formaldehyde and ammonia output voltage signals, benzene compounds output weak current signals, VOCs output frequency signals, and temperature and humidity directly output digital signals. The differentiated signal output characteristics correspond to independent signal conditioning circuits at the back end.

[0025] The main control processing unit uses a 32-bit automotive-grade STM32 series microcontroller as the main control chip for the vehicle MCU (hereinafter referred to as MCU). It has an onboard flash memory storage unit with a capacity of no less than 16GB, a 16-bit precision 100Hz sampling frequency AD acquisition circuit, and a complete set of internal operating programs. These programs include sensor power-on baseline calibration programs, sensor-specific temperature, humidity, and flow compensation algorithms, a 12-dimensional feature extraction program, an odor classification model based on a random forest algorithm, national standard pollution level determination logic, multi-priority vehicle linkage control programs, and actuator fault identification programs. The large-capacity flash memory storage unit is designed to persistently store sensor calibration parameters, national standard concentration thresholds, machine learning model training parameters, and at least 30 days of complete historical detection data, ensuring data integrity even after power failure and enabling long-term traceability of detection records. The AD acquisition circuit receives the analog raw signals output by the sensor array and transmits them to the built-in signal processing unit, which is customized for different sensor signals. Independent amplification and filtering schemes are used. Formaldehyde's 0-100mV voltage signal is amplified by 50 to 100 times gain and then filtered using a 10Hz cutoff frequency second-order low-pass filter. Benzene's 0-20μA weak current signal is amplified by 200 to 500 times high gain and simultaneously filtered using a 10Hz second-order low-pass filter to eliminate vehicle electromagnetic interference. Ammonia's 0-500mV voltage signal is amplified by 20 to 50 times gain and filtered using a 10Hz second-order low-pass filter. VOC's 1-5kHz frequency signal is filtered using a 50Hz first-order low-pass filter. Odor level's 0-1V voltage signal is filtered using a 0.1 to 10Hz second-order bandpass filter to extract effective odor fluctuation signals. The independent signal conditioning design for each channel amplifies weak but effective detection signals, filters electromagnetic interference noise generated by in-vehicle electrical appliances and motors, improves the signal-to-noise ratio, and avoids misjudgments of concentration caused by noise. The filtered and amplified analog signal is converted into a digital signal by an AD acquisition circuit and sent to the MCU for temperature and humidity compensation correction. The calculation formula is as follows:

[0026] in, This is the corrected concentration value. K represents the odor concentration value detected by the corresponding sensor, H represents the temperature correction coefficient of the corresponding sensor, T represents the ambient temperature collected in real time by the temperature and humidity sensor, and RH represents the ambient humidity collected in real time by the temperature and humidity sensor.

[0027] Different sensors are matched with specific correction coefficients. Among them, the temperature correction coefficient of the electrochemical sensor is 0.006 and the humidity correction coefficient is 0.002; the temperature correction coefficient of the QCM quartz crystal microbalance sensor is 0.004 and the humidity correction coefficient is 0.002; and the temperature correction coefficient of the MOS metal oxide semiconductor sensor is 0.005 and the humidity correction coefficient is 0.003.

[0028] The MCU uses a 30-second detection cycle to extract 12-dimensional features from the calibrated sensor data. These features include the sensor's peak response, peak value, response time to peak, signal attenuation slope within 30 seconds after peak, average steady-state response within 10 seconds of peak stability, and average ambient temperature and humidity values ​​within the 30-second cycle. All features are normalized from minimum to maximum to a range of 0 to 1. Normalization eliminates differences in the magnitude of signals from different sensors, improving the classification accuracy of the machine learning model. The processed 12-dimensional features are then input into an odor classification model based on a random forest algorithm to identify odors. The model parameters are set to 100 independent decision trees, with a maximum depth of 15 layers per tree, and random sampling is performed each time. Eight-dimensional features are used for node splitting, and the Gini coefficient criterion is used to divide nodes. Finally, the entire decision tree is fused by majority voting to output a unique odor classification result. The odor classification model is equipped with a dedicated training dataset. The dataset collection environment covers temperatures from -20℃ to 60℃, humidity from 0 to 99%RH, and wind speed from 0 to 0.5m / s. The collection time periods include three working conditions: 5 minutes after vehicle start-up, 30 minutes of driving, and 1 hour of parking in a closed space. The collection location is a closed vehicle interior space. The sample collection standard is that the number of samples collected for each odor category is no less than 500 sets, and the total number of samples for all odors is no less than 2500 sets. An additional 100 sets of odorless blank control samples are provided. The dataset is divided into training set, validation set, and test set in a 7:2:1 ratio to ensure the model's generalization ability.

