Flight training evaluation system fusing electroencephalogram characteristics and physiological indexes

Abstract

The invention relates to the technical field of flight training evaluation, and discloses an electroencephalogram feature and physiological index fused flight training evaluation system. The system comprises a physiological signal acquisition module which synchronously captures multichannel electroencephalogram original signals and body surface physiological index data, and the body surface physiological index data comprises an electrocardiograph R-R interval sequence, respiratory wave frequency amplitude and galvanic skin response amplitude; the multi-modal fusion module is used for analyzing an electrocardiograph R-R interval sequence to generate a heart rate variability feature vector and establishing dynamic association mapping of an electroencephalogram entropy value and a physiological feature vector; the cognitive state modeling module is used for generating a cognitive load index according to the dynamic association mapping and constructing a cognitive stability quantization matrix; the self-adaptive feedback module is used for receiving related data and dynamically adjusting simulated flight scene parameters; and the evaluation output module is used for integrating the data to generate a comprehensive training evaluation report containing a neurophysiological coordination degree score and an operation accuracy rating. According to the system, comprehensive evaluation and dynamic training adjustment of the cognitive state of the pilot are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of flight training assessment, and in particular to a flight training assessment system integrating electroencephalogram (EEG) characteristics with physiological indicators. Background Art

[0002] In the field of flight training, the evaluation of pilot training effectiveness has long relied on operational trajectory data and subjective ratings, an approach with significant limitations. Traditional evaluation systems focus primarily on pilots' operational accuracy, such as external performance indicators like route deviation and operational response time, but struggle to capture their internal cognitive state. For example, during flight missions simulating complex weather conditions, pilots may experience cognitive fatigue due to sustained, intense concentration. However, operational data alone often fails to capture these changes in state in a timely manner, leading to biased evaluation results. Existing assessment systems attempt to incorporate single physiological indicators, such as heart rate or galvanic skin response. However, these indicators, when used in isolation, fail to fully capture a pilot's cognitive processes. Heart rate changes can be caused by both physiological stress and increased cognitive load, and without the support of other indicators, cognitive status can be misjudged. Furthermore, while EEG signals can reflect brain activity, existing processing methods are often limited to single-frequency analysis, ignoring the dynamic correlation between signals across different frequency bands and physiological indicators, making it impossible to construct a comprehensive model of cognitive status. Traditional assessment systems often rely on static feedback mechanisms, providing fixed feedback based on preset thresholds. This makes them incapable of adapting to the demands of different flight phases. Pilots face significant differences in cognitive load during takeoff and cruise, making static feedback difficult to implement targeted training adjustments, hindering training effectiveness. These shortcomings leave room for improvement in existing assessment systems in terms of accuracy, comprehensiveness, and dynamic adaptability. Summary of the Invention

[0003] The purpose of the present invention is to provide a flight training evaluation system that integrates EEG characteristics with physiological indicators to solve the problems raised in the above background technology.

[0004] To achieve the above objectives, the present invention provides a flight training assessment system that integrates EEG characteristics with physiological indicators, the system comprising: A physiological signal acquisition module is used to synchronously capture the pilot's multi-channel EEG raw signals and body surface physiological index data during the simulated flight mission. The body surface physiological index data at least includes the ECG RR interval sequence, respiratory wave frequency amplitude, and skin galvanic response amplitude; A multimodal fusion module is used to perform time-frequency decomposition processing on the original EEG signal to extract energy entropy values ​​of different frequency bands, and simultaneously parse the ECG RR interval sequence to generate heart rate variability feature vectors, and establish a dynamic correlation mapping between EEG entropy values ​​and physiological feature vectors through a cross-modal attention mechanism; a cognitive state modeling module, configured to generate a cognitive load index reflecting the pilot's concentration based on the dynamic association map, and to construct a cognitive stability quantification matrix based on the respiratory wave frequency amplitude fluctuation rate; An adaptive feedback module, configured to receive the cognitive load index and the cognitive stability quantification matrix and dynamically adjust the visual stimulation intensity and joystick force feedback coefficient of the simulated flight scene according to preset key operation nodes of the flight phase; The evaluation output module is used to integrate the cognitive load index, cognitive stability quantification matrix and simulated flight operation trajectory data to generate a comprehensive training evaluation report that includes neurophysiological coordination score and operation accuracy rating.

