Vehicle interior air quality self-adaptive control method and device considering motion sickness risk

Abstract

The invention provides an in-vehicle air quality self-adaptive control method and device considering motion sickness risks. The in-vehicle air quality self-adaptive control method comprises the following steps that S1, data are acquired in real time through a data acquisition module; s2, establishing a motion sickness risk assessment model, and outputting a motion sickness risk score in real time; s3, the motion sickness risk score is compared with a set threshold value, and the current motion sickness risk grade of the passenger is obtained; s4, the anti-motion sickness device automatically matches an intervention mode according to the motion sickness risk level; s5, calculating the cumulative impact quantity of the current vehicle to match the initial intervention time length, and dynamically updating the actual intervention time; and S6, dynamically correcting the control gain and execution parameters of the anti-corona device through a double-layer fuzzy PID adjustment algorithm, and performing self-learning correction of the control rule in combination with a weighted comprehensive error integral. According to the method, predictive optimization of the cabin environment can be realized, and the motion sickness risk of passengers is effectively inhibited or eliminated while the air quality is guaranteed.

Description

Technical Field

[0001] This invention relates to the fields of vehicle intelligent control, occupant physiological health monitoring, and in-vehicle cabin environment adjustment, and in particular to an adaptive control method and device for in-vehicle air quality that takes into account the risk of motion sickness. Background Technology

[0002] Motion sickness is an discomfort induced by a conflict between the motion information received by the vestibular, visual, and proprioceptive systems of passengers during travel, often manifesting as dizziness, nausea, and vomiting. In the field of smart cockpits, improving passenger comfort and effectively preventing motion sickness has become one of the core research directions.

[0003] However, existing motion sickness relief techniques have the following shortcomings:

[0004] 1. Delayed Triggering Logic: Current intervention measures mostly adopt a passive triggering mode, meaning that the system only initiates intervention after the passenger experiences significant discomfort and actively reports it. This reactive approach cannot effectively prevent motion sickness symptoms before they occur, resulting in poor motion sickness prevention outcomes.

[0005] 2. Isolated Dimensions in Motion Sickness Prevention Strategies: Motion sickness is not determined by a single factor, but rather is the result of a multi-dimensional coupling of factors such as vehicle dynamics parameters, air quality, cabin odor, and temperature and humidity. Current technologies typically treat air quality control and motion sickness prevention control as independent systems, failing to consider the non-linear increase in occupant sensitivity to odors and CO2 concentration under extreme dynamic driving conditions (such as continuous curves). This dimensional isolation results in control strategies that cannot achieve comprehensive prediction and advance compensation of in-vehicle environmental parameters under complex operating conditions.

[0006] 3. Lack of flexible adjustment in intervention intensity: Existing motion sickness intervention programs often adopt fixed intervention patterns and fail to dynamically adjust based on real-time motion sickness risk levels. In aspects such as air quality regulation, if the adjustment intensity is too high or there are sudden changes in wind speed, it may even induce secondary discomfort in passengers due to changes in air pressure or sound field, exacerbating the degree of motion sickness.

[0007] Therefore, how to establish a closed-loop control system that can assess the risk of motion sickness in passengers in real time and adjust the in-vehicle environment accordingly has become a key technological bottleneck for improving the smart cockpit riding experience. Summary of the Invention

[0008] In view of this, the purpose of the present invention is to provide an adaptive control method and device for in-vehicle air quality that takes into account the risk of motion sickness. It can fully consider the head movement characteristics of the occupants, motion sickness assessment indicators and in-vehicle environmental parameters, quantify the motion sickness risk level of the occupants in real time, and dynamically correct the in-vehicle air quality adjustment logic and intervention intensity accordingly, so as to achieve predictive optimization and flexible control of in-vehicle environmental parameters and improve the ride comfort of the smart cockpit.

[0009] To achieve the above objectives, the present invention adopts the following technical solution: an adaptive control method for in-vehicle air quality considering the risk of motion sickness, comprising the following steps;

[0010] S1. Real-time acquisition of occupant head posture data and in-vehicle environmental parameters via data acquisition module;

[0011] S2. Establish a motion sickness risk assessment model and input the occupant's head posture data into the motion sickness risk assessment model to output motion sickness risk scores in real time;

[0012] S3. Compare the motion sickness risk score with the set threshold to quantify and classify the motion sickness risk of the occupants and obtain the current motion sickness risk level of the occupants.

[0013] S4. The anti-motion sickness device automatically matches the intervention mode according to the motion sickness risk level and activates the progressive multi-dimensional intervention measures corresponding to each mode.

[0014] S5. Calculate the cumulative impact of the current vehicle to match the initial intervention duration, and dynamically update the actual intervention time;

[0015] S6. Through a dual-layer fuzzy PID control algorithm, the control gain and execution parameters of the anti-dizziness device are dynamically corrected, and the control rules are self-learned and corrected by combining the weighted comprehensive error integral.

[0016] In a preferred embodiment, the occupant head posture data in step S1 includes the acceleration of the head in the longitudinal, lateral, and vertical directions, as well as the angular velocity about each axis; the in-vehicle environmental parameters include the CO2 concentration, PM2.5 concentration, temperature, and humidity inside the vehicle.

[0017] In a preferred embodiment: the motion sickness risk assessment model described in S2 adopts an improved six-degree-of-freedom subjective vertical conflict model based on environmental tolerance correction. It calculates environmental stress factors and dynamically corrects the model's pathogenic threshold, thereby integrating in-vehicle environmental parameters to jointly assess motion sickness risk. The specific calculation includes the following steps:

[0018] S21: Using the otolith organ and semicircular canal dynamic model, process the acquired occupant head posture data, calculate the conflict error between the actual perceived vertical vector and the subjective vertical vector of the inner ear vestibular system, denoted as ||Δv||.