[0029] In complex odors, due to the large number of types involved, the criteria for determining musty odor are: concentration of mold metabolites ≥ 0.01 mg / m³; smoke odor is: nicotine concentration ≥ 0.005 mg / m³; food odor is: fatty acid concentration ≥ 0.02 mg / m³; leather odor is: leather volatiles concentration ≥ 0.015 mg / m³; and sweat odor is: ammonia concentration ≥ 0.15 mg / m³. Samples with multiple odors are individually labeled as mixed odors. This dual labeling standard avoids classification errors caused by a single determination method. The odor classification model output can distinguish six odor types: musty, smoke, food, leather, sweat, and mixed odors, and simultaneously classify the odor intensity into three levels: weak, medium, and strong.

[0030] The MCU pre-stores national standard-limited safe concentration thresholds for pollutants: formaldehyde ≤ 0.1 mg / m³, benzene ≤ 0.11 mg / m³, toluene ≤ 1.10 mg / m³, xylene ≤ 1.50 mg / m³, total volatile organic compounds (TVOCs) ≤ 0.60 mg / m³, and ammonia ≤ 0.20 mg / m³. By comparing the temperature and humidity corrected concentrations of each pollutant with the thresholds, the system automatically classifies the in-vehicle pollution level into three categories: acceptable, slightly exceeding the standard, and severely exceeding the standard. The national standard-quantified grading design replaces manual subjective judgment, making it easier for users to intuitively determine the risk level of in-vehicle air quality.

[0031] The human-machine interaction module includes a display screen, a tri-color LED (warning light), and a buzzer. The display screen can be an LCD screen integrated into the vehicle's infotainment panel, which displays the pollutant concentrations of the four sampling zones, the types of odors identified, and the current pollution level in real time. The visual display design allows drivers and passengers to monitor the odor pollution situation inside the vehicle in real time. The tri-color LED is divided into green, yellow, and red, and is equipped with a graded sound and light warning logic. When the air quality inside the vehicle is within acceptable limits, the green LED is constantly lit and there is no buzzer noise. In the case of slightly exceeding the standard, the yellow LED flashes continuously at a frequency of 1Hz, and the buzzer emits an intermittent alarm every 2 seconds to avoid continuous noise interference to the driver and passengers. In the case of severely exceeding the standard, the red LED flashes at a high frequency of 2Hz, and the buzzer sounds an alarm continuously. At the same time, a warning message is pushed to the vehicle's infotainment system and mobile APP through the wireless communication module. The graded warning design takes into account the effectiveness of the warning and the need for driving noise reduction, and distinguishes between fault and normal exceeding the standard alarm to avoid user misjudgment.

[0032] The wireless communication module integrates CAN bus communication units, Bluetooth communication units, and Wi-Fi communication units. Its multi-communication protocol compatibility design adapts to both OEM and aftermarket application scenarios. The CAN bus connects to the original vehicle's in-vehicle network, synchronously transmitting real-time detection data, odor identification results, and pollution levels to the original vehicle's central control screen, enabling native vehicle interaction. The Bluetooth and Wi-Fi channels connect to the user's mobile terminal APP, supporting remote viewing of real-time air quality data and retrieval of historical detection records stored in the flash memory unit. Users can monitor the accumulation of odors in a parked, closed vehicle even when not in the car. This remote interaction design addresses the shortcomings of existing devices that only offer local viewing and lack remote monitoring capabilities.