[0005] Preferably, the physiological signal acquisition module includes an EEG signal preprocessing unit for performing power frequency noise filtering and eye movement artifact correction on the multi-channel EEG original signal, and separating effective EEG components using independent component analysis; The physiological signal acquisition module also includes a physiological signal synchronization unit, which binds the corrected EEG components, ECG RR interval series, and skin electrodermal response amplitude to the same time coordinate system through a timestamp alignment mechanism to form a time-domain aligned neurophysiological synchronization data set.

[0006] Preferably, the multimodal fusion module includes a frequency domain decomposition engine, which performs wavelet packet transform on the EEG components to divide the frequency bands into Delta wave, Theta wave, Alpha wave and Beta wave, and calculates the energy entropy value sequence of each frequency band; The multimodal fusion module also includes a physiological feature analysis unit, which analyzes the sample entropy value and fractal dimension of the ECG RR interval sequence through a nonlinear dynamics algorithm to generate a heart rate variability feature vector.

[0007] Preferably, the cognitive state modeling module includes a cognitive load index generation unit, which receives a frequency band energy entropy value sequence and a heart rate variability feature vector, uses a convolutional neural network model to learn the nonlinear association between the energy entropy value and the heart rate variability feature vector, and outputs a cognitive load index curve that changes with flight mission time; The cognitive state modeling module also includes a stability quantification unit that integrates the respiratory wave frequency amplitude fluctuation rate and the number of sudden changes in the galvanic skin response amplitude to construct a cognitive stability quantification matrix that reflects the autonomic nervous system activation state.

[0008] Preferably, the adaptive feedback module includes a scene control decision maker, which triggers a visual stimulation intensity adjustment instruction and a joystick force feedback coefficient adjustment instruction according to the numerical threshold of the cognitive load index curve at the key operation node of the preset flight phase; The scene control decision maker also receives data on the number of mutations in the galvanic skin response amplitude in the cognitive stability quantization matrix, and when the number of mutations exceeds a baseline threshold, increases the damping coefficient of the joystick force feedback coefficient.

[0009] Preferably, the evaluation output module includes a neurophysiological coordination calculation unit, which performs normalized weighted fusion on the cognitive load index curve and the cognitive stability quantification matrix to generate a neurophysiological coordination score reflecting the degree of coordination between EEG characteristics and physiological indicators; The evaluation output module also includes an operation accuracy analysis unit, which dynamically time-regularizes the simulated flight operation trajectory data with the standard flight trajectory template, calculates the mean heading angle deviation and the altitude control error rate, and outputs the operation accuracy rating.

[0010] Preferably, the evaluation output module further comprises a comprehensive evaluation report generating unit, which associates the neurophysiological coordination score with the operation accuracy rating, and marks the flight training stage as a high-risk state when the neurophysiological coordination score is lower than a critical value and the operation accuracy rating is unqualified; The comprehensive assessment report generating unit further extracts abnormal peak values ​​of respiratory wave frequency amplitude fluctuation rate in the cognitive stability quantification matrix and associates them with time nodes when high-risk states occur.

[0011] Preferably, the system further comprises a training optimization suggestion module, which receives the comprehensive training evaluation report generated by the evaluation output module and analyzes the correlation pattern between the neurophysiological coordination score and the operation accuracy rating; The training optimization suggestion module generates a special training program for visual attention allocation or physiological stress regulation according to the correlation pattern, and feeds the special training program back to the adaptive feedback module to update the visual stimulation intensity adjustment rule.

[0012] Preferably, the training optimization suggestion module includes a program customization engine, which generates a visual tracking training program for enhancing situational awareness when it detects that the neurophysiological coordination score is continuously lower than the baseline; The solution customization engine also receives steep drop point data of the cognitive load index curve, and generates anti-interference physiological adjustment training instructions based on the flight operation type corresponding to the steep drop point.

[0013] Preferably, the system is deployed with a data persistence storage module to archive the neurophysiological synchronization data set of the physiological signal acquisition module, the cognitive stability quantification matrix of the cognitive state modeling module, and the comprehensive training evaluation report of the evaluation output module in time series; The data persistence storage module is also associated with the execution records of the special training program generated by the training optimization suggestion module to form an evolution tracking database of the pilot's long-term neurophysiological adaptability.

