Cockpit lighting dimming test method and system based on multi-sensor fusion

Abstract

This invention discloses a cockpit lighting dimming test method and system based on multi-sensor fusion, belonging to the field of aircraft lighting testing technology. This invention solves the problem that existing methods cannot meet the dimming test requirements of aircraft cockpit lighting. By constructing a test scenario, it uses brightness acquisition equipment and multiple sensors to collect lighting and physiological data within the aircraft cockpit. A decision tree model is used to perform cluster analysis on the data to obtain the pilot's corresponding reaction label. Based on the pilot's physiological reaction, it determines whether the current lighting conditions meet the pilot's requirements, thereby determining whether the dimming test requirements are met. Compared to observing light intensity or illumination angle with the naked eye, the above operation not only ensures the effectiveness of the cockpit lighting dimming test method but also improves its accuracy, thus helping to make better dimming decisions based on current lighting conditions.

Description

Technical Field

[0001] This invention relates to the field of aircraft lighting testing technology, specifically to a cockpit lighting dimming test method and system based on multi-sensor fusion. Background Technology

[0002] Multi-sensor fusion technology is a technique that combines multiple sensors to improve the accuracy and robustness of a system by collecting and fusing data from different sensors. In modern aircraft, cockpit lighting and dimming systems are critical safety features. The purpose of cockpit lighting is to provide pilots with a clear and comfortable visual environment to ensure flight safety. Dimming systems, on the other hand, can select different brightness levels according to different scenarios to avoid causing visual fatigue to the pilot.

[0003] However, in existing technologies, traditional cockpit lighting dimming testing methods mostly rely on observing light intensity or the angle of light illumination. But with the application of various sensors, traditional methods can no longer meet the requirements of modern aircraft for cockpit lighting dimming testing.

[0004] Therefore, it does not meet the existing requirements. In response, we propose a cockpit lighting dimming test method and system based on multi-sensor fusion. Summary of the Invention

[0005] The purpose of this invention is to provide a cockpit lighting dimming test method and system based on multi-sensor fusion. By constructing a test scenario, lighting and physiological data within the aircraft cockpit are collected using brightness acquisition equipment and multiple sensors. A decision tree model is used to perform cluster analysis on the data to obtain the pilot's corresponding reaction label. Based on the pilot's physiological reaction, the current lighting conditions are determined to see if they meet the pilot's requirements, thereby determining whether the dimming test is satisfactory. Compared to observing light intensity or illumination angle with the naked eye, this method not only ensures the effectiveness of the cockpit lighting dimming test method but also improves its accuracy, thus helping to make better dimming decisions based on current lighting conditions and solving the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A cockpit lighting dimming test method based on multi-sensor fusion includes the following steps: Step 1: Determine the test scene and lighting parameters by controlling the portable simulation terminal through the host of the measurement and control system, including: using the portable simulation terminal to construct diverse test scenes and integrating the intelligent headlights in the aircraft cockpit into the test scene through test software; Step 2: Collect light information inside the aircraft cockpit using a brightness acquisition device, and collect the pilot's physiological data in real time through multi-sensor fusion to obtain the actual situation inside the cockpit. Step 3: Read the data output from the portable simulation terminal, brightness acquisition device and multiple sensors online, and preprocess the acquired data, including filtering, noise reduction and smoothing. Step 4: Extract feature information from lighting data and physiological data, and use principal component analysis to fuse the extracted lighting data features and physiological data features to obtain a new feature space; Step 5: Construct a decision tree model to perform cluster analysis on the lighting status and the driver's physiological state in the cockpit, and obtain the driver's physiological reactions and subjective feelings under different lighting conditions; apply the trained decision tree model to the actual cockpit lighting dimming test, and feed the cluster analysis results back to the test software; Step Six: The testing software, based on a computer, displays real-time lighting data and driver physiological data in the cockpit. The decision tree model performs cluster analysis on the current lighting conditions based on the driver's physiological response, and conducts dimming tests on the cockpit lighting conditions based on the analysis results.

