Domain control optimization method and system based on cockpit internal environment recognition

Abstract

The invention discloses a domain control optimization method and system based on cockpit internal environment recognition, and relates to the field of domain control optimization, and the method comprises the steps: building a dynamic correlation model of the change of real-time environment parameters outside a cockpit and the expected change of environment parameters in the cockpit; constructing a pre-adjustment optimization feedback loop, and carrying out feedback closed-loop primary adjustment on the environmental parameters in the cabin; performing unified modeling on the historical multi-source environment parameters, and establishing an environment situation recognition model in the cabin; based on the identified environment situation type and the corresponding historical data of the human physiological state, establishing an association relationship between the situation type and the human physiological state; the method comprises the following steps: predicting a human physiological state based on expected environmental parameter prediction in a cabin, environmental situation recognition and a correlation between a situation type and the human physiological state, adaptively generating a domain control strategy by screening historical domain control strategy records, and carrying out secondary adjustment on environmental parameters in the cabin.

Description

Technical Field

[0001] The present invention relates to the field of domain control optimization, and in particular to a domain control optimization method and system based on cabin environment recognition. Background Art

[0002] Smart cockpits usually adjust environmental parameters based on internal environmental sensors, but this approach ignores the dynamic impact of the external environment on the cabin environment. Factors such as temperature, humidity, and light in the external environment can significantly affect the stability of the cabin environment. In addition, traditional cabin environment physical data statically regulates the cabin environment, ignoring the feedback of human physiological data on environmental control needs. The human physiological state will produce different reactions in different environments. Therefore, the introduction of physiological data perception combined with the regulation of multimodal environmental data can achieve more personalized cabin environment management.

[0003] Therefore, how to realize a domain control optimization method that can adjust the environmental parameters inside the cabin in advance based on a dynamic correlation prediction model of the environment inside and outside the cabin, and at the same time perform secondary domain control optimization of the environmental parameters inside the cabin based on the correlation between environmental data and passengers' physiological data, so as to enhance the smart cockpit's ability to respond to environmental changes in advance, is an urgent problem that needs to be solved. Summary of the invention

[0004] In order to solve the above technical problems, a domain control optimization method based on cabin environment recognition is provided. This technical solution solves the problems raised in the above background technology.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is: A domain control optimization method based on cockpit environment recognition includes: Collect real-time environmental parameters inside and outside the cabin, and establish a dynamic correlation model between the real-time environmental parameter changes outside the cabin and the expected changes in the environmental parameters inside the cabin; Based on the deviation between the expected environmental parameters in the cabin and the ideal environmental parameters, a pre-adjustment optimization feedback loop is constructed to perform a closed-loop feedback adjustment of the environmental parameters in the cabin. The historical multi-source environmental parameters are uniformly modeled to establish an in-cabin environmental situation recognition model to identify the in-cabin environmental situation type and situation matching degree; Based on the identified environmental situation types and the corresponding historical data of human physiological states, establish the association relationship between situation types and human physiological states; The human physiological state is predicted based on the prediction of expected environmental parameters in the cockpit, environmental situation recognition, and the correlation between situation type and human physiological state. The domain control strategy is adaptively generated by screening historical domain control strategy records to make secondary adjustments to the environmental parameters in the cockpit.

[0006] Preferably, the collecting of real-time environmental parameters inside and outside the cabin and establishing a dynamic correlation model between changes in real-time environmental parameters outside the cabin and expected changes in environmental parameters inside the cabin specifically includes: Collect the external environment parameters of the cabin and the real-time environment parameters of the cabin at set time intervals to establish a synchronous data sequence of the internal and external environment status of the cabin; Preprocess and normalize the synchronous data sequence, and use the processed data as the initial input of the dynamic correlation model of the cabin internal and external environmental parameters; A regression model is used to correlate the changes in the external environment of the cabin with the response of the internal environment. A dynamic correlation model of the internal and external environment parameters of the cabin is established through a multivariate linear regression model. The sliding time window for training the dynamic correlation model of the cabin internal and external environmental parameters is set, and the synchronous data sequence of the cabin internal and external environmental states collected within the sliding time window is brought into the multivariate linear regression model to calculate and adjust the regression coefficient in real time.