[0014] Compared with the prior art, the present invention has the following beneficial effects: By integrating EEG features with multi-dimensional physiological indicators, a comprehensive characterization of the pilot's cognitive state is achieved. Compared to single-indicator evaluation methods, the integration of multimodal data can more accurately reflect the complex changes in the cognitive process. The time-frequency decomposition of EEG signals and the extraction of energy entropy values ​​in different frequency bands can capture the dynamic characteristics of the brain during information processing. The heart rate variability feature vectors generated by the ECG RR interval sequence can reflect the regulatory state of the autonomic nervous system. The correlation mapping established by the two through the cross-modal attention mechanism reveals the inherent connection between cognitive activity and physiological responses, providing a deeper and more comprehensive assessment of cognitive state. The cognitive load index and cognitive stability quantification matrix generated by the cognitive state modeling module provide a quantitative basis for assessment. The cognitive load index directly reflects the pilot's attentional focus, while the quantitative matrix constructed from the respiratory wave frequency amplitude fluctuation rate reflects the stability of the cognitive state. The combination of the two can comprehensively assess the pilot's cognitive performance in different mission scenarios. For example, when executing an emergency avoidance mission, it can determine whether the pilot maintains a high level of concentration and whether their cognitive state is stable, avoiding the one-sidedness of single-metric assessment. The adaptive feedback module dynamically adjusts simulation scenario parameters based on key operational nodes during each flight phase, enhancing the relevance and effectiveness of training. Different flight phases place varying cognitive demands on pilots. Takeoff and landing require higher levels of concentration, while cruising emphasizes cognitive stability. Dynamic adjustments to visual stimulation intensity and joystick force feedback coefficients simulate operating environments under varying cognitive loads, helping pilots adapt to complex situations during training and improving their adaptability in actual flight. The assessment output module integrates multiple data points to generate a comprehensive report, including neurophysiological coordination scores and operational accuracy ratings, providing a comprehensive reference for evaluating training effectiveness. The neurophysiological coordination score reflects the synergy between cognitive state and physiological response, while the operational accuracy rating reflects actual operational performance. The combination of the two provides a more comprehensive assessment, enabling trainers to develop targeted training plans, optimize training content, and promote the overall improvement of pilot skills. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a working principle diagram of the flight training evaluation system integrating EEG characteristics with physiological indicators according to the present invention; Figure 2 Flowchart of preprocessing and synchronization of physiological signal acquisition module; Figure 3 Flowchart for index generation and stability quantification of the cognitive state modeling module; Figure 4 Flowchart for the evaluation output module score calculation and accuracy analysis; Figure 5 Flowchart for generating and providing feedback on training optimization suggestion modules. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] See also Figure 1 The present invention provides a flight training evaluation system integrating EEG characteristics with physiological indicators, the system comprising: By synchronously collecting EEG signals and surface physiological indicators, combined with multimodal data fusion and cognitive state modeling techniques, the system enables dynamic assessment and feedback adjustment of pilot training status. The system comprises a physiological signal acquisition module, a multimodal fusion module, a cognitive state modeling module, an adaptive feedback module, and an evaluation output module. The physiological signal acquisition module uses multi-channel EEG electrodes and surface physiological sensors to synchronously capture raw EEG signals and physiological indicators such as ECG, respiratory waves, and skin conductance. The multimodal fusion module decomposes the EEG signals in the time-frequency domain, extracting energy entropy values ​​in the Delta, Theta, Alpha, and Beta frequency bands. It also analyzes the ECG RR interval sequence to generate a heart rate variability feature vector. A dynamic correlation mapping between EEG and physiological indicators is established through a cross-modal attention mechanism. The cognitive state modeling module generates a cognitive load index based on the dynamic correlation mapping and constructs a cognitive stability quantification matrix based on the respiratory wave frequency and amplitude fluctuation rate. The adaptive feedback module dynamically adjusts the visual stimulation intensity and joystick force feedback coefficient of the simulated flight scenario based on cognitive state parameters. The evaluation output module integrates neurophysiological data and operation trajectory data to generate a comprehensive training evaluation report that includes neurophysiological coordination scores and operation accuracy ratings.