[0007] Furthermore, in step four, principal component analysis is used to fuse the extracted lighting data features and physiological data features, specifically including the following steps: The extracted lighting and physiological data features are normalized, and the correlation coefficient matrix between them is calculated to measure the degree of linear correlation. Based on the correlation coefficient matrix, the top k features with the highest correlation are selected as the final fusion features. Principal component analysis is performed on the selected fusion features to obtain k new principal components. The cockpit state is divided into three categories: alertness, fatigue, and anxiety, and these categories are mapped to the newly generated k principal components to obtain a new feature space.

[0008] Furthermore, in step five, a decision tree model is constructed to perform cluster analysis on the lighting status and the driver's physiological state in the cockpit, specifically including the following steps: The processed feature data is clustered using a decision tree algorithm to generate a cluster tree. Different physiological response states are defined based on the generated cluster tree. The real-time lighting conditions and driver physiological states in the cockpit are matched with the defined states and classified. The trained decision tree model is used to predict the state of the feature data, and the prediction results represent the impact of different lighting conditions in the cockpit on the driver's physiology. The accuracy and stability of the model are evaluated by comparing the actual state with the predicted state. Based on the model evaluation results, the parameters of the decision tree are updated to optimize the model. The optimized decision tree model is then applied to actual cockpit lighting dimming tests, and the optimized decision tree model is used to predict the state of new data. The prediction results serve as the basis for cockpit lighting dimming.

[0009] A cockpit lighting dimming test system based on multi-sensor fusion includes: The scene construction unit is used to determine the test scene and lighting parameters through a portable simulation terminal. The test scene is used to simulate various situations that may occur during actual flight of the aircraft, and the lighting parameters are used to simulate the type, color, brightness and dimming mode of the lighting system in the aircraft cockpit. The system control unit, which is the host of the measurement and control system, is used to provide a CPCI controller for the portable simulation terminal and control the portable simulation terminal to perform dimming tests on the cockpit lighting according to specified steps; The data acquisition unit consists of a brightness acquisition device and a multi-sensor device, used to collect real-time lighting data and pilot physiological data in the aircraft cockpit; The data fusion unit is used to extract feature information from lighting data and physiological data, and fuse the feature information of the two to obtain a new feature space; The model processing unit is used to construct a decision tree model and collect lighting data and physiological data from historical cockpit lighting dimming tests. The historical data is used to train and optimize the decision tree model to obtain an optimized decision tree model. The model prediction unit is used to perform cluster analysis on new lighting and physiological data using an optimized decision tree model to obtain the prediction analysis results corresponding to different lighting conditions in the cockpit and the driver's physiological response; and to conduct dimming tests on the lighting conditions in the cockpit based on the prediction analysis results.

[0010] Furthermore, it also includes: The data transmission unit provides a data transmission link for the dimming test system, enabling the data output by the portable simulation terminal, brightness acquisition device, multiple sensors, and decision tree model to be fed back to the test software for display in real time.

[0011] Furthermore, the data acquisition unit includes: The filtering module is used to remove high-frequency noise and unwanted vibrations from lighting and physiological data using a low-pass filter. The noise reduction module is used to remove trends and outliers from the filtered lighting and physiological data. The smoothing module is used to smooth the noise-reduced lighting and physiological data using a moving average technique.

[0012] Furthermore, the data fusion unit includes: The feature extraction module is used to select feature variables with permeability correlation from lighting data and physiological data, as feature information of lighting data and physiological data; The fusion processing module is used to fuse the extracted lighting data features and physiological data features using principal component analysis to obtain a new feature space.

[0013] Furthermore, the model processing unit includes: The sample processing module is used to collect historical lighting data and physiological data, clean the data, and then divide the data into training set and test set; The model training module is used to build decision tree models. It uses training and testing sets to train and test the decision tree models, and verifies whether the test results are the same as the historical actual results. If they are the same, a valid decision tree model is obtained.

[0014] Furthermore, the model processing unit also includes: The model optimization module is used to update and iterate the decision tree model using new lighting and physiological data, continuously optimizing the clustering ability of the decision tree model; The tag setting module is used to construct corresponding reaction tags based on the various physiological responses of drivers in historical physiological data.