[0019] S22: Extract the deviations of the current in-vehicle CO2 concentration, PM2.5 concentration, temperature, and humidity from the preset comfort thresholds; calculate the environmental stress factor E using a weighted normalization method. stress The calculation formula is as follows:

[0020]

[0021] In the formula, X j These represent the real-time values ​​of CO2, PM2.5, temperature, and humidity inside the vehicle, respectively; X target,j X represents the target comfort value for the corresponding indicator. limit,j This represents the tolerance limit value for the corresponding indicator; w j This represents the dizziness-inducing weighting coefficient for the corresponding indicator.

[0022] S23: Based on the physiological principle that the harsher the environment, the lower the human body's tolerance, utilizing environmental stress factor E stress The pathogenicity threshold parameter of the Hill function is dynamically adjusted to obtain the real-time dynamic threshold b', calculated using the following formula:

[0023]

[0024] In the formula, b is the baseline pathogenicity threshold under standard conditions; λ is the environmental sensitivity attenuation coefficient.

[0025] S24: Substitute the real-time dynamic threshold b' into the Hill function, perform nonlinear mapping on the sensory conflict vector error ||Δv|| to obtain the sensory conflict conversion quantity h, and input h into the motion sickness transfer function model. After quantization, the final motion sickness index MSI is calculated. The calculation formulas are as follows:

[0026]

[0027]

[0028] In the formula, n is the nonlinear shape exponent, P is the gain coefficient, τ1 is the time constant, and s is the Laplace operator.

[0029] In a preferred embodiment: the progressive multidimensional intervention measures described in S4 include;

[0030] S41: Primary intervention: Activate ventilation function, and make routine adjustments to air direction and air volume according to the deviation of PM2.5 and CO2 to stabilize the indicators within the target threshold and suppress early discomfort;

[0031] S42: Intermediate Intervention: Based on ventilation, the fragrance release unit is automatically triggered to smoothly adjust the center value of the airflow direction, causing the airflow to shift towards the occupant's head; the airflow is appropriately increased according to the motion sickness rate; at the same time, the sound and light alarm unit emits short beeps for intermittent warnings, working together to alleviate the physiological conflicts in the early stages of motion sickness.

[0032] S43: Advanced Intervention: The proportional coefficient is significantly reduced according to the level of motion sickness. The actuator movement is made extremely smooth by weakening the adjustment intensity. At the same time, the audible and visual alarm unit continuously sounds an alarm to remind the driver to drive carefully and reduce the physical impact on the vehicle.

[0033] In a preferred embodiment, in S5, the cumulative impact amount is obtained by weighted calculation of occupant head posture data; the actual intervention duration is matched to an initial value based on the cumulative impact amount and dynamically attenuated according to the rate of change of motion sickness risk, and the calculation steps are as follows:

[0034] Single impact I single (t) represents the impact received by the head at the current moment, and the cumulative impact amount I. total (t) Using a preset time T as the statistical time window, the single impact quantity I... single (t) is calculated using the practice-weighted cumulative method, and the formulas are as follows:

[0035]

[0036]

[0037] In the formula, P i Represents occupant head posture data; ω i λ represents the weighting factor of the corresponding parameter. i Representative dimensionless conversion coefficient; cumulative impact I total (t) represents the integral of a single impact force over a preset time T, which can also be understood as the total impact on the occupant's head over the preset time T. Since the degree of dizziness is a cumulative result over a period of time, continuous integration cannot be achieved in actual engineering; therefore, a discretized rectangular method is used to approximate the integration. Thus, the overall calculation formula can be understood as the integral of the single impact force I... single (t) is the integral over the time interval from tT to t. Δt is the base time step, representing the time difference between two adjacent data acquisitions, and can also be understood as the time width of each integration unit.

[0038] Preset mild, moderate, and severe cumulative impact thresholds, and initial intervention duration T of the device. base Matching is performed based on a comparison of the cumulative impact amount and a threshold; after intervention is initiated, the remaining intervention duration is calculated in real time starting from the time the intervention is initiated. The duration decay factor γ is determined by the motion sickness rate of change ΔMSI. The larger the motion sickness rate of change ΔMSI, the slower the duration decays, and vice versa. The formula for calculating ΔMSI is as follows:

[0039]

[0040] Actual intervention time T of the device actual The formula for calculating (t) is:

[0041]

[0042] Intervention shall be terminated immediately if any of the following conditions are met: remaining intervention duration ≤ 0, cumulative impact amount is below the first threshold, or the motion sickness level is risk-free.

[0043] In a preferred embodiment, in S6, the input variables of the dual-layer fuzzy PID control algorithm include PM2.5 deviation, CO2 deviation, temperature deviation, humidity deviation, and motion sickness index (MSI); the output variables include PID correction values ​​for dynamically adjusting controller parameters, and adjustment parameters for controlling the air conditioning actuator.

[0044] The dual-layer fuzzy PID control algorithm adopts a dual-layer fuzzy inference architecture, specifically including:

[0045] Air quality inference layer: Taking PM2.5 deviation, PM2.5 deviation change rate, CO2 deviation, and CO2 deviation change rate as inputs, the basic PID parameter correction amount and basic execution adjustment amount are inferred through core fuzzy rules;

[0046] Scene adaptation correction layer: Taking MSI, temperature deviation and humidity deviation as input, the output correction coefficient dynamically corrects the air quality inference layer results. Among them, the MSI correction outputs the correction coefficient according to its fuzziness level.

[0047] In a preferred embodiment, the two-layer fuzzy PID control algorithm has PID parameter correction rules adapted to the scenario, including:

[0048] (1) In scenarios with high risk of motion sickness, reduce the proportional and integral effects and enhance the differential effect to improve smoothness;

[0049] (2) In scenarios where air quality is severely exceeded, the proportional and integral effects are enhanced to accelerate regulation;

[0050] (3) In steady-state scenarios, the proportional and differential coefficients are kept stable, and the integral action is increased to eliminate steady-state error.

[0051] In a preferred embodiment, in S6, constraints are imposed on the control of the actuator, including: limiting the airflow level, airflow angle, and single adjustment range when the motion sickness risk level is high; and limiting the external circulation ratio and single mode adjustment range.