[0007] Preferably, the constructing of a pre-adjustment optimization feedback loop based on the deviation between the expected environmental parameters in the cabin and the ideal environmental parameters, and performing a feedback closed-loop adjustment on the environmental parameters in the cabin specifically includes: The system uses a proportional-integral-derivative controller as the core controller of the feedback loop. It dynamically adjusts the cabin's internal environmental parameters based on the real-time deviation between the ideal environmental parameters and the expected environmental parameters in the cabin. Specifically: Set the ideal environment parameters in the cockpit; The current real-time external environment parameters of the cockpit are brought into the multivariate linear regression model to obtain the expected internal environment parameters of the cockpit, which are used to describe the predicted values ​​of the environment parameters at the next moment; By evaluating the immediate response to the real-time deviation between the expected environmental parameters and the ideal environmental parameters inside the cockpit, the cumulative impact of historical deviations, and the changing trend of future deviations, the proportional gain, integral gain, and differential gain corresponding to the feedback adjustment are determined respectively; The feedback formula is established through proportional gain, integral gain and differential gain to determine the adjustment range of various environmental parameters in the cockpit. The feedback formula is: ; In the formula, is the feedback adjustment amplitude of the jth environmental parameter inside the cabin at time t, is the real-time deviation between the jth expected environmental parameter and the ideal environmental parameter inside the cockpit at time t, , , They are proportional gain, integral gain and differential gain parameters respectively; By calculating the adjustment range of each environmental parameter in real time, various environmental parameters in the cabin are adjusted in real time, and the adjustment process is marked as one adjustment; When it is detected that the real-time adjustment amplitude of each environmental parameter controlled by the feedback loop is less than the set minimum adjustment amplitude threshold and tends to be stable, an adjustment is ended.

[0008] Preferably, the unified modeling of historical multi-source environmental parameters to establish a cockpit environment situation recognition model for identifying the cockpit environment situation type and situation matching degree specifically includes: Set the time window for collecting historical multi-source environmental parameters; After one adjustment is completed, the historical multi-source environmental parameters inside the cockpit within the historical multi-source environmental parameter collection time window are obtained, and the features of the historical multi-source environmental parameters are extracted based on the parameter type; Based on the multimodal data fusion model, the historical multi-source environmental parameters are uniformly modeled and the cabin environmental situation scores corresponding to the historical multi-source environmental parameters are calculated; Combine multi-source environmental parameter features with environmental context scores to establish a historical environmental context dataset, and set the feature center of all parameter feature types in the historical environmental context dataset; Based on the clustering algorithm, an association recognition model of environmental parameter feature classification corresponding to environmental situation type is established to identify the environmental situation type and situation matching degree in the cabin.

[0009] Preferably, the establishing of the association relationship between the situation type and the human physiological state based on the identified environmental situation type and the corresponding human physiological state historical data specifically includes: Obtain the human physiological state parameters collected at the same time corresponding to each data record in the historical environmental situation data set; Based on the environmental situation type recognition results of the historical environmental situation data set, the correlation between environmental situation type and human physiological state is determined through a regression analysis model; Continue to adjust the size of the historical environmental scenario data set, and measure the real-time fit of the regression analysis model by calculating the model's coefficient of determination, until it is detected that the accuracy of the regression analysis model evaluation reaches the preset standard and the adjustment is terminated.

[0010] Preferably, the predicting of human physiological state based on the expected environmental parameter prediction in the cockpit, environmental context recognition, and the correlation between context type and human physiological state, adaptively generating domain control strategy by screening historical domain control strategy records, and performing secondary adjustment on the environmental parameters in the cockpit specifically includes: Obtain the real-time environmental parameters of the cabin's exterior and interior, bring the real-time environmental parameters into the dynamic correlation model of the cabin's interior and exterior environmental parameters, and obtain the expected cabin interior environmental parameters; The clustering algorithm model is used to identify the expected environmental scenario type and scenario matching degree at the next moment corresponding to the expected environmental parameters inside the cabin; The expected situation type and situation matching degree of the next moment are introduced into the regression analysis model of the correlation between the situation type and the human physiological state to obtain the expected human physiological state parameters of the next moment; Through similarity calculation, identify the historical human physiological state parameter records whose similarity with the expected human physiological state parameters at the next moment exceeds the set range, and then match and obtain the domain control policy records corresponding to the historical human physiological state parameters in the historical database; After screening out similar historical human physiological state parameter records, dynamically weight the domain control strategy corresponding to the historical human physiological state parameter records based on the feedback of historical human physiological state parameter changes; All domain control policies are weighted and averaged to calculate the final weight of the domain control policy. The cabin environmental parameters corresponding to the domain control policy are adjusted secondary according to the distribution of the final weight of the domain control policy.