[0015] Furthermore, the model prediction unit includes: The data prediction module is used to input the feature information from the extracted lighting and physiological data into the decision tree model for cluster analysis. Based on the current lighting conditions in the cockpit and the driver's physiological response, the cluster prediction determines whether the current lighting conditions meet the requirements of the dimming test.

[0016] Compared with the prior art, the beneficial effects of the present invention are: This invention constructs diverse test scenarios using a portable simulation terminal and integrates intelligent headlights from an aircraft cockpit into these scenarios. Lighting and physiological data from the aircraft cockpit are collected using brightness acquisition devices and multiple sensors. A decision tree model is used to cluster and analyze the data, obtaining the corresponding reaction labels for the pilot. Based on the pilot's physiological reactions, the current lighting conditions are determined to see if they meet the pilot's requirements, thus determining whether the dimming test is suitable. Compared to observing light intensity or illumination angle, this method not only ensures the effectiveness of the cockpit lighting dimming test method but also improves its accuracy, thereby helping to make better dimming decisions based on current lighting conditions. Attached Figure Description

[0017] Figure 1 This is a diagram of the cockpit lighting dimming test system based on multi-sensor fusion according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] To address the limitations of existing technologies, traditional cockpit lighting dimming testing methods, which largely rely on observing light intensity or angle, are no longer sufficient to meet the technical requirements of modern aircraft cockpit lighting dimming testing due to the application of various sensors. Please refer to [link to relevant documentation]. Figure 1 This embodiment provides the following technical solution: A cockpit lighting dimming test method based on multi-sensor fusion includes the following steps: Step 1: To ensure the validity and reliability of the test results, the portable simulation terminal is first controlled by the control system host to determine the test scenario and lighting parameters. This includes: constructing diverse test scenarios using the portable simulation terminal, including daytime, nighttime, various flight conditions, and weather conditions, to ensure the comprehensiveness and authenticity of the test; integrating the intelligent headlights in the aircraft cockpit into the test scenario through test software, such as the distribution and control settings of the headlights, to ensure the accuracy of the aircraft cockpit lighting dimming test. Step 2: Collect light information inside the aircraft cockpit using a brightness acquisition device, and collect the pilot's physiological data in real time through multi-sensor fusion, including: collecting light intensity, color temperature, air humidity, temperature, and carbon dioxide concentration inside the aircraft cockpit through light sensors, color temperature sensors, temperature and humidity sensors, and gas sensors; and collecting the pilot's physiological indicators, behavioral data, and subjective evaluations in real time through heart rate sensors, breathing sensors, eye trackers, and button switches on the pilot's panel to obtain the actual situation inside the cockpit. Step 3: Read the data output from the portable simulation terminal, brightness acquisition device and multiple sensors online, and preprocess the acquired data, including filtering, noise reduction and smoothing. Step 4: Extract feature information from lighting and physiological data. For example, extract and record indicators such as light intensity, color temperature, and usage frequency from the lighting data to reflect the overall lighting situation in the cockpit; normalize the data such as light intensity, color temperature, and usage frequency for subsequent data fusion analysis; extract heart rate, respiratory rate, brain wave spectrum, and skin conductance from the physiological data; use principal component analysis to fuse the extracted lighting and physiological data features to obtain a new feature space. This specifically includes the following steps: The extracted lighting and physiological data features are normalized to eliminate dimensional differences and distribution biases. A correlation coefficient matrix is ​​calculated between the lighting and physiological data features to measure their linear correlation. Based on the correlation coefficient matrix, the top k features with the highest correlation are selected as the final fusion features; the value of k needs to be appropriately chosen, generally between 5 and 10. Principal component analysis is performed on the selected fusion features to obtain k new principal components, which effectively reflect the combined information of the lighting and physiological data features. The cockpit state is divided into three categories: alertness, fatigue, and anxiety, and mapped to the newly generated k principal components to obtain a new feature space. Based on this, lighting and physiological data features can be considered simultaneously, thus achieving their fusion. Through the above fusion process, principal component analysis can effectively combine the originally independent lighting and physiological data features, enabling effective prediction of cockpit state and providing a scientific and accurate basis for cockpit lighting dimming.