[0018] Example 1: See Figure 2The EEG signal preprocessing unit of the physiological signal acquisition module uses a multi-stage filtering architecture to process the raw EEG signals. High-frequency noise is initially eliminated through hardware-level analog filtering, followed by a 50Hz power frequency notch in the digital signal processing phase, and a bidirectional IIR filter design is used to avoid phase distortion. To address eye movement artifact interference, the system combines independent component analysis with regression correction algorithms. First, the FastICA algorithm is used to separate independent components containing eye movement artifacts. Component identification is then performed based on the time-frequency feature templates of typical eye movement artifacts. Finally, the least squares method is used to construct an artifact regression model and remove it from the original signal. Wavelet threshold denoising is used to suppress electromyographic interference, using the sym8 wavelet basis function for an 8-layer decomposition and soft thresholding of high-frequency coefficients. The preprocessed EEG signal retains the effective frequency band of 0.5-45Hz, significantly improving the signal-to-noise ratio.

[0019] The physiological signal synchronization unit adopts a hierarchical time synchronization strategy. At the hardware level, FPGA is used to generate synchronization pulse signals accurate to the microsecond level, driving all acquisition devices to work in master-slave mode. At the data level, a timestamp alignment mechanism based on the NTP protocol is used to embed a high-precision time stamp in each frame of the data packet. For signal sources with different sampling rates, the system uses cubic spline interpolation to achieve sampling rate normalization, and uniformly resamples the ECG signal from 1000Hz and the respiratory signal from 250Hz to the same 1000Hz sampling rate as the EEG signal. During the time alignment process, the system continuously monitors the clock drift of each signal channel. When the cumulative deviation exceeds 1ms, the timestamp correction program is automatically triggered to maintain the time consistency of multimodal data.

[0020] The frequency domain decomposition engine of the multimodal fusion module uses an improved wavelet packet transform algorithm to process EEG signals. The decomposition process uses the db4 wavelet basis function for a six-layer recursive decomposition, forming 64 subbands covering 0.5-30 Hz. Based on typical cognitive states during flight missions, the system focuses on extracting four characteristic frequency bands: Delta waves (0.5-4 Hz) corresponding to deep relaxation, Theta waves (4-8 Hz) associated with working memory load, Alpha waves (8-13 Hz) reflecting resting arousal, and Beta waves (13-30 Hz) characterizing cognitive activity. Energy entropy is calculated using a sliding time window analysis using a Hamming window function. A window length of 2 seconds and a step size of 500 ms balance the requirements for temporal resolution and spectral stability. The energy of each frequency band is converted into a probability distribution through normalization, and the Shannon entropy is then calculated to form the time series features.

[0021] The physiological feature analysis unit performs multi-scale feature extraction on the ECG RR interval series. Time-domain analysis includes conventional metrics such as standard deviation (SDNN) and the root mean square (RMSSD) of the difference between adjacent RR intervals. Nonlinear dynamic analysis uses the sample entropy algorithm with an embedding dimension set to 2 and a similarity tolerance of 0.2 times the time series standard deviation to effectively quantify the complexity of the RR interval series. Fractal dimension calculation uses the Higuchi algorithm with a maximum subsequence length set to 1 / 4 of the total number of RR intervals. The slope of the fractal curve is obtained through least squares fitting. Frequency-domain analysis uses the Lomb-Scargle periodogram to estimate the ratio of high-frequency power to low-frequency power, adapting to non-uniformly sampled RR interval series. The resulting 12-dimensional heart rate variability feature vector contains three types of characteristic metrics: time domain, frequency domain, and nonlinearity, comprehensively reflecting the activity of the autonomic nervous system.