[0033] The linkage control interface is used to output control signals to drive the original vehicle's air conditioning, windows, air purifier, and in-vehicle fragrance system to remove odors. The linkage control interface uses a relay and a TLP521-4 four-channel opto-isolation circuit design, with an isolation voltage ≥2500Vrms. The relay is an Omron G5V-2-H1-12VDC double-pole double-throw miniature signal relay, with a 1N4007 freewheeling diode connected in parallel across the relay coil. The opto-isolation circuit is used because the detection device's control circuit is a 3.3V low-voltage digital circuit, while the original vehicle's actuator is a 12V high-voltage in-vehicle circuit. Direct connection would generate electromagnetic interference and back electromotive force, potentially burning out the main control MCU. Optical isolation completely cuts off the odor. Electrical connections on both sides block interference signals. The freewheeling diode absorbs the reverse electromotive force of the relay coil when it is de-energized, protecting the circuit components. The circuit connection logic is that the GPIO pin of the main control MCU outputs a 3.3V high and low level control signal to the input side of the optocoupler. The output side of the optocoupler drives the relay coil to switch on and off. The relay contacts output a 12V switching signal to control the original vehicle actuator. Each actuator has an independent set of optocoupler and relay channels, which do not interfere with each other. The actuator synchronously feeds back the operating status switching signal to the main control. The MCU continuously monitors the feedback signal. If the low-level fault feedback signal does not change for 3 seconds, it determines that the corresponding equipment is faulty and immediately triggers the audible and visual alarm of the human-machine interaction module.

[0034] The actuator control signals are divided into three categories: the vehicle air conditioner has two control signals, one for switching between internal and external circulation modes and the other for adjusting the fan speed; the window lift has one control signal, with a high-level trigger opening the window by 5cm to 10cm for ventilation; the vehicle air purifier has two control signals, one for controlling on / off and the other for switching between existing formaldehyde decomposition / smoke removal / mold removal and antibacterial purification modes; the vehicle fragrance system has an independent timing control loop, using the existing intermittent release mode. The fragrance release quantification control logic is differentiated according to the pollution level. When the air quality is acceptable, the fragrance system is in standby mode and does not release; when the air quality is slightly above the standard, the fragrance system releases for 5 seconds at a time, with a 1-second interval. For 5 minutes, a single application of 0.1 ml is used, with the release rate controlled at 5 ml / h to 10 ml / h. When the odor exceeds the standard, the fragrance system releases for 8 seconds at a time, with an interval of 8 minutes, and a single application of 0.2 ml, with the release rate controlled at 10 ml / h to 15 ml / h. When multiple odors exceed the standard simultaneously, the fragrance system releases for 10 seconds at a time, with an interval of 5 minutes, with an hourly release rate of 15 ml / h to 20 ml / h. After the odor in the car returns to the acceptable standard, the fragrance system continues to run intermittently for 10 minutes before entering standby mode. The intermittent quantitative release design avoids the secondary air pollution caused by the superposition of fragrances caused by a large amount of fragrance release at one time, while also decomposing residual odors in a long-term manner.

[0035] The vehicle's coordinated actions are executed according to a fixed priority, following the principle of prioritizing health hazards and the degree of exceedance. The priorities, from highest to lowest, are: formaldehyde exceeding the standard, smoke odor, musty odor, food odor, leather odor, and sweat odor. When the pollution level is within acceptable limits, no automatic actions are executed; the vehicle's original air conditioning, windows, air purifier, and fragrance system remain operational. In cases of mild exceedance, the air purifier's corresponding purification mode is activated based on the highest priority odor. In cases of severe exceedance, the purifier's enhanced purification mode is activated, simultaneously controlling window ventilation. When multiple high-risk odors are present, a superimposed enhanced treatment logic is executed. Three consecutive tests determine the severity of the odor. If the odor level exceeds the standard, the wireless communication module pushes a parking ventilation reminder to the vehicle's infotainment system and mobile app. The entire linkage program is set with manual intervention priority logic. After the user manually operates the air conditioning, windows, air purifier, and fragrance system, the program immediately suspends all automatic linkage control and executes the user's manually set commands. There are three trigger conditions for restoring automatic linkage: when the odor level in the vehicle changes, automatic control is temporarily restored and the user's manually set core parameters are retained; automatic control is restored when the user manually activates the automatic linkage function in the vehicle or through the mobile app; and automatic linkage control is automatically restored when the odor level remains severely exceeded and there is no manual intervention for a long time.