[0020] Step 5: Construct a decision tree model to perform cluster analysis on the lighting conditions and the driver's physiological state in the cockpit, obtaining the driver's physiological reactions and subjective feelings under different lighting conditions; apply the trained decision tree model to actual cockpit lighting dimming tests, and feed the cluster analysis results back to the testing software; conduct adjustment tests on the cockpit lighting conditions by real-time monitoring of environmental parameters, driver physiological indicators, behavioral data, and subjective evaluations; specifically including the following steps: The processed feature data is clustered using a decision tree algorithm to generate a cluster tree. Different physiological response states, such as alertness, fatigue, and anxiety, are defined based on these cluster trees. Real-time cockpit lighting conditions and the driver's physiological state are matched with these defined states and classified accordingly. The trained decision tree model is then used to predict the state of the feature data, with the predictions representing the impact of different lighting conditions on the driver's physiology. The accuracy and stability of the model are evaluated by comparing the actual and predicted states, allowing for optimization and improvement. Based on the evaluation results, the decision tree parameters are updated to optimize the model and improve the accuracy of state predictions. The optimized decision tree model is then applied to actual cockpit lighting dimming tests, using the optimized model to predict the state of new data. The prediction results serve as the basis for cockpit lighting dimming.

[0021] Step Six: The testing software, based on a computer, displays real-time lighting data and driver physiological data in the cockpit. The decision tree model performs cluster analysis on the current lighting conditions based on the driver's physiological response, and conducts dimming tests on the cockpit lighting conditions based on the analysis results.

[0022] A cockpit lighting dimming test system based on multi-sensor fusion includes: The scenario construction unit is used to determine the test scenario and lighting parameters via a portable simulation terminal. The test scenario simulates various situations that may occur during actual flight, while the lighting parameters simulate the type, color, brightness, and dimming mode of the cockpit lighting system. Specifically, the scenario construction unit is an important component of the testing method. It determines the test scenario and related lighting parameters using a portable simulation terminal, setting them according to possible situations during actual flight, such as different conditions during day and night, different weather conditions, and different flight conditions. By simulating different scenarios, the performance of the cockpit lighting system under different conditions can be more comprehensively evaluated. The lighting parameters need to be set according to the specific lighting system, such as simulating different light types (incandescent, LED, fluorescent, etc.), different light colors (warm, cool, neutral, etc.), and different dimming modes (manual, automatic, timed dimming, etc.). By controlling these parameters, various lighting conditions can be simulated to facilitate a comprehensive evaluation of the cockpit lighting system.

[0023] The system control unit, which is the host of the measurement and control system, provides a CPCI controller for the portable simulation terminal and controls the portable simulation terminal to perform dimming tests on the cockpit lighting according to specified steps. Specifically, after the test scenario is constructed through the portable simulation terminal, the system control unit sends various control commands to the portable simulation terminal, such as turning a light on / off, changing the color, brightness, and dimming mode of the light, to control the operation of the portable simulation terminal, thereby achieving a comprehensive evaluation of the cockpit lighting system.

[0024] The data acquisition unit, comprising a brightness acquisition device and a multi-sensor device, is used to collect real-time lighting data and pilot physiological data within the aircraft cockpit; the data acquisition unit includes: The filtering module is used to remove high-frequency noise and unnecessary vibrations from lighting and physiological data using a low-pass filter. Specifically, the low-pass filter can effectively remove noise sources such as air conditioning vibrations in the cockpit and turbulence during flight. The commonly used low-pass filter cutoff frequency is 10-20Hz, and the band-stop filter cutoff frequency is 100-200Hz.

[0025] The noise reduction module is used to remove trends and outliers from filtered lighting and physiological data, which can help remove temporary anomalies caused by equipment drift or other reasons. The smoothing module is used to smooth the noise-reduced lighting and physiological data using moving average technology. This can remove short-term fluctuations in the data and improve its visibility and stability. The processing steps for the above three types of modules are as follows: first, filtering is performed to remove noise and vibration; then, denoising is performed to remove trends and outliers; and finally, smoothing is performed to remove short-term fluctuations.