[0022] An adaptive baseline correction method is used for galvanic skin response signal processing. The system dynamically tracks the slowly changing trend components of the galvanic skin signal and extracts the baseline using a moving average filter with a window length set to 30 seconds. For transient galvanic skin response events, the detection algorithm comprehensively examines the signal slope change and amplitude threshold, defining a valid response as a rising edge slope exceeding 0.05μS / s and an amplitude exceeding 10% of the baseline. Respiratory wave signal processing focuses on extracting features in two dimensions: frequency and amplitude. Instantaneous frequency estimation uses zero-crossing detection combined with median filtering, and the amplitude envelope is extracted using the Hilbert transform. The system pays special attention to the fluctuation characteristics of the respiratory wave frequency, calculating the standard deviation of the respiratory frequency per minute as an indicator of the respiratory wave frequency amplitude fluctuation rate.

[0023] The time alignment accuracy of multimodal data is ensured through a cross-validation mechanism. The system regularly injects standard test pulse signals to verify the time stamp accuracy of each channel. When inter-channel delay exceeds 1ms, a time compensation routine is automatically initiated, adjusting the data buffer using a digital delay line. To address potential packet loss during signal transmission, the system uses linear prediction to compensate for data loss, with a maximum tolerance of three consecutive frames. Time synchronization status is displayed in real time through a visual interface, including a clock deviation curve for each channel and a synchronization error statistical histogram.

[0024] The calculation process of frequency band energy entropy values ​​incorporates a quality control mechanism. The system monitors the signal-to-noise ratio (SNR) of each frequency band in real time. When the noise contamination of a particular frequency band exceeds a threshold, the entropy value for that period is automatically recalculated or marked as invalid. For periods of signal distortion caused by motion artifacts, the system compensates for these distortions by linear interpolation of the preceding and following valid data segments, with a maximum compensation duration of 1 second. Before output, the energy entropy sequence is smoothed using a moving median filter with a window width of 5 sampling points to eliminate the influence of isolated outliers.

[0025] Strict quality control is implemented during the heart rate variability feature extraction process. The R-wave detection algorithm combines amplitude thresholds with morphological template matching, and each detected R-wave is manually annotated and reviewed. In the event of consecutive missed or false detections, the system automatically adjusts detection parameters or switches to a backup detection algorithm. For periods containing arrhythmias such as premature atrial beats, the system automatically excludes these data segments from feature calculation. After feature vector generation, the Mahalanobis distance test is performed to eliminate feature combinations that fall outside the normal physiological range.

[0026] The signal preprocessing process implements multi-level quality monitoring. The raw signal is first visualized through a visual interface to display waveform quality, allowing the operator to manually mark periods of abnormality. An automated quality assessment algorithm calculates the signal-to-noise ratio, baseline drift, and peak-to-peak performance for each period, generating a signal quality index curve. For periods where the quality index falls below a threshold, the system automatically triggers a reacquisition process or prompts the operator to check electrode contact. The preprocessed signal then undergoes an independent quality verification module, based on criteria such as the rationality of frequency band energy distribution and the degree of residual artifacts.

[0027] The time synchronization system utilizes a redundant design to ensure reliability. The master clock module is equipped with dual hot-backup crystal oscillators. Automatically switching to the backup oscillator when the main crystal frequency drift exceeds 1ppm is detected. Timestamp recording utilizes both FPGA hardware timing and software timing in parallel, with discrepancies regularly compared. The Network Time Protocol service is configured in multi-server redundant mode, obtaining reference times from multiple time sources and using a majority vote mechanism to determine the final synchronization reference. Time synchronization status is integrated into the system health monitoring system, and any anomalies trigger a tiered alarm mechanism.

[0028] Example 2: See Figure 3 The cognitive load index generation unit of the cognitive state modeling module uses a deep convolutional neural network architecture to process multimodal physiological features. The network input layer receives a 76-dimensional composite feature consisting of a 64-dimensional EEG energy entropy sequence and a 12-dimensional heart rate variability feature vector. Zero-mean normalization is used to normalize the feature dimensions to the same scale. The main network structure consists of three convolutional blocks, each consisting of a one-dimensional convolutional layer, a batch normalization layer, and a LeakyReLU activation layer. The convolution kernel size is set to 5 sampling points along the time axis, and the number of output channels of each convolutional block is 32, 64, and 128, respectively. After the last convolutional block, the feature map is compressed to a 128-dimensional feature vector through a global average pooling layer. Finally, a fully connected layer and a sigmoid activation function are used to output a cognitive load index in the range [0, 1]. The network is trained using the Adam optimizer, and the loss function is defined as the mean squared error between the predicted index and the expert annotation. A dropout regularization strategy with a rate of 0.3 is applied during training to prevent overfitting.