[0052] The self-learning optimization strategy is as follows: calculate the weighted comprehensive error integral J at a preset period to evaluate the control effect; and adaptively correct the output center value of the fuzzy control according to the deviation between J and the target value and the deviation change pattern until the algorithm converges.

[0053] The present invention also provides an adaptive control device for in-vehicle air quality that takes into account the risk of motion sickness, for executing the aforementioned adaptive control method for in-vehicle air quality that takes into account the risk of motion sickness, comprising:

[0054] Data acquisition module: used to collect occupant head posture data and in-vehicle environmental parameters in real time;

[0055] Decision control module: used to run the motion sickness risk assessment model and the in-vehicle air quality adaptive control method that takes motion sickness risk into account, and output motion sickness risk level, intervention duration and control instructions;

[0056] Anti-motion sickness module: Controls in-vehicle air quality, fragrance release, and audible and visual alarms based on the motion sickness risk level and anti-motion sickness execution commands;

[0057] User interaction module: provides passengers with three working modes: automatic, manual, and preventative;

[0058] Storage module: Used to store motion sickness model parameters, historical data, and learning logs.

[0059] In a preferred embodiment: the device has three motion sickness intervention modes;

[0060] Automatic mode: The device operates automatically based on input parameters and real-time motion sickness assessment results;

[0061] Manual mode: Based on your own discomfort, manually select the specific anti-dizziness intervention mode, adjust the fragrance release concentration, and set the intervention duration;

[0062] Prevention mode: Remotely preset using mobile phone Bluetooth or IoT terminal to activate environmental optimization before boarding.

[0063] Compared with the prior art, the present invention has the following beneficial effects:

[0064] (1) It solves the problem of dimensional isolation in multi-objective regulation. The dual-layer fuzzy logic can dynamically balance air purification efficiency and anti-dizziness comfort according to the level of motion sickness risk, and achieve the global optimality of the strategy.

[0065] (2) It has extremely strong intervention flexibility. The dynamic softening of PID parameters and the step size limit constraint of execution under high risk effectively eliminate the secondary stimulation caused by step adjustment.

[0066] (3) It has long-term self-evolution capability. The self-learning correction strategy can automatically optimize the center value of the fuzzy control rule for the sensitivity differences of different passengers and system degradation such as the aging of the air conditioning filter.

[0067] (4) Progressive multi-dimensional collaborative intervention: The anti-dizziness device can coordinate with various means such as air quality adjustment, fragrance release, and sound and light alarms. It automatically matches the progressive intervention mode according to the level of motion sickness risk, matches the initial intervention time based on the cumulative impact of the vehicle, and dynamically updates the actual intervention time according to the motion sickness change rate, thus avoiding over-intervention or under-intervention.

[0068] (5) It has high integration and strong real-time performance. The algorithm complexity is adapted to the lightweight operation requirements of the vehicle MCU. It can monitor the degree of motion sickness of passengers in real time and make reasonable interventions. It has broad prospects for industrial application. Attached Figure Description

[0069] Appendix Figure 1 A flowchart of an adaptive control method and device for in-vehicle air quality that takes into account the risk of motion sickness;

[0070] Appendix Figure 2 This is a schematic diagram of the motion sickness risk assessment model framework;

[0071] Appendix Figure 3 A flowchart of a progressive multidimensional intervention method;

[0072] Appendix Figure 4 Flowchart of the dynamic intervention duration algorithm;

[0073] Appendix Figure 5 Here is a flowchart of the two-layer fuzzy PID control algorithm;

[0074] Appendix Figure 6 Flowchart for self-learning correction strategy;

[0075] Appendix Figure 7 A schematic diagram of the device framework for an adaptive control method of in-vehicle air quality that takes into account motion sickness risk feedback. Detailed Implementation

[0076] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0077] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0078] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations according to this application; as used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise; furthermore, it should be understood that when the terms “comprising” and / or “including” are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0079] This invention proposes an adaptive control method and device for in-vehicle air quality that takes into account the risk of motion sickness, as shown in the attached figure. Figure 1-7 As shown, the method includes the following steps;

[0080] S1. Real-time acquisition of occupant head posture data and in-vehicle environmental parameters via data acquisition module;

[0081] S2. Establish a motion sickness risk assessment model and input the occupant's head posture data into the model to output a motion sickness risk score in real time.

[0082] S3. Compare the motion sickness risk score with the set threshold to quantify and classify the motion sickness risk of the occupants and obtain the current motion sickness risk level of the occupants.

[0083] S4. The anti-motion sickness device automatically matches the intervention mode according to the motion sickness risk level and activates the progressive multi-dimensional intervention measures corresponding to each mode.

[0084] S5. Calculate the cumulative impact of the current vehicle to match the initial intervention duration, and dynamically update the actual intervention time;

[0085] S6. Through a dual-layer fuzzy PID control algorithm, the control gain and execution parameters of the anti-dizziness device are dynamically corrected, and the control rules are self-learned and corrected by combining the weighted comprehensive error integral.

[0086] The occupant head posture data in step S1 includes longitudinal acceleration, lateral acceleration, vertical acceleration, roll rate, pitch rate, and yaw rate of the occupant's head; the in-vehicle environmental parameters include CO2 concentration, PM2.5 concentration, temperature, and humidity.

[0087] In this embodiment, the motion sickness risk assessment model, based on the six-degree-of-freedom subjective vertical conflict model, introduces a dynamic adjustment mechanism of environmental stress factors on the human vestibular tolerance threshold. The specific calculation process includes the following steps:

[0088] S2.1 Calculating the Perception Conflict Vector Error: Using the otolith organ and semicircular canal dynamic model, the occupant head posture data obtained in step S1 is processed to calculate the conflict error between the actual perceived vertical vector and the subjective vertical vector of the inner ear vestibular system, denoted as ||Δv||. Specifically, the semicircular canal dynamic model is an existing technology and a relatively common dynamic model, frequently used in motion sickness assessment models involving the otolith organ and semicircular canals. The OTO and SCC in the provided papers correspond to the otolith organ and semicircular canals, respectively: 1) A Digital Human Model for Symptom Progression of Vestibular Motion Sickness based on Subjective Vertical Conflict Theory; 2) Computation of the Vestibula-Ocular Reflex for eye closure based on the 6DOF-SVC model; 3) Research on Motion Sickness Modeling and Objective Evaluation Methods for Car Occupant Motion Sickness; 4) Research on Trajectory Planning and Control Methods for Autonomous Vehicles Considering Motion Sickness; This model is mentioned in all of the above literature.