[0026] The data fusion unit is used to extract feature information from lighting data and physiological data, and fuse the feature information of the two to obtain a new feature space; the data fusion unit includes: The feature extraction module is used to select feature variables with permeability correlation from lighting and physiological data as feature information for the lighting and physiological data. Specifically, the selected features should have a strong correlation with the target variable, such as the optimization result of lighting dimming, to facilitate subsequent data analysis and modeling. Secondly, the selected features should have a certain degree of interpretability to facilitate the analysis of the reasons behind the data. At the same time, the selected features should have sufficient diversity to cover different aspects and details in the lighting and physiological data. Finally, the selected features should be able to represent the overall picture and characteristics of the lighting and physiological data to ensure the accuracy of model predictions.

[0027] The fusion processing module is used to fuse the extracted lighting data features and physiological data features using principal component analysis to obtain a new feature space. Specifically, the lighting data features and physiological data features are projected into two different subspaces using principal component analysis, resulting in two new feature spaces: the lighting data feature space and the physiological data feature space. The features in these feature spaces represent important information about the lighting system and the physiological system, respectively. Then, a weighted average method is used to fuse them to obtain a new feature space that includes all the information from the lighting data features and physiological data features.

[0028] The model processing unit is used to construct a decision tree model and collect lighting and physiological data from historical cockpit lighting dimming tests. It uses this historical data to train and optimize the decision tree model, resulting in an optimized model. The model processing unit includes: The sample processing module collects historical lighting and physiological data, cleans the data, and then divides it into training and test sets. Specifically, data cleaning ensures data quality and integrity. Cleaning steps include outlier removal, missing value imputation, and data transformation. The data is then divided into training and test sets. The training set is used to train the decision tree model, and the test set is used to test the performance of the decision tree model, thus comprehensively validating its performance.

[0029] The model training module is used to build decision tree models. It trains and tests these models using training and testing sets, verifying that the test results match historical real-world results. If they match, the decision tree model is considered validated. Specifically, by defining the type of decision tree model (e.g., SVM, a kernel-based classification algorithm), cross-validation is used to tune hyperparameters during training to achieve optimal model performance. During testing, the model's performance on the test set is evaluated and compared to historical real-world results. If the test results closely match historical real-world results, the model is considered effective and suitable for cockpit lighting control. Finally, the trained model is saved and used in subsequent cockpit lighting dimming tests to achieve real-time control of the cockpit lighting.

[0030] The model optimization module is used to update and iterate the decision tree model using new lighting and physiological data, continuously optimizing the clustering ability of the decision tree model. Specifically, since the constantly collected new lighting and physiological data may contain specific information or patterns, the new information or patterns can help in judging the state of the cockpit. By continuously updating and improving the decision tree model based on new data and feedback, its clustering ability in complex cockpit states can be improved.

[0031] The label setting module is used to construct corresponding reaction labels based on various physiological responses of drivers in historical physiological data. Specifically, it classifies reactions according to common physiological response types, such as heart rate, blood pressure, respiration, and muscle tension, and determines the dimensions of each reaction label to facilitate learning and recognition by the decision tree model. These dimensions can be combinations of multiple dimensions; for example, "heart rate between 90-130 beats per minute" can be used as a label dimension. Next, thresholds for physiological response labels are determined to allow the decision tree model to select appropriate labels when making state judgments. When determining the thresholds, multiple factors need to be considered, such as the magnitude, duration, and location of the physiological response. Finally, the labels are validated and optimized to ensure their accuracy and completeness.

[0032] The model prediction unit is used to perform cluster analysis on new lighting and physiological data using an optimized decision tree model to obtain predictive analysis results corresponding to different lighting conditions in the cockpit and the driver's physiological responses; and based on the predictive analysis results, to conduct dimming tests on the cockpit lighting conditions. The model prediction unit includes: The data prediction module inputs feature information from the extracted lighting and physiological data into a decision tree model for cluster analysis. Based on the current cockpit lighting conditions and the driver's physiological responses, it predicts whether the current lighting conditions meet the dimming test requirements. Specifically, it aggregates the feature information from the extracted lighting and physiological data to form a new feature vector matrix. This matrix includes all potentially influencing factors, such as light intensity, color temperature, location, and the driver's physiological responses. This new feature vector matrix is ​​then input into the decision tree model for cluster analysis. The clustering results are divided into several clusters, each representing a different physiological state. For each cluster, its average physiological state indicators, such as average heart rate and average blood flow, are calculated to reflect the current lighting status in the cockpit. Based on these average physiological state indicators, the clusters are further divided into two categories: those meeting the dimming test requirements and those not meeting them. Finally, based on the classification results, corresponding suggestions are output, such as increasing or decreasing lighting brightness or changing lighting color, to help the cockpit lighting dimming test system make appropriate lighting adjustments according to the actual situation.