[0036] The following is a complete description of the entire process of using this method, from vehicle startup to offline monitoring after engine shutdown. After the vehicle is started, the on-board power supply module is connected to the ACC12V power supply, and all modules of the whole machine are powered on and started. The first step is to perform a power-on self-test and baseline calibration process. The main control MCU sequentially checks the hardware connectivity status of the sensor array, air circuit solenoid valve, air pump, communication circuit, and linkage control circuit. Simultaneously, baseline zero-point calibration is performed on all sensors to eliminate long-term static drift errors of the sensors. After the self-test calibration is completed, the system automatically enters the zone-based and time-based gas sampling process. The main control outputs a control signal to drive the electronically controlled solenoid valve group to switch the four sampling branches. The micro air pump starts and the sampling flow rate is stabilized by PWM pulse speed regulation.At a rate of 8L / min, four sampling ports sequentially collect air samples from the driver's seat, front passenger seat, rear seats, and trunk. After entering the pipeline, the air samples first pass through a two-stage filtration unit to remove dust and moisture, then are sent to a constant-temperature chamber for pre-treatment. The MCU compares the concentration detected in the four areas in real time, identifies high-pollution areas, and adaptively adjusts the sampling duration of each branch. After completing a single 20-second cycle of sampling, the raw signal output from the multimodal sensor array unit is sent to the signal processing unit. Amplification and filtering are performed according to the independent rules of each sensor, and the signal is converted into a digital signal. This digital signal is then input into the temperature and humidity correction formula to complete concentration compensation, resulting in accurately corrected pollutant concentration data. Multiple sets of calibration data are collected in 30-second intervals, and 12-dimensional features are extracted and input into an odor classification model based on a random forest algorithm. This automatically identifies the type and intensity of odors in the vehicle and simultaneously determines the pollution level as acceptable, slightly exceeding the standard, or severely exceeding the standard by comparing it with national standard thresholds. After the determination, the MCU transmits the concentration, odor type, and pollution level of the four zones to the display screen in the human-machine interaction module for real-time display, and simultaneously triggers a three-color LED and buzzer-based tiered sound and light warning. Subsequently, the whole vehicle linkage treatment process begins, where the MCU matches the corresponding purification strategy according to the priority of formaldehyde > smoke > musty odor > other odors, controlling the fragrance system and air The air purifier and window lift module initiate their respective removal programs, while the fragrance releases decomposing fragrance according to a quantified intermittent rule. The actuators simultaneously report their operating status to the main controller, triggering a fault alarm immediately upon detection. The wireless communication module synchronizes all detection, warning, and linkage data to the original vehicle's CAN bus and the user's mobile app in real time, while the flash memory unit continuously stores all historical detection data. During normal vehicle operation, the entire system continuously cycles through sampling, correction, identification, warning, and linkage processes, dynamically adjusting the purification strategy in real time. When passengers manually adjust the air conditioning, windows, or air purifier, the automatic linkage program immediately pauses, preserving user settings to meet their needs. After conditions are restored, automatic treatment restarts. When the vehicle is turned off, the ACC power supply is disconnected, and the onboard power module automatically switches to the built-in lithium battery power supply. The entire unit enters offline monitoring mode, continuously cycling through zone sampling and odor detection processes, recording accumulated odor data under sealed, static vehicle conditions. Users can remotely view the air quality inside the vehicle after the engine is turned off via a mobile app. When the vehicle is restarted, the entire unit re-performs a power-on self-test calibration, repeating the entire cycle. If the device malfunctions in sensors, air pumps, or linkage actuators, the system immediately triggers alternating red and yellow fault alarms, simultaneously pushing fault information to the vehicle's infotainment system and mobile terminal, reminding the user to inspect and maintain the device.