[0089] S2.2 Calculation of environmental stress factor E stress This function extracts the deviations of the current in-vehicle CO2 concentration, PM2.5 concentration, temperature, and humidity from preset comfort thresholds. Considering the different weights of various environmental factors in inducing motion sickness, a weighted normalization method is used to calculate the environmental stress factor.

[0090]

[0091] In the formula, j=1,2,3,4 correspond to the four indicators of CO2, PM2.5, temperature, and humidity inside the vehicle, respectively, and X j These represent the real-time values ​​of CO2, PM2.5, temperature, and humidity inside the vehicle, respectively; X target For comfort target value; X limit This is the tolerance limit; w j The dizziness-inducing weighting coefficients for each indicator (w in this example) j Values: CO2 = 0.3, PM2.5 = 0.25, Temperature = 0.25, Humidity = 0.2). This factor The larger the value, the greater the physiological stress on the human body from the environment.

[0092] S2.3. Based on the physiological principle that the harsher the environment, the lower the human body's tolerance, the environmental stress factor E is utilized. stress The half-criteria threshold parameter b of the Hill function is dynamically adjusted to obtain the real-time dynamic threshold b':

[0093]

[0094] In the formula, b is the basic pathogenicity threshold under standard conditions; λ is the environmental sensitivity attenuation coefficient (range 0~1); b' is the corrected real-time pathogenicity threshold.

[0095] S2.4 Substitute the dynamic threshold b' into the Hill function to perform a nonlinear mapping on the perceived conflict vector error ||Δv||.

[0096]

[0097] In the formula, h is the sensory conflict conversion amount; n is the nonlinear shape index (n=2 in this embodiment).

[0098] The motion sickness probability (MSI) is calculated by inputting h into the motion sickness transfer function model, passing it through a first-order inertial element, and then quantizing it to calculate the final motion sickness index (MSI) (range 0~100):

[0099]

[0100] In the formula, P is the gain coefficient, τ1 is the time constant, and s is the Laplace operator.

[0101] The process of comparing motion sickness risk scores with set thresholds is as follows:

[0102] Based on the current MSI value, it is quantitatively classified into four risk levels: no risk, low risk, medium risk, and high risk. No risk does not warrant a matching intervention model; low risk triggers primary intervention; medium risk triggers intermediate intervention; and high risk triggers advanced intervention.

[0103] In this embodiment, step S3 specifically includes:

[0104] Step S3 specifically includes setting a four-level risk threshold based on the MSI value in this embodiment to adapt to the intervention needs in different scenarios. The grading rules are as follows:

[0105] MSI < 5 indicates no risk; 5 ≤ MSI < 10 indicates low risk; 10 ≤ MSI ≤ 30 indicates medium risk; MSI > 30 indicates high risk.

[0106] Based on the current MSI value, it is quantitatively classified into four risk levels: no risk, low risk, medium risk, and high risk. No risk does not warrant a matching intervention model; low risk triggers primary intervention; medium risk triggers intermediate intervention; and high risk triggers advanced intervention.

[0107] The threshold values ​​for classifying motion sickness risk levels are merely example values ​​adapted to a specific vehicle model. In practical applications, these values ​​can be calibrated and adjusted through real-vehicle experiments based on the target vehicle's dynamic characteristics and the motion sickness sensitivity of the target occupant group, to meet the anti-motion sickness needs in different scenarios.

[0108] In this embodiment, step S4 specifically includes matching a three-level intervention model according to the risk level. (See attached...) Figure 3 As shown, each level of intervention is independent and progressive; initiating a higher-level intervention does not terminate a lower-level intervention, ensuring continuity of intervention effects.

[0109] (1) Primary intervention

[0110] When the ventilation function is activated, the fuzzy PID algorithm mainly uses the core inference of the air quality inference layer. The wind direction (D) and air volume (V) are adjusted in a routine manner according to the deviation of PM2.5 and CO2, so as to stabilize the indicators within the target threshold and suppress early discomfort.

[0111] (2) Intermediate intervention

[0112] In addition to ventilation, the fragrance release unit is automatically triggered. A fuzzy logic-based scene adaptation layer intervenes. The algorithm begins to smoothly adjust the center value of the airflow direction (D), shifting the airflow towards the occupant's head; the airflow volume (V) is moderately increased according to the rate of motion sickness. Simultaneously, the control module issues short, intermittent beeps as a warning. This collaboratively alleviates the physiological conflicts in the initial stages of motion sickness.

[0113] (3) Advanced intervention

[0114] Ventilation mode enters motion sickness suppression mode. The algorithm significantly reduces the scaling factor K based on the MSI level. p By reducing the intensity of adjustment, the actuator's movement is ensured to be extremely smooth, avoiding air pressure or sound field fluctuations caused by sudden changes in wind speed. Hard constraints are applied to ensure that single airflow adjustment is ≤1 level and single airflow direction adjustment is ≤30°, preventing violent mechanical movements from triggering secondary sensitivity in the occupant's vestibular system. An audible and visual alarm unit is used to remind the driver to drive cautiously, reducing physical impact on the vehicle.

[0115] When the risk level is no risk, no intervention is performed; when the risk level is low risk, the corresponding intervention mode is primary intervention; when the risk level is medium risk, the corresponding intervention mode is intermediate intervention; when the risk level is high risk, the corresponding intervention mode is advanced intervention.