[0011] Furthermore, a domain control optimization system based on in-cabin environment recognition is proposed, which is used to implement the domain control optimization method based on in-cabin environment recognition as described above, including: An environmental parameter change prediction module, which establishes a synchronous data sequence of the internal and external environmental states of the cabin, uses a regression model to associate the changes in the external environment of the cabin with the response of the internal environment, and establishes a dynamic correlation model of the internal and external environmental parameters of the cabin through a multivariate linear regression model; A domain control primary adjustment module, wherein the domain control primary adjustment module uses a proportional-integral-differential controller as a core controller of a feedback loop, establishes a feedback formula through proportional gain, integral gain and differential gain to determine the adjustment amplitude of various environmental parameters in the cabin, and adjusts the internal environmental parameters of the cabin once based on the real-time deviation value between the ideal environmental parameters and the expected environmental parameters in the cabin, until the real-time adjustment amplitude of various environmental parameters controlled by the feedback loop is less than the set minimum adjustment amplitude threshold and tends to be stable; The environmental situation type recognition module uniformly models the historical multi-source environmental parameters based on the multimodal data fusion model, calculates the cabin environmental situation scores corresponding to the historical multi-source environmental parameters, and establishes an association recognition model corresponding to the environmental situation type of the environmental parameter feature classification based on the clustering algorithm, which is used to identify the environmental situation type in the cabin and the situation matching degree.

[0012] A human physiological state prediction module, which determines the correlation between the environmental situation type and the human physiological state through a regression analysis model based on the environmental situation type recognition result of the historical environmental situation data set; A domain control secondary adjustment module predicts the human physiological state at the next moment based on the prediction of expected environmental parameters in the cabin, environmental situation recognition, and the correlation between situation types and human physiological states. Based on the feedback of historical human physiological state parameter changes, the module dynamically sets weights for the domain control strategies corresponding to the historical human physiological state parameter records and performs secondary adjustments to the cabin environmental parameters corresponding to the domain control strategies.

[0013] Compared with the prior art, the present invention has the following beneficial effects: A dynamic correlation model is established between the real-time environmental parameter changes outside the cockpit and the expected changes in environmental parameters inside the cockpit. A pre-adjustment optimization feedback loop is constructed to perform a closed-loop feedback adjustment on the environmental parameters inside the cockpit. Historical multi-source environmental parameters are uniformly modeled to establish an environmental context recognition model inside the cockpit. Based on the identified environmental context types and the corresponding historical data on human physiological states, the relationship between context types and human physiological states is established. Based on the expected environmental parameter prediction inside the cockpit, environmental context recognition, and the relationship between context types and human physiological states, the human physiological state is predicted. By screening historical domain control strategy records, a domain control strategy is adaptively generated to make secondary adjustments to the environmental parameters inside the cockpit.

[0014] It realizes the advance adjustment of the environmental parameters inside the cabin based on the dynamic correlation between the environment inside and outside the cabin, and then performs secondary domain control optimization of the environmental parameters inside the cabin based on the correlation between environmental data and passengers' physiological data, so as to achieve continuous intelligent adaptive optimization of the cabin environment and provide higher comfort and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of a domain control optimization method based on cabin environment recognition of the present invention; Figure 2 A flow chart of a dynamic correlation model for collecting real-time environmental parameters inside and outside the cabin and establishing changes in real-time environmental parameters outside the cabin and expected changes in environmental parameters inside the cabin for the present invention; Figure 3 The flowchart of the present invention is to construct a pre-adjustment optimization feedback loop based on the deviation between the expected environmental parameters in the cabin and the ideal environmental parameters, and to perform a feedback closed-loop adjustment on the environmental parameters in the cabin; Figure 4 The present invention unifies historical multi-source environmental parameters into a model, establishes a cockpit environment situation recognition model, and is used to identify the cockpit environment situation type and situation matching degree flow chart; Figure 5 A flowchart of establishing an association relationship between a situation type and a human physiological state based on the identified environmental situation type and the corresponding human physiological state historical data of the present invention; Figure 6A flowchart of the present invention for adaptively generating a domain control strategy by screening historical domain control strategy records and performing secondary adjustment on the environmental parameters in the cabin; Figure 7 The present invention is a structural schematic diagram of a domain control optimization system based on cabin environment recognition. DETAILED DESCRIPTION