[0037] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific technical solutions and / or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the technical solutions of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for detecting and removing odors inside a car, characterized in that, The system includes an on-board power supply module, a sampling preprocessing module, a multimodal sensor array unit, a main control processing unit, a human-machine interaction module, a wireless communication module, and a linkage control interface. It also includes the following steps: S1. Vehicle power-on self-test and baseline calibration: After power-on, the main controller completes the hardware connectivity self-test of the whole machine and performs zero-point baseline calibration on all sensors in the multi-modal sensor array unit. The vehicle power supply module includes the vehicle circuit and the built-in lithium battery. When the vehicle is ignited, it is powered by the vehicle circuit. After the vehicle is turned off, the vehicle power supply module automatically switches to the built-in lithium battery and supports offline continuous sampling. S2. Zonal and Time-Based Adaptive Gas Sampling: The sampling preprocessing module includes an electronically controlled solenoid valve group, a micro air pump, a flow regulating valve, a filter unit, a constant temperature air chamber, and four independent sampling branches. The four sampling branches are used to collect air samples from the driver's seat, passenger seat, rear seats, and trunk, respectively. The sampling branches are switched in stages by the electronically controlled solenoid valve group. The micro air pump is speed-regulated by the main control PWM pulse, and the sampling flow rate is constant from 0.5L / min to 1.0L / min. The sampling time for a single area is 5s, and the sampling cycle of the four sampling branches is fixed at 20s. The constant temperature air chamber maintains the internal temperature at 30±5℃. S3. Signal Conditioning and Temperature and Humidity Compensation Correction: The multimodal sensor array unit is set in a constant temperature chamber, including temperature and humidity sensors, as well as MOS metal oxide semiconductor sensors, QCM quartz crystal microbalance sensors and electrochemical sensors for detecting odor types. Different types of sensor signals output by the multimodal sensor array unit are sent to the main control processing unit. The main control processing unit performs gain amplification and filtering processing on different types of sensor signals, and then converts them into digital signals through an AD converter with 16-bit precision and 100Hz sampling frequency. The following formula is used for temperature and humidity compensation correction: in, This is the corrected concentration value. K is the temperature correction coefficient of the corresponding sensor, H is the humidity correction coefficient of the corresponding sensor, T is the ambient temperature collected in real time by the temperature and humidity sensor, and RH is the ambient humidity collected in real time by the temperature and humidity sensor. S4. Feature Extraction and Odor Intelligent Recognition Steps: The main control processing unit has a built-in odor classification model based on the random forest algorithm. The main control processing unit extracts the 12-dimensional feature vector of the corrected concentration value in a 30-second complete detection cycle and inputs it into the odor classification model to identify and classify the odor. S5. Odor Classification Judgment: The main control unit pre-stores the national standard safety thresholds for odor components: formaldehyde ≤ 0.1 mg / m³, benzene ≤ 0.11 mg / m³, toluene ≤ 1.10 mg / m³, xylene ≤ 1.50 mg / m³, TVOCs ≤ 0.60 mg / m³, and ammonia ≤ 0.20 mg / m³. Based on the corrected concentration values ​​of each odor, the pollution level is classified as qualified, slightly exceeding the standard, or severely exceeding the standard. S6. Data display and graded sound and light warning: The human-machine interaction module includes a display screen, a three-color LED and a buzzer. The display screen shows the odor concentration, odor type and pollution level of each zone in real time. When the level is qualified, the three-color LED is always lit in green. When the level is slightly exceeded, the three-color LED is lit in yellow and flashes at a frequency of 1Hz. The buzzer alarms intermittently for 2 seconds. When the level is severely exceeded, the three-color LED is lit in red and flashes continuously at a frequency of 2Hz. The buzzer alarms continuously. S7. Vehicle-mounted equipment linkage for odor removal: The linkage control interface uses a relay and four-way opto-isolated circuits to output switch control signals. The linkage control interface is used to control the vehicle-mounted equipment for odor removal. The vehicle-mounted equipment includes a fragrance system, an air purifier, and a window lift module. The odor removal priority is from high to low as follows: formaldehyde, smoke, musty, food, leather, sweat, and compound odors. S8. Data storage and remote interaction: The main control processing unit has a built-in flash memory storage unit with a capacity of no less than 16GB, which can store at least 30 days of historical detection data. The wireless communication module integrates CAN bus, Bluetooth and Wi-Fi. The CAN bus synchronizes data to the vehicle's infotainment system, and Bluetooth and Wi-Fi connect to mobile terminal applications, supporting remote viewing of real-time data and retrieval of historical records.