[0116] In this embodiment, as shown in the appendix Figure 4 The step S5 described herein specifically includes the following steps:

[0117] S5.1 Calculation of cumulative head impact on occupant, the formula for calculating single impact is as follows:

[0118]

[0119] In the formula, P i Represents occupant head posture data, ω i λ is the weighting factor for the corresponding parameter. i These are the dimensional conversion coefficients.

[0120] The cumulative impact amount is calculated using a 5-second statistical time window as follows:

[0121]

[0122] S5.2 Matching the initial duration. In this example, the initial duration matching rule is as follows: When I total The initial intervention duration is 10 seconds when I1 ≤ I1, and when I1 < I total When I2 < I2, the initial intervention duration is 30 seconds; when I2 ≤ I total <I3 is 60s, I total ≥I3 is 120s.

[0123] S5.3 determines the magnitude of the duration decay factor, which is determined based on the motion sickness rate of change ΔMSI. The formula for calculating ΔMSI is as follows:

[0124]

[0125] When ΔMSI is less than 0, γ is 0.05s. -1 When ΔMSI is between 0 and 0.01, γ is 0.01s. -1 When ΔMSI is greater than 0.01, γ is 0.

[0126] S5.4 Calculation of actual intervention duration, the formula for calculating actual intervention duration is as follows: In the formula, T base t represents the initial intervention duration, γ is the duration decay factor, and t inter The intervention has been in effect for a certain period of time; after the intervention is initiated, the remaining intervention duration is calculated in real time. The duration decay factor is determined by the motion sickness rate of change ΔMSI. The larger the motion sickness rate of change ΔMSI, the slower the duration decays, and vice versa. Over-intervention should be avoided. The duration decay factor γ is determined based on the motion sickness rate of change ΔMSI; the intervention duration tinter is calculated, and timing begins from the start of the current intervention.

[0127] S5.5 will read the current cumulative impact amount and re-match the intervention duration at the end of the previous intervention to prepare for the calculation of the intervention duration for the next intervention cycle.

[0128] The intervention must be terminated immediately if any of the following conditions are met: First, the remaining intervention duration is less than or equal to T. actual ≤0; secondly, the cumulative impact amount Itotal The motion sickness level is reduced to <I1; and the motion sickness level is no risk.

[0129] In this embodiment, as shown in the appendix Figure 6 As shown, step S6 specifically includes:

[0130] S6.1 Input Variable Acquisition and Quantization

[0131] Calculate the deviations of each input variable from the target thresholds and the rate of change of deviations for PM2.5 and CO2, where the target thresholds for PM2.5 are ≤35 μg / m³, CO2 are ≤1000 ppm, temperature is 24℃, humidity is 50%RH, and MSI low-risk threshold is <5; and quantize all inputs to a unified fuzzy universe of discourse [-3,3]. The quantization formula is as follows:

[0132]

[0133] ΔX represents the actual deviation value of each input variable. max The values ​​are PM2.5 = 100 μg / m³, CO2 = 2000 ppm, temperature = 10℃, humidity = 50% RH, and MSI = 100. The sampling period is uniformly set to 1 second. The output variables include two parts: PID parameter correction and vehicle air conditioning control parameters.

[0134] PID parameter corrections: including proportional coefficient correction ΔKp, integral coefficient correction ΔKi, and derivative coefficient correction ΔKd, all mapped to the [-1,1] fuzzy domain to meet the parameter fine-tuning requirements in automotive scenarios;

[0135] The air conditioning operation adjustment parameters include air volume adjustment ΔV, air direction adjustment ΔD, and circulation mode adjustment ΔC, all of which are mapped to the fuzzy universe of discourse [-n,n].

[0136] S6.2 Two-layer fuzzy structure inference

[0137] In step 6.2, the two-layer fuzzy structure is as follows:

[0138] Air quality inference layer: based on PM2.5 deviation (e PM ), PM2.5 deviation change rate (ec PM CO2 deviation (e) CO2 ), CO2 deviation change rate (ec CO2 Using ) as input, the basic PID parameter correction and basic execution adjustment are inferred through the core fuzzy rules;

[0139] Scene adaptation correction layer: based on MSI, temperature deviation (e T ), humidity deviation (e HUsing this as input, the MSI and temperature and humidity are corrected, and the results of the air quality inference layer are further intervened based on real-time feedback and environmental comfort.

[0140] For different air conditioning adjustment targets, different initial PID parameters (Kp0, Ki0, Kd0) are set for the adjustment of air volume, air direction, and circulation mode.

[0141] Based on the adjustment requirements of different scenarios, formulate independent PID parameter correction rules:

[0142] Based on the fuzziness level of MSI, output the corresponding correction coefficient (k). p k i k d k v k d0 k c When MSI is in PM or PB mode, set k p ≤0.5, k d ≥1.5, k v ≤0.6, achieving both weakening of regulation intensity and improvement of regulation smoothness.

[0143] The temperature and humidity correction process is as follows: the basic adjustment values ​​ΔV, ΔD, and ΔC are fine-tuned by ±0.5 levels, ±0.5°, and ±0.5%, respectively. The fine-tuning process does not interfere with the core air quality regulation logic.

[0144] The preset fuzzy rule base contains 49 core rules for reasoning, outputting the basic PID parameter correction amount (ΔK). p1 ΔK i1 ΔK d1 ) and the basic execution adjustment (ΔV, ΔD, ΔC).

[0145] The basic execution adjustment values ​​(ΔV, ΔD, ΔC) correspond to the following physical quantity ranges: air volume 1-5, air direction 0-180°, and circulation mode 0-100%.