[0016] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0017] Reference Figure 1 As shown, a domain control optimization method based on cabin environment recognition includes: Collect real-time environmental parameters inside and outside the cabin, and establish a dynamic correlation model between the real-time environmental parameter changes outside the cabin and the expected changes in the environmental parameters inside the cabin; Based on the deviation between the expected environmental parameters in the cabin and the ideal environmental parameters, a pre-adjustment optimization feedback loop is constructed to perform a closed-loop feedback adjustment of the environmental parameters in the cabin. The historical multi-source environmental parameters are uniformly modeled to establish an in-cabin environmental situation recognition model to identify the in-cabin environmental situation type and situation matching degree; Based on the identified environmental situation types and the corresponding historical data of human physiological states, establish the association relationship between situation types and human physiological states; The human physiological state is predicted based on the prediction of expected environmental parameters in the cockpit, environmental situation recognition, and the correlation between situation type and human physiological state. The domain control strategy is adaptively generated by screening historical domain control strategy records to make secondary adjustments to the environmental parameters in the cockpit.

[0018] Reference Figure 2 As shown, real-time environmental parameters inside and outside the cockpit are collected, and a dynamic correlation model between the real-time environmental parameter changes outside the cockpit and the expected changes in the environmental parameters inside the cockpit is established. Specifically, it includes: At set time interval Collect the external environment parameters of the cockpit and the real-time environment parameters of the cockpit, and establish a synchronous data sequence of the environment status inside and outside the cockpit; Preprocess and normalize the synchronous data sequence, and use the processed data as the initial input of the dynamic correlation model of the cabin internal and external environmental parameters; The multiple linear regression model is used to associate the changes in the external environment of the cockpit with the response of the internal environment. Through the multiple linear regression model, a dynamic correlation model between the external environment parameters and the internal environment parameters is established. The multiple regression equation is expressed as follows: ; In the formula, is the data of the i-th environmental parameter in the cabin at time t, is the data of the mth environmental parameter outside the cabin at the previous moment of time t, is the regression coefficient of the i-th cabin environment parameter corresponding to the m-th cabin environment parameter; The sliding time window for training the dynamic correlation model of the cabin internal and external environmental parameters is set, and the synchronous data sequence of the cabin internal and external environmental states collected within the sliding time window is brought into the multivariate linear regression model to calculate and adjust the regression coefficient in real time.

[0019] Reference Figure 3 As shown, based on the deviation between the expected environmental parameters in the cabin and the ideal environmental parameters, a pre-adjustment optimization feedback loop is constructed, and a feedback closed-loop adjustment of the environmental parameters in the cabin is performed, specifically including: The system uses a proportional-integral-derivative controller as the core controller of the feedback loop. It dynamically adjusts the cabin's internal environmental parameters based on the real-time deviation between the ideal environmental parameters and the expected environmental parameters in the cabin. Specifically: Set the ideal environment parameters in the cockpit; The current real-time external environment parameters of the cockpit are brought into the multivariate linear regression model to obtain the expected internal environment parameters of the cockpit, which are used to describe the predicted values ​​of the environment parameters at the next moment; By evaluating the immediate response to the real-time deviation between the expected environmental parameters and the ideal environmental parameters inside the cockpit, the cumulative impact of historical deviations, and the changing trend of future deviations, the proportional gain, integral gain, and differential gain corresponding to the feedback adjustment are determined respectively; The gain parameters can be calculated based on the multiple regression model, and the immediate response of the deviation (parameter real-time data) and the cumulative impact of the historical deviation (parameter historical average data) (parameter time series differential) are respectively brought into the multiple regression model to determine the regression parameters.