[0029] The fluctuation rate of respiratory wave frequency amplitude is calculated using sliding window statistical analysis. The system calculates the standard deviation of the instantaneous frequency sequence of the respiratory wave signal with a window length of 10 seconds and a step size of 1 second. The instantaneous frequency estimation uses an improved peak detection algorithm. The respiratory wave signal is first low-pass filtered at 4Hz to eliminate high-frequency noise. The respiratory cycle boundaries are then determined by finding the zero crossing point of the signal's first-order derivative. The fluctuation rate within a single window is calculated using the following formula: in, Indicates the respiratory wave frequency amplitude fluctuation rate of the current window, is the number of complete respiratory cycles detected within the window, For the The instantaneous frequency of a respiratory cycle, The system performs median filtering on the calculated volatility sequence, with the window width set to 5 data points to eliminate the impact of transient abnormal fluctuations.

[0030] Detection of sudden changes in GSR amplitude utilizes a two-level criteria mechanism. Primary detection is based on the first-order difference of the signal, triggering a potential event marker when the difference between three consecutive sampling points exceeds three standard deviations of the baseline noise level. Secondary verification combines the amplitude change rate and duration criteria, requiring that a valid sudden change have a rising edge slope that lasts at least 0.5 seconds and an amplitude change that exceeds 15% of the baseline value. The system establishes a personalized baseline reference value for each pilot, derived from the statistical characteristics of the GSR signal during a resting state five minutes before training. The timestamp and amplitude change of the sudden change event are recorded as a two-dimensional feature vector, which contributes to the subsequent construction of a cognitive stability quantification matrix.

[0031] The cognitive stability quantification matrix organizes physiological indicators using a six-dimensional feature space. The rows of the matrix correspond to the time dimension, dividing the training process into equal-length segments at 30-second intervals. The columns contain six feature dimensions: mean respiratory wave frequency fluctuation, number of electrodermal changes, heart rate variability sample entropy, EEG theta / alpha power ratio, EEG global complexity index, and pupil diameter change rate. Each matrix cell stores the normalized value of each feature within the corresponding segment. Normalization uses the z-score method, with parameters derived from the statistical distribution of a large pilot database. Strict quality control is implemented during the matrix filling process. If the missing data rate for a segment exceeds 20%, the segment is marked as invalid and compensated by interpolation from adjacent segments.

[0032] The scenario control decision maker in the adaptive feedback module implements rule-based fuzzy logic control. Input variables include the first-order difference value of the cognitive load index, the current row vector of the cognitive stability matrix, and the flight phase code. The fuzzification process converts the continuous variable into three linguistic variables: cognitive load changes are categorized as "decreasing," "stable," and "increasing," and stability states are categorized as "turbulent," "fluctuating," and "stable." The knowledge base contains 27 production rules, such as "If cognitive load increases and stability is disturbed, then enhance visual stimulation." Defuzzification uses the center of gravity method to calculate the precise output value, ultimately generating a visual stimulation adjustment coefficient in the range [0, 1] and a force feedback scaling factor in the range [0.5, 2].

[0033] Visual stimulus intensity regulation implements multi-parameter coordinated control. The system adjusts the three visual dimensions of the head-up display symbols: brightness, contrast, and flicker frequency. Brightness adjustment follows the CIE 1931 brightness perception curve, with a base brightness setting of 100 cd / m² and a maximum adjustment range of ±30%. Contrast enhancement utilizes a histogram stretching algorithm, aiming for a display contrast ratio of 10:1 or greater for key instruments. The flicker frequency is controlled within a range of 4-8 Hz to avoid inducing photosensitivity reactions. The combination of adjustment parameters is dynamically selected based on the flight phase, with brightness adjustment prioritized during the cruise phase and contrast adjustment prioritized during the landing phase. All adjustment operations utilize a gradual transition, with a single adjustment range not exceeding 10% of the baseline value to avoid sudden shifts in attention.