[0146] S6.3 Parameter Synthesis and Execution

[0147] The independent correction rule for PID parameters in the two-layer fuzzy PID control algorithm is as follows:

[0148] (1) When in a high MSI scenario, reduce the proportional gain to decrease the adjustment strength, weaken the integral action, avoid over-adjustment, increase the derivative gain, and enhance the smoothness and predictive ability of the adjustment; ΔK p The value range is NS~ZO, ΔK i The value range is NS~ZO, ΔK d The value range is PS to PM;

[0149] (2) When the in-vehicle air quality is severely substandard, increase the proportional coefficient to accelerate intervention, take an appropriate integral value, eliminate steady-state error and appropriately differentiate to suppress overshoot; ΔK p The value range is PS~PB, ΔK i The value range is ZO~PS, ΔK d The value range is ZO~PS;

[0150] (3) When in a steady-state scenario, maintain the proportional coefficient to ensure control stability, incrementally increase the integral action to eliminate residual steady-state error, and maintain the derivative coefficient to avoid introducing noise interference; ΔK p =ZO、ΔK i =PS、ΔK d =ZO;

[0151] The final PID parameters are obtained by superimposing the initial parameters and the correction factor. The update formula is as follows:

[0152]

[0153]

[0154]

[0155] In the formula, ΔK p ΔK i ΔK d k is the final correction value for the output of the two-layer inference. p k i k d Correction coefficients for PID parameters output by the scene adaptation correction layer.

[0156] The centroid method is used to defuzzify the fuzzy quantity output by fuzzy inference to obtain the precise adjustment quantity (ΔK). p ΔK i ΔK d All input and output variables are divided into multiple fuzzy sets (ΔV, ΔD, ΔC); the membership functions of each fuzzy set are triangularly distributed.

[0157] The centroid method is used to defuzzify the fuzzy quantities output by fuzzy inference to obtain the precise adjustment quantity. The defuzzification formula is as follows:

[0158]

[0159] In the formula, μ i Let x be the membership degree of the i-th rule, and take the minimum membership degree of each input variable under that rule. i Output the center value of the fuzzy set for the i-th fuzzy rule, where n is the number of fuzzy rules activated in the current scene.

[0160] Incremental PID is used to calculate the execution increment. The formula for calculating the execution increment is as follows:

[0161]

[0162] In the formula, e(t) is the deviation at the current time, and e(t-1) and e(t-2) are the deviations at the previous two times.

[0163] U(t) is the current control variable, and the formula for calculating U(t) is:

[0164]

[0165] In the formula, U(t-1) is the control quantity at the previous moment, to avoid control quantity jumps caused by MCU abnormalities;

[0166] The incremental control output of the incremental PID controller is converted into a physical quantity adjustment value for the air conditioning actuator, and vehicle-mounted scenario adaptation constraints are applied:

[0167] 1. When MSI ≥ 30, implement flexible risk control constraints: airflow level ≤ 3, wind direction angle ≥ 100°, single airflow adjustment range ≤ 1 level;

[0168] 2. The external air circulation rate in the recirculation mode is ≤80% to avoid excessive wind noise at high speeds;

[0169] 3. The single wind direction adjustment range is ≤30° and the single cycle mode adjustment range is ≤20%, to avoid overload operation of the actuator and reduce the stimulation of the adjustment action on the occupants.

[0170] S6.4 Execution Intervention and Self-Learning Optimization

[0171] As attached Figure 6 As shown, the anti-motion sickness execution module receives and executes the generated control commands, runs a self-learning correction strategy to continuously optimize performance, and calculates the weighted comprehensive error integral J over a preset time T (e.g., 10 seconds) to evaluate the air quality compliance effect and motion sickness suppression effect. The calculation formula is as follows:

[0172]

[0173] In the formula, ω PM ω CO2 ω MSI These are the evaluation weight coefficients for each indicator, e PM (t) and e CO2 ΔMSI(t) reflects the degree to which air quality meets standards, and ΔMSI(t) reflects the inhibitory effect of the intervention on motion sickness.

[0174] When J > Jtarget and the deviation directions are consistent, it is determined that the adjustment force is insufficient, and the output center value xi is shifted in the positive direction. The correction formula is as follows:

[0175]

[0176] In the formula, η is the self-learning rate factor. This represents the average deviation within the cycle. It's the base output weight for automatically adjusting the airflow or the proportion of external circulation.

[0177] When J > Jtarget and the deviations alternate frequently, it is determined to be over-adjustment or oscillation. The fuzzy center value xi is then compressed and corrected using the following formula:

[0178]

[0179] In the formula, β is the smoothing correction coefficient, and MSIlevel is the weight of the current motion sickness risk level. By lowering the output center value xi, the air conditioning adjustment action is made smoother, prioritizing the reduction of stimulation to the occupant's vestibular system.

[0180] When J≤Jtarget for multiple consecutive statistical periods, the algorithm is considered to have converged, rule base correction is stopped, and the current optimal control characteristics are maintained.

[0181] This invention provides an adaptive air quality control device for in-vehicle systems that considers the risk of motion sickness. The device mainly includes:

[0182] Data acquisition module: Equipped with a nine-axis inertial measurement unit and an environmental sensor group, it collects occupant head posture data and in-vehicle environmental parameters in real time.

[0183] Decision control module: Receives data from the acquisition module, performs coordinate transformation, data quantification, motion sickness risk assessment model calculation and anti-motion sickness control algorithm, and outputs motion sickness risk level, intervention duration and anti-motion sickness execution instructions.

[0184] Anti-dizziness execution module: includes intelligent risk control unit, fragrance release unit, and sound and light alarm unit, which executes anti-dizziness commands.

[0185] User interaction module: Provides three working modes: automatic, manual, and preventative. It allows occupants to set intervention modes, intervention intensity, and timed start commands via a display screen or remote terminal.

[0186] Storage module: Saves motion sickness assessment model parameters, occupant head posture data, in-vehicle environment parameters, occupant motion sickness level, historical data, and rule base correction logs.

[0187] The device frame is attached. Figure 7 As shown, the modules interact with each other via the vehicle's CAN bus or a dedicated communication link.

[0188] The data acquisition module is used to acquire occupant head posture data and in-vehicle environmental parameters in real time. It typically includes:

[0189] Nine-axis inertial measurement unit (IMU): Acquires occupant head posture data at a preset sampling period, specifically including longitudinal acceleration, lateral acceleration, vertical acceleration, roll rate, pitch rate, and yaw rate.