[0020] Among them, the proportional gain directly reflects the magnitude of the current environmental parameter error and immediately amplifies it to respond quickly. This immediate response can predict the continued trend of the current error in a short period of time.

[0021] The integral gain accumulates environmental parameters over a period of time, identifies steady-state errors, and provides a prediction of long-term trends for future adjustments. By eliminating accumulated errors, the system can achieve stable target parameters in the long term.

[0022] The differential gain amplifies the error change rate, allowing the system to predict future trends in environmental parameters. The differential gain responds before future trends occur, reducing system fluctuations and overshoots through advance control.

[0023] The feedback formula is established through proportional gain, integral gain and differential gain to determine the adjustment range of various environmental parameters in the cockpit. The feedback formula is: ; In the formula, is the feedback adjustment amplitude of the jth environmental parameter inside the cabin at time t, is the real-time deviation between the jth expected environmental parameter and the ideal environmental parameter inside the cockpit at time t, , , They are proportional gain, integral gain and differential gain parameters respectively; By calculating the adjustment range of each environmental parameter in real time, various environmental parameters in the cabin are adjusted in real time, and the adjustment process is marked as one adjustment; When it is detected that the real-time adjustment amplitude of each environmental parameter controlled by the feedback loop is less than the set minimum adjustment amplitude threshold and tends to be stable, an adjustment is ended.

[0024] Reference Figure 4 As shown, the historical multi-source environmental parameters are uniformly modeled to establish an in-cabin environmental situation recognition model, which is used to identify the in-cabin environmental situation type and situation matching degree, including: Set the time window for collecting historical multi-source environmental parameters; After one adjustment is completed, the historical multi-source environmental parameters inside the cockpit within the historical multi-source environmental parameter collection time window are obtained, and the features of the historical multi-source environmental parameters are extracted based on the parameter type; Based on the multimodal data fusion model, the historical multi-source environmental parameters are uniformly modeled and the cabin environmental situation scores corresponding to the historical multi-source environmental parameters are calculated; Combine multi-source environmental parameter features with environmental context scores to establish a historical environmental context dataset, and set the feature center of all parameter feature types in the historical environmental context dataset; Based on the clustering algorithm, an association recognition model of environmental parameter feature classification corresponding to environmental situation type is established to identify the environmental situation type and situation matching degree in the cabin.

[0025] Among them, the situation type classification formula is: ; In the formula, The type of situation to be identified, For situation type The characteristic center, rate the environmental context; Each scenario type represents a specific combination of environmental conditions, for example, “high temperature and high humidity”, “low light and low air quality” and other scenario categories. The scenario matching degree indicates the degree of conformity of the specific parameters of the environmental condition combination to the scenario category.

[0026] Reference Figure 5 As shown, based on the identified environmental situation type and the corresponding human physiological state historical data, establishing the association relationship between the situation type and the human physiological state specifically includes: Obtain the human physiological state parameters collected at the same time corresponding to each data record in the historical environmental situation data set; Based on the environmental situation type recognition results of the historical environmental situation data set, the correlation between environmental situation type and human physiological state is determined through a regression analysis model; The size of the historical environmental scenario data set is continuously adjusted, and the real-time fit of the regression analysis model is measured by calculating the determination coefficient of the model. The determination coefficient calculation formula is: ; In the formula, is the coefficient of determination, The actual physiological state data of the human body in group a, is the physiological state data predicted by the model, is the average value of the actual physiological state data.

[0027] The adjustment is terminated until it is detected that the accuracy of the regression analysis model assessment reaches the preset standard.

[0028] Based on historical data analysis, the system deeply establishes the correlation between environmental situation types and human physiological states. By using regression analysis and correlation calculation, the system can quantify the response characteristics of human physiological states under different situation types, thereby providing more precise adaptive support for cabin environment control.