[0033] The data transmission unit provides a data transmission link for the dimming test system, enabling the data output by the portable simulation terminal, brightness acquisition device, multiple sensors, and decision tree model to be fed back to the test software for display in real time. Specifically, the main function of the data transmission unit is to efficiently and stably transmit data between the various units, so that the units can share data and resources, while also ensuring the accuracy and integrity of the data.

[0034] The beneficial effects achieved by the above methods are as follows: The effectiveness of the cockpit lighting dimming test method is ensured, and the accuracy of the cockpit lighting dimming test method is improved, thereby helping to make better dimming decisions based on the current lighting conditions.

[0035] Working principle: Diverse test scenarios are constructed through a portable simulation terminal, and the intelligent headlights in the aircraft cockpit are integrated into the test scenarios; lighting data and physiological data in the aircraft cockpit are collected through brightness acquisition devices and multiple sensors, and the data are clustered and analyzed using a decision tree model to obtain the corresponding reaction labels of the pilot; and the current lighting conditions are judged based on the pilot's physiological reaction to determine whether they meet the requirements of the dimming test.

[0036] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0037] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A cockpit lighting dimming test method based on multi-sensor fusion, characterized in that, Includes the following steps: Step 1: Determine the test scene and lighting parameters by controlling the portable simulation terminal through the host of the measurement and control system, including: using the portable simulation terminal to construct diverse test scenes and integrating the intelligent headlights in the aircraft cockpit into the test scene through test software; Step 2: Collect light information inside the aircraft cockpit using a brightness acquisition device, and collect the pilot's physiological data in real time through multi-sensor fusion to obtain the actual situation inside the cockpit. Step 3: Read the data output from the portable simulation terminal, brightness acquisition device and multiple sensors online, and preprocess the acquired data, including filtering, noise reduction and smoothing. Step 4: Extract feature information from lighting data and physiological data, and use principal component analysis to fuse the extracted lighting data features and physiological data features to obtain a new feature space; Step 5: Construct a decision tree model to perform cluster analysis on the lighting status and the driver's physiological state in the cockpit, and obtain the driver's physiological reactions and subjective feelings under different lighting conditions; apply the trained decision tree model to the actual cockpit lighting dimming test, and feed the cluster analysis results back to the test software; Step Six: The testing software, based on a computer, displays real-time lighting data and driver physiological data in the cockpit. The decision tree model performs cluster analysis on the current lighting conditions based on the driver's physiological response, and conducts dimming tests on the cockpit lighting conditions based on the analysis results.

2. The cockpit lighting dimming test method based on multi-sensor fusion according to claim 1, characterized in that: In step four, principal component analysis is used to fuse the extracted lighting data features and physiological data features, specifically including the following steps: The extracted lighting and physiological data features are normalized, and the correlation coefficient matrix between them is calculated to measure the degree of linear correlation. Based on the correlation coefficient matrix, the top k features with the highest correlation are selected as the final fusion features. Principal component analysis is performed on the selected fusion features to obtain k new principal components. The cockpit state is divided into three categories: alertness, fatigue, and anxiety, and these categories are mapped to the newly generated k principal components to obtain a new feature space.