[0190] Environmental sensor group: including PM2.5 sensor, CO2 sensor, temperature and humidity sensor, used to synchronously collect environmental parameters such as PM2.5 concentration, CO2 concentration, temperature and humidity inside the vehicle.

[0191] The decision control module is typically implemented by an in-vehicle domain controller or a dedicated microcontroller unit (MCU). The decision control module receives raw data, performs motion sickness risk assessment, risk classification, dynamic calculation of intervention duration, two-layer fuzzy PID algorithm operation and self-learning optimization, and finally outputs motion sickness risk level, intervention duration and anti-motion sickness control command.

[0192] The anti-dizziness execution module is responsible for converting control commands into physical actions, mainly including:

[0193] Intelligent air control unit: Adjusts air volume, air direction and internal / external circulation mode through the vehicle air conditioning control unit.

[0194] Fragrance release unit: controls the concentration and release duration of specific fragrances (such as peppermint and ginger extracts).

[0195] Audible and visual alarm unit: generates visual or auditory cues of varying intensities based on the risk level.

[0196] The user interaction module provides a display screen or mobile terminal application interface, supporting occupants to select three working modes: automatic, manual, and preventative, and to set personalized parameters.

[0197] The storage module is used to save motion sickness assessment model parameters, time-series data of the vehicle and environment, personalized historical records of occupants, and rule base correction logs generated by algorithm self-learning.

[0198] The anti-motion sickness device has three motion sickness intervention modes: automatic mode, manual mode, and prevention mode. The descriptions of each mode are as follows:

[0199] 1. Automatic Mode: Based on real-time parameters acquired by the data acquisition module, the system outputs a Motion Sickness Risk Assessment (MSI) score through a motion sickness risk assessment model. The device automatically matches the corresponding intervention mode according to the risk level of the MSI, and uses a dual-layer fuzzy PID algorithm to adjust the airflow, direction, and circulation mode in real time, while also adjusting the cumulative impact amount I. total The actual intervention duration T is dynamically updated with the motion sickness change rate ΔMSI.actual .

[0200] 2. Manual Mode: Occupants can interact via the device's display screen based on their subjective perception of discomfort. This mode allows occupants to bypass system tiers and manually select specific anti-motion sickness intervention modes, adjust fragrance release concentration, and set intervention duration. In this mode, flexible risk control constraints still apply.

[0201] 3. Prevention Mode: Passengers can remotely preset the mode using their mobile phone via Bluetooth or an IoT device. This mode supports pre-trip pre-treatment, activating a primary intervention mode in advance to optimize in-vehicle air quality, release fragrances, and eliminate odors and other triggering factors, achieving predictive optimization of in-vehicle environmental parameters.

Claims

1. An adaptive control method for in-vehicle air quality considering motion sickness risk, characterized in that, Includes the following steps; S1. Real-time acquisition of occupant head posture data and in-vehicle environmental parameters via data acquisition module; S2. Establish a motion sickness risk assessment model and input the occupant's head posture data into the motion sickness risk assessment model to output motion sickness risk scores in real time; S3. Compare the motion sickness risk score with the set threshold to quantify and classify the motion sickness risk of the occupants and obtain the current motion sickness risk level of the occupants. S4. The anti-motion sickness device automatically matches the intervention mode according to the motion sickness risk level and activates the progressive multi-dimensional intervention measures corresponding to each mode. S5. Calculate the cumulative impact of the current vehicle to match the initial intervention duration, and dynamically update the actual intervention time; S6. Through a dual-layer fuzzy PID control algorithm, the control gain and execution parameters of the anti-dizziness device are dynamically corrected, and the control rules are self-learned and corrected by combining the weighted comprehensive error integral.

2. The adaptive control method for in-vehicle air quality considering motion sickness risk according to claim 1, characterized in that, The occupant head posture data in step S1 includes the acceleration of the head in the longitudinal, lateral, and vertical directions, as well as the angular velocity around each axis; the in-vehicle environmental parameters include the CO2 concentration, PM2.5 concentration, temperature, and humidity inside the vehicle.

3. The adaptive control method for in-vehicle air quality considering motion sickness risk according to claim 1, characterized in that: The motion sickness risk assessment model described in S2 adopts an improved six-degree-of-freedom subjective vertical conflict model based on environmental tolerance correction. It calculates environmental stress factors and dynamically corrects the pathogenic threshold of the model, thereby integrating in-vehicle environmental parameters to jointly assess motion sickness risk. The specific calculation includes the following steps: S21: Using the otolith organ and semicircular canal dynamic model, process the acquired occupant head posture data, calculate the conflict error between the actual perceived vertical vector and the subjective vertical vector of the inner ear vestibular system, denoted as ||Δv||. S22: Extract the deviations of the current in-vehicle CO2 concentration, PM2.5 concentration, temperature, and humidity from the preset comfort thresholds; calculate the environmental stress factor E using a weighted normalization method. stress The calculation formula is as follows: In the formula, X j These represent the real-time values ​​of CO2, PM2.5, temperature, and humidity inside the vehicle, respectively; X target,j X represents the target comfort value for the corresponding indicator. limit,j This represents the tolerance limit value for the corresponding indicator; w j This represents the dizziness-inducing weighting coefficient for the corresponding indicator. S23: Based on the physiological principle that the harsher the environment, the lower the human body's tolerance, utilizing environmental stress factor E stress The pathogenicity threshold parameter of the Hill function is dynamically adjusted to obtain the real-time dynamic threshold b', calculated using the following formula: In the formula, b is the baseline pathogenicity threshold under standard conditions; λ is the environmental sensitivity attenuation coefficient. S24: Substitute the real-time dynamic threshold b' into the Hill function, perform nonlinear mapping on the sensory conflict vector error ||Δv|| to obtain the sensory conflict conversion quantity h, and input h into the motion sickness transfer function model. After quantization, the final motion sickness index MSI is calculated. The calculation formulas are as follows: In the formula, n is the nonlinear shape exponent, P is the gain coefficient, τ1 is the time constant, and s is the Laplace operator.