[0029] Reference Figure 6 As shown, based on the prediction of expected environmental parameters in the cockpit, environmental situation recognition, and the relationship between situation type and human physiological state, the human physiological state is predicted, and the domain control strategy is adaptively generated by screening historical domain control strategy records. The environmental parameters in the cockpit are adjusted twice, including: Obtain the real-time environmental parameters of the cabin's exterior and interior, bring the real-time environmental parameters into the dynamic correlation model of the cabin's interior and exterior environmental parameters, and obtain the expected cabin interior environmental parameters; The clustering algorithm model is used to identify the expected environmental scenario type and scenario matching degree at the next moment corresponding to the expected environmental parameters inside the cabin; The expected situation type and situation matching degree of the next moment are introduced into the regression analysis model of the correlation between the situation type and the human physiological state to obtain the expected human physiological state parameters of the next moment; Through similarity calculation, identify the historical human physiological state parameter records whose similarity with the expected human physiological state parameters at the next moment exceeds the set range, and then match and obtain the domain control policy records corresponding to the historical human physiological state parameters in the historical database; After screening out similar historical human physiological state parameter records, the domain control strategy corresponding to the historical human physiological state parameter records is dynamically weighted based on the feedback of historical human physiological state parameter changes. The strategy weight setting expression is: ; In the formula, is the dynamic weight of the current strategy, , are the set ideal physiological state parameters and the actual physiological state parameters, is the feedback adjustment coefficient, in which the ideal physiological state parameters can be set according to the human health and safety standards; All domain control policies are weighted and averaged to calculate the final weight of the domain control policy. The cabin environmental parameters corresponding to the domain control policy are adjusted secondary according to the distribution of the final weight of the domain control policy.

[0030] Reference Figure 7 As shown, a domain control optimization system based on cabin environment recognition includes: An environmental parameter change prediction module, which establishes a synchronous data sequence of the internal and external environmental states of the cabin, uses a regression model to associate the changes in the external environment of the cabin with the response of the internal environment, and establishes a dynamic correlation model of the internal and external environmental parameters of the cabin through a multivariate linear regression model; A domain control primary adjustment module, wherein the domain control primary adjustment module uses a proportional-integral-differential controller as a core controller of a feedback loop, establishes a feedback formula through proportional gain, integral gain and differential gain to determine the adjustment amplitude of various environmental parameters in the cabin, and adjusts the internal environmental parameters of the cabin once based on the real-time deviation value between the ideal environmental parameters and the expected environmental parameters in the cabin, until the real-time adjustment amplitude of various environmental parameters controlled by the feedback loop is less than the set minimum adjustment amplitude threshold and tends to be stable; The environmental situation type recognition module uniformly models the historical multi-source environmental parameters based on the multimodal data fusion model, calculates the cabin environmental situation scores corresponding to the historical multi-source environmental parameters, and establishes an association recognition model corresponding to the environmental situation type of the environmental parameter feature classification based on the clustering algorithm, which is used to identify the environmental situation type in the cabin and the situation matching degree.

[0031] A human physiological state prediction module, which determines the correlation between the environmental situation type and the human physiological state through a regression analysis model based on the environmental situation type recognition result of the historical environmental situation data set; A domain control secondary adjustment module predicts the human physiological state at the next moment based on the prediction of expected environmental parameters in the cabin, environmental situation recognition, and the correlation between situation types and human physiological states. Based on the feedback of historical human physiological state parameter changes, the module dynamically sets weights for the domain control strategies corresponding to the historical human physiological state parameter records and performs secondary adjustments to the cabin environmental parameters corresponding to the domain control strategies.

[0032] Furthermore, the present solution also proposes a storage medium for a domain control optimization method based on in-cabin environment recognition, on which a computer-readable program is stored. When the computer-readable program is called, the above-mentioned domain control optimization method based on in-cabin environment recognition is executed.

[0033] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).

[0034] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A domain control optimization method based on cabin environment recognition, characterized in that: include: Collect real-time environmental parameters inside and outside the cabin, and establish a dynamic correlation model between the real-time environmental parameter changes outside the cabin and the expected changes in the environmental parameters inside the cabin; Based on the deviation between the expected environmental parameters in the cabin and the ideal environmental parameters, a pre-adjustment optimization feedback loop is constructed to perform a closed-loop feedback adjustment of the environmental parameters in the cabin. The historical multi-source environmental parameters are uniformly modeled to establish an in-cabin environmental situation recognition model to identify the in-cabin environmental situation type and situation matching degree; Based on the identified environmental situation types and the corresponding historical data of human physiological states, establish the association relationship between situation types and human physiological states; The human physiological state is predicted based on the prediction of expected environmental parameters in the cockpit, environmental situation recognition, and the correlation between situation type and human physiological state. The domain control strategy is adaptively generated by screening historical domain control strategy records to make secondary adjustments to the environmental parameters in the cockpit.