2. The method for detecting and removing odors inside a car according to claim 1, characterized in that: In step S2, when the odor concentration in any area exceeds the standard and is 30% or more higher than that in other areas, the sampling time for that area is extended to 8 seconds, and the sampling time for other qualified areas is shortened to 3 seconds. When multiple areas exceed the standard at the same time, the areas are sorted from high to low odor concentration, and the sampling time for the most polluted area can be adjusted up to 10 seconds.

3. The method for detecting and removing odors inside a car according to claim 2, characterized in that: The constant temperature chamber is equipped with a heating element, a semiconductor cooling element, and a cooling fan. When the temperature in the constant temperature chamber is below 25°C, the heating element is activated to raise the temperature. When the temperature is above 35°C, the semiconductor cooling element and the cooling fan are activated to lower the temperature.

4. The method for detecting and removing odors inside a car according to claim 1, characterized in that: In step S3, the MOS metal oxide semiconductor sensor detects VOCs, nitrogen-containing odor gases, and complex odors; QCM quartz crystal microbalance sensor detects high-boiling-point VOCs, aldehydes and ketones, benzene compounds, nitrogen- and sulfur-containing odor gases, and complex odors; Electrochemical sensors detect ppb-level irritating odors from aldehydes, nitrogen- and sulfur-containing odor gases, some VOCs, and complex odors.

5. A method for detecting and removing odors inside a car according to claim 4, characterized in that: In step S3, the temperature correction coefficient of the electrochemical sensor is 0.006 and the humidity correction coefficient is 0.002; the temperature correction coefficient of the QCM quartz crystal microbalance sensor is 0.004 and the humidity correction coefficient is 0.002; and the temperature correction coefficient of the MOS metal oxide semiconductor sensor is 0.005 and the humidity correction coefficient is 0.

003.

6. A method for detecting and removing odors inside a car according to claim 1, characterized in that: In step S4, the 12-dimensional features include the peak response, response duration, attenuation slope, steady-state response value, and ambient temperature and humidity of each sensor. The 12-dimensional features are input into the odor classification model to identify and classify odors. The random forest algorithm contains 100 decision trees with a maximum depth of 15 layers per tree. Eight-dimensional features are randomly sampled in a single run. The Gini coefficient is used to split nodes. The classification results of all decision trees are fused using the majority voting method, and the odor classification results are output. Simultaneously, the intensity level of the odor is divided according to step S5.

7. A method for detecting and removing odors inside a car according to claim 6, characterized in that: In step S7, the fragrance system adopts an intermittent release mode. When the level is within acceptable limits, the fragrance system is in standby mode. When the level is slightly exceeded, the fragrance system releases for 5 seconds at a time, with an interval of 15 minutes and a single dose of 0.1 ml, with the release rate controlled at 5 ml / h to 10 ml / h. When the level is severely exceeded, the fragrance system releases for 8 seconds at a time, with an interval of 8 minutes and a single dose of 0.2 ml, with the release rate controlled at 10 ml / h to 15 ml / h. When multiple odors are severely exceeded simultaneously, the fragrance system releases for 10 seconds at a time, with an interval of 5 minutes and an hourly release rate of 15 ml / h to 20 ml / h. After the level is restored to acceptable, the fragrance system continues to run for 10 minutes before going into standby mode.

8. A method for detecting and removing odors inside a car according to claim 7, characterized in that: When the level of odor exceeds the standard slightly, the corresponding purification mode of the air purifier will be activated. When the level of odor exceeds the standard severely, the purification mode of the air purifier will be activated and the window lift module will be controlled to open the window by 5cm to 10cm for auxiliary ventilation. When the user manually operates the in-vehicle equipment, the linkage control interface will suspend the control of the in-vehicle equipment until any of the following conditions occur: the pollution level of the odor in the car changes, the odor in the car continues to exceed the standard severely, or the user manually activates the automatic mode. In these three cases, the linkage control interface will resume the control of the in-vehicle equipment.

Images