4. The adaptive control method for in-vehicle air quality considering motion sickness risk according to claim 1, characterized in that: The progressive multidimensional intervention measures described in S4 include: S41: Primary intervention: Activate ventilation function, and make routine adjustments to air direction and air volume according to the deviation of PM2.5 and CO2 to stabilize the indicators within the target threshold and suppress early discomfort; S42: Intermediate Intervention: Based on ventilation, the fragrance release unit is automatically triggered to smoothly adjust the center value of the airflow direction, causing the airflow to shift towards the occupant's head; the airflow is appropriately increased according to the motion sickness rate; at the same time, the sound and light alarm unit emits short beeps for intermittent warnings, working together to alleviate the physiological conflicts in the early stages of motion sickness. S43: Advanced Intervention: The proportional coefficient is significantly reduced according to the level of motion sickness. The actuator movement is made extremely smooth by weakening the adjustment intensity. At the same time, the audible and visual alarm unit continuously sounds an alarm to remind the driver to drive carefully and reduce the physical impact on the vehicle.

5. The adaptive control method for in-vehicle air quality considering motion sickness risk according to claim 3, characterized in that, In S5, the cumulative impact amount is obtained by weighted calculation of occupant head posture data; the actual intervention duration is matched to the initial value based on the cumulative impact amount and dynamically decayed according to the rate of change of motion sickness risk. The calculation steps are as follows: Single impact I single (t) represents the impact received by the head at the current moment, and the cumulative impact amount I. total (t) Using a preset time T as the statistical time window, the single impact quantity I... single (t) is calculated using the practice-weighted cumulative method, and the formulas are as follows: In the formula, P i Represents occupant head posture data; ω i The weighting factor represents the corresponding parameter; λ i Representative dimensionless conversion coefficients; Δt is the basic time step; Preset mild, moderate, and severe cumulative impact thresholds, and initial intervention duration T of the device. base Matching is performed based on a comparison of the cumulative impact amount and a threshold; after intervention is initiated, the remaining intervention duration is calculated in real time starting from the time the intervention is initiated. The duration decay factor γ is determined by the motion sickness rate of change ΔMSI. The larger the motion sickness rate of change ΔMSI, the slower the duration decays, and vice versa. The formula for calculating ΔMSI is as follows: Actual intervention time T of the device actual The formula for calculating (t) is: Intervention shall be terminated immediately if any of the following conditions are met: remaining intervention duration ≤ 0, cumulative impact amount is below the first threshold, or the motion sickness level is risk-free.

6. The adaptive control method for in-vehicle air quality considering motion sickness risk according to claim 1, characterized in that, In S6, the input variables of the dual-layer fuzzy PID control algorithm include PM2.5 deviation, CO2 deviation, temperature deviation, humidity deviation, and motion sickness index (MSI). The output variables include the PID correction value used to dynamically adjust the controller parameters, and the adjustment parameters used to control the air conditioning actuator; The dual-layer fuzzy PID control algorithm adopts a dual-layer fuzzy inference architecture, specifically including: Air quality inference layer: Taking PM2.5 deviation, PM2.5 deviation change rate, CO2 deviation, and CO2 deviation change rate as inputs, the basic PID parameter correction amount and basic execution adjustment amount are inferred through core fuzzy rules; Scene adaptation correction layer: Taking MSI, temperature deviation and humidity deviation as input, the output correction coefficient dynamically corrects the air quality inference layer results. Among them, the MSI correction outputs the correction coefficient according to its fuzziness level.

7. The adaptive control method for in-vehicle air quality considering motion sickness risk according to claim 6, characterized in that, The dual-layer fuzzy PID control algorithm has PID parameter correction rules adapted to the scenario, including: (1) In scenarios with high risk of motion sickness, reduce the proportional and integral effects and enhance the differential effect to improve smoothness; (2) In scenarios where air quality is severely exceeded, the proportional and integral effects are enhanced to accelerate regulation; (3) In steady-state scenarios, the proportional and differential coefficients are kept stable, and the integral action is increased to eliminate steady-state error.

8. The adaptive control method for in-vehicle air quality considering motion sickness risk according to claim 1, characterized in that, In S6, constraints are imposed on the control of the actuator, including: when in a high motion sickness risk level, limiting the air volume level, air direction angle, and single adjustment range; and limiting the external circulation ratio and single mode adjustment range. The self-learning optimization strategy is as follows: calculate the weighted comprehensive error integral J at a preset period to evaluate the control effect; and adaptively correct the output center value of the fuzzy control according to the deviation between J and the target value and the deviation change pattern until the algorithm converges.

9. An adaptive control device for in-vehicle air quality considering motion sickness risk, used to execute the adaptive control method for in-vehicle air quality considering motion sickness risk according to any one of claims 1-8, characterized in that... include: Data acquisition module: used to collect occupant head posture data and in-vehicle environmental parameters in real time; Decision control module: used to run the motion sickness risk assessment model and the in-vehicle air quality adaptive control method that takes motion sickness risk into account, and output motion sickness risk level, intervention duration and control instructions; Anti-motion sickness module: Controls in-vehicle air quality, fragrance release, and audible and visual alarms based on the motion sickness risk level and anti-motion sickness execution commands; User interaction module: provides passengers with three working modes: automatic, manual, and preventative; Storage module: Used to store motion sickness model parameters, historical data, and learning logs.

10. The in-vehicle air quality adaptive control device considering motion sickness risk according to claim 9 is characterized in that: The device has three motion sickness intervention modes; Automatic mode: The device operates automatically based on input parameters and real-time motion sickness assessment results; Manual mode: Based on your own discomfort, manually select the specific anti-dizziness intervention mode, adjust the fragrance release concentration, and set the intervention duration; Prevention mode: Remotely preset using mobile phone Bluetooth or IoT terminal to activate environmental optimization before boarding.

Images