2. The domain control optimization method based on cabin environment recognition according to claim 1 is characterized in that: The collecting of real-time environmental parameters inside and outside the cabin and establishing a dynamic correlation model between changes in real-time environmental parameters outside the cabin and expected changes in environmental parameters inside the cabin specifically includes: Collect the external environment parameters of the cabin and the real-time environment parameters of the cabin at set time intervals to establish a synchronous data sequence of the internal and external environment status of the cabin; Preprocess and normalize the synchronous data sequence, and use the processed data as the initial input of the dynamic correlation model of the cabin internal and external environmental parameters; A regression model is used to correlate the changes in the external environment of the cabin with the response of the internal environment. A dynamic correlation model of the internal and external environment parameters of the cabin is established through a multivariate linear regression model. The sliding time window for training the dynamic correlation model of the cabin internal and external environmental parameters is set, and the synchronous data sequence of the cabin internal and external environmental states collected within the sliding time window is brought into the multivariate linear regression model to calculate and adjust the regression coefficient in real time.

3. The domain control optimization method based on cabin environment recognition according to claim 2 is characterized in that: The construction of a pre-adjustment optimization feedback loop based on the deviation between the expected environmental parameters in the cabin and the ideal environmental parameters, and the feedback closed-loop adjustment of the environmental parameters in the cabin specifically include: The system uses a proportional-integral-derivative controller as the core controller of the feedback loop. It dynamically adjusts the cabin's internal environmental parameters based on the real-time deviation between the ideal environmental parameters and the expected environmental parameters in the cabin. Specifically: Set the ideal environment parameters in the cockpit; The current real-time external environment parameters of the cockpit are brought into the multivariate linear regression model to obtain the expected internal environment parameters of the cockpit, which are used to describe the predicted values ​​of the environment parameters at the next moment; By evaluating the immediate response to the real-time deviation between the expected environmental parameters and the ideal environmental parameters inside the cockpit, the cumulative impact of historical deviations, and the changing trend of future deviations, the proportional gain, integral gain, and differential gain corresponding to the feedback adjustment are determined respectively; The feedback formula is established through proportional gain, integral gain and differential gain to determine the adjustment range of various environmental parameters in the cockpit. The feedback formula is: ; In the formula, is the feedback adjustment amplitude of the jth environmental parameter inside the cabin at time t, is the real-time deviation between the jth expected environmental parameter and the ideal environmental parameter inside the cockpit at time t, , , They are proportional gain, integral gain and differential gain parameters respectively; By calculating the adjustment range of each environmental parameter in real time, various environmental parameters in the cabin are adjusted in real time, and the adjustment process is marked as one adjustment; When it is detected that the real-time adjustment amplitude of each environmental parameter controlled by the feedback loop is less than the set minimum adjustment amplitude threshold and tends to be stable, an adjustment is ended.

4. The domain control optimization method based on cabin environment recognition according to claim 3 is characterized in that: The unified modeling of historical multi-source environmental parameters and the establishment of a cockpit environment situation recognition model for identifying the type of cockpit environment situation and the situation matching degree specifically include: Set the time window for collecting historical multi-source environmental parameters; After one adjustment is completed, the historical multi-source environmental parameters inside the cockpit within the historical multi-source environmental parameter collection time window are obtained, and the features of the historical multi-source environmental parameters are extracted based on the parameter type; Based on the multimodal data fusion model, the historical multi-source environmental parameters are uniformly modeled and the cabin environmental situation scores corresponding to the historical multi-source environmental parameters are calculated; Combine multi-source environmental parameter features with environmental context scores to establish a historical environmental context dataset, and set the feature center of all parameter feature types in the historical environmental context dataset; Based on the clustering algorithm, an association recognition model of environmental parameter feature classification corresponding to environmental situation type is established to identify the environmental situation type and situation matching degree in the cabin.