3. The cockpit lighting dimming test method based on multi-sensor fusion according to claim 1, characterized in that: In step five, a decision tree model is constructed to perform cluster analysis on the lighting status and the driver's physiological state in the cockpit. This specifically includes the following steps: The processed feature data is clustered using a decision tree algorithm to generate a cluster tree. Different physiological response states are defined based on the generated cluster tree. The real-time lighting conditions and driver physiological states in the cockpit are matched with the defined states and classified. The trained decision tree model is used to predict the state of the feature data, and the prediction results represent the impact of different lighting conditions in the cockpit on the driver's physiology. The accuracy and stability of the model are evaluated by comparing the actual state with the predicted state. Based on the model evaluation results, the parameters of the decision tree are updated to optimize the model. The optimized decision tree model is then applied to actual cockpit lighting dimming tests, and the optimized decision tree model is used to predict the state of new data. The prediction results serve as the basis for cockpit lighting dimming.

4. A cockpit lighting dimming test system based on multi-sensor fusion, applied to the cockpit lighting dimming test method based on multi-sensor fusion as described in any one of claims 1-3, characterized in that: include: The scene construction unit is used to determine the test scene and lighting parameters through a portable simulation terminal. The test scene is used to simulate various situations that may occur during actual flight of the aircraft, and the lighting parameters are used to simulate the type, color, brightness and dimming mode of the lighting system in the aircraft cockpit. The system control unit, which is the host of the measurement and control system, is used to provide a CPCI controller for the portable simulation terminal and control the portable simulation terminal to perform dimming tests on the cockpit lighting according to specified steps; The data acquisition unit consists of a brightness acquisition device and a multi-sensor device, used to collect real-time lighting data and pilot physiological data in the aircraft cockpit; The data fusion unit is used to extract feature information from lighting data and physiological data, and fuse the feature information of the two to obtain a new feature space; The model processing unit is used to construct a decision tree model and collect lighting data and physiological data from historical cockpit lighting dimming tests. The historical data is used to train and optimize the decision tree model to obtain an optimized decision tree model. The model prediction unit is used to perform cluster analysis on new lighting and physiological data using an optimized decision tree model to obtain the prediction analysis results corresponding to different lighting conditions in the cockpit and the driver's physiological response. Based on the predictive analysis results, a dimming test was conducted on the lighting conditions in the cockpit.

5. The cockpit lighting dimming test system based on multi-sensor fusion according to claim 4, characterized in that: Also includes: The data transmission unit provides a data transmission link for the dimming test system, enabling the data output by the portable simulation terminal, brightness acquisition device, multiple sensors, and decision tree model to be fed back to the test software for display in real time.

6. The cockpit lighting dimming test system based on multi-sensor fusion according to claim 4, characterized in that: The data acquisition unit includes: The filtering module is used to remove high-frequency noise and unwanted vibrations from lighting and physiological data using a low-pass filter. The noise reduction module is used to remove trends and outliers from the filtered lighting and physiological data. The smoothing module is used to smooth the noise-reduced lighting and physiological data using a moving average technique.

7. The cockpit lighting dimming test system based on multi-sensor fusion according to claim 4, characterized in that: The data fusion unit includes: The feature extraction module is used to select feature variables with permeability correlation from lighting data and physiological data, as feature information of lighting data and physiological data; The fusion processing module is used to fuse the extracted lighting data features and physiological data features using principal component analysis to obtain a new feature space.

8. The cockpit lighting dimming test system based on multi-sensor fusion according to claim 4, characterized in that: The model processing unit includes: The sample processing module is used to collect historical lighting data and physiological data, clean the data, and then divide the data into training set and test set; The model training module is used to build decision tree models. It uses training and testing sets to train and test the decision tree models, and verifies whether the test results are the same as the historical actual results. If they are the same, a valid decision tree model is obtained.

9. The cockpit lighting dimming test system based on multi-sensor fusion according to claim 8, characterized in that: The model processing unit further includes: The model optimization module is used to update and iterate the decision tree model using new lighting and physiological data, continuously optimizing the clustering ability of the decision tree model; The tag setting module is used to construct corresponding reaction tags based on the various physiological responses of drivers in historical physiological data.

10. The cockpit lighting dimming test system based on multi-sensor fusion according to claim 4, characterized in that: The model prediction unit includes: The data prediction module is used to input the feature information from the extracted lighting and physiological data into the decision tree model for cluster analysis. Based on the current lighting conditions in the cockpit and the driver's physiological response, the cluster prediction determines whether the current lighting conditions meet the requirements of the dimming test.

Images