5. The domain control optimization method based on cabin environment recognition according to claim 4 is characterized in that: The establishing of the association relationship between the situation type and the human physiological state based on the identified environmental situation type and the corresponding human physiological state historical data specifically includes: Obtain the human physiological state parameters collected at the same time corresponding to each data record in the historical environmental situation data set; Based on the environmental situation type recognition results of the historical environmental situation data set, the correlation between environmental situation type and human physiological state is determined through a regression analysis model; Continue to adjust the size of the historical environmental scenario data set, and measure the real-time fit of the regression analysis model by calculating the model's coefficient of determination, until it is detected that the accuracy of the regression analysis model evaluation reaches the preset standard and the adjustment is terminated.

6. The domain control optimization method based on cabin environment recognition according to claim 5 is characterized in that: The method of predicting the human physiological state based on the expected environmental parameter prediction in the cockpit, environmental situation recognition, and the correlation between the situation type and the human physiological state, adaptively generating the domain control strategy by screening the historical domain control strategy records, and performing secondary adjustment on the environmental parameters in the cockpit specifically includes: Obtain the real-time environmental parameters of the cabin's exterior and interior, bring the real-time environmental parameters into the dynamic correlation model of the cabin's interior and exterior environmental parameters, and obtain the expected cabin interior environmental parameters; The clustering algorithm model is used to identify the expected environmental scenario type and scenario matching degree at the next moment corresponding to the expected environmental parameters inside the cabin; The expected situation type and situation matching degree of the next moment are introduced into the regression analysis model of the correlation between the situation type and the human physiological state to obtain the expected human physiological state parameters of the next moment; Through similarity calculation, identify the historical human physiological state parameter records whose similarity with the expected human physiological state parameters at the next moment exceeds the set range, and then match and obtain the domain control policy records corresponding to the historical human physiological state parameters in the historical database; After screening out similar historical human physiological state parameter records, dynamically weight the domain control strategy corresponding to the historical human physiological state parameter records based on the feedback of historical human physiological state parameter changes; All domain control policies are weighted and averaged to calculate the final weight of the domain control policy. The cabin environmental parameters corresponding to the domain control policy are adjusted secondary according to the distribution of the final weight of the domain control policy.

7. A domain control optimization system based on in-cabin environment recognition, used to implement a domain control optimization method based on in-cabin environment recognition as claimed in any one of claims 1 to 6, characterized in that: include: An environmental parameter change prediction module, which establishes a synchronous data sequence of the internal and external environmental states of the cabin, uses a regression model to associate the changes in the external environment of the cabin with the response of the internal environment, and establishes a dynamic correlation model of the internal and external environmental parameters of the cabin through a multivariate linear regression model; A domain control primary adjustment module, wherein the domain control primary adjustment module uses a proportional-integral-differential controller as a core controller of a feedback loop, establishes a feedback formula through proportional gain, integral gain and differential gain to determine the adjustment amplitude of various environmental parameters in the cabin, and adjusts the internal environmental parameters of the cabin once based on the real-time deviation value between the ideal environmental parameters and the expected environmental parameters in the cabin, until the real-time adjustment amplitude of various environmental parameters controlled by the feedback loop is less than the set minimum adjustment amplitude threshold and tends to be stable; An environmental situation type recognition module, which unifies historical multi-source environmental parameters based on a multimodal data fusion model, calculates the cabin environmental situation scores corresponding to the historical multi-source environmental parameters, and establishes an association recognition model corresponding to the environmental situation type of the environmental parameter feature classification based on a clustering algorithm, so as to identify the environmental situation type in the cabin and the situation matching degree; A human physiological state prediction module, which determines the correlation between the environmental situation type and the human physiological state through a regression analysis model based on the environmental situation type recognition result of the historical environmental situation data set; A domain control secondary adjustment module predicts the human physiological state at the next moment based on the prediction of expected environmental parameters in the cabin, environmental situation recognition, and the correlation between situation types and human physiological states. Based on the feedback of historical human physiological state parameter changes, the module dynamically sets weights for the domain control strategies corresponding to the historical human physiological state parameter records and performs secondary adjustments to the cabin environmental parameters corresponding to the domain control strategies.

Images