[0034] Adjusting the joystick force feedback coefficient involves reconstructing the dynamic parameters. The basic force feedback model is a second-order mass-spring-damper system, and the damping coefficient in its dynamic equation is and stiffness coefficient Dynamically updated according to cognitive state. When the cognitive stability quantification matrix shows an increase in the level of autonomic nervous system activation, the system adjusts the parameters according to the following strategy: Damping coefficient Increase the default value from 0.4N·s / m to 0.7N·s / m to increase the motion resistance of the joystick; the stiffness coefficient The force at the center of the force is reduced from 20N / m to 15N / m, minimizing the return force. Critical damping conditions are used during parameter adjustment to maintain consistent dynamic response. Force feedback commands are implemented through a real-time control loop with an update frequency of 1kHz, ensuring continuous and accurate force feedback.

[0035] A dual-judgment mechanism is used to identify key operational nodes during the flight phase. Static nodes are preset according to standard operating procedures, including standard moments such as the start of the takeoff roll, the moment of liftoff, and altitude layer transitions. Dynamic nodes are determined through real-time analysis of flight parameters. A temporary key node marker is triggered when the airspeed change rate exceeds 0.5m / s² or the heading angle deviation exceeds 5°. The node time window is set from 10 seconds before to 20 seconds after the event. During this period, the sampling frequency of cognitive state parameters is increased from 1Hz to 4Hz. The system maintains a priority queue to manage multiple concurrent nodes to ensure that adjustment instructions from high-priority nodes can be responded to in a timely manner.

[0036] The real-time visualization of cognitive status parameters adopts a multi-level presentation design. The main monitor displays the trend curve of the cognitive load index, and the background color changes gradually according to the numerical range: green (0-0.3), yellow (0.3-0.7), and red (0.7-1). The auxiliary panel displays the current status of the cognitive stability quantification matrix in the form of a radar chart, and the numerical values ​​of the six characteristic dimensions are connected to form an irregular hexagon. The emergency warning adopts a three-level prompt: when the cognitive load index exceeds 0.8 for 30 consecutive seconds, the system triggers a sound alarm and flashes a red border; when more than three dimensions of the stability matrix are abnormal, an orange warning sign is displayed in the corresponding flight phase marking area. The position and size of all visual elements meet ergonomic standards to ensure that instructors do not need to frequently adjust their visual focus during the monitoring process.

[0037] The priority management of parameter adjustment instructions adopts dynamic weight allocation. The system assigns a basic weight to each type of adjustment instruction, with visual stimulation adjustment being 0.6 and force feedback adjustment being 0.4. The weight is dynamically adjusted according to the flight phase, and the force feedback adjustment weight is increased to 0.7 during takeoff and landing. When a command conflict is detected, the arbitration module calculates the comprehensive priority based on the flight safety margin. When the safety margin is less than 30%, the force feedback enhancement instruction is forced to be executed first. The instruction queue implements a sliding time window limit, and the interval between continuous adjustment instructions must not be less than 2 seconds to avoid parameter oscillation. All adjustment operations record detailed logs, including timestamps, adjustment types, parameter changes, and decision basis, to support post-analysis and system optimization.

[0038] Example 3: See Figure 4 The neurophysiological coordination calculation unit of the evaluation output module uses a multi-dimensional feature fusion algorithm to process cognitive state parameters. The input data includes a time-aligned cognitive load index curve and a cognitive stability quantification matrix. Both are first normalized to eliminate dimensional differences. The normalization process uses a dynamic range compression method, using the maximum and minimum values ​​in the historical data as reference boundaries. The neurophysiological coordination score is calculated using the following formula: in, represents the neurophysiological coordination score, is the normalized cognitive load index of the current period, is the deviation index of the cognitive stability quantification matrix, and The default values ​​are 0.6 and 0.4 respectively. The Mahalanobis distances between the six characteristic dimensions of the stability matrix and their ideal values ​​are calculated to reflect the overall degree of deviation. The scoring is performed using a sliding 10-second time window with a 50% window overlap, ultimately generating a coordination curve with a time resolution of 5 seconds.

[0039] The operational accuracy analysis unit implements multi-level comparison of flight trajectories. The standard flight trajectory template contains a sequence of ideal values ​​in three dimensions: heading angle, altitude, and airspeed, with a sampling interval of 1 second. The dynamic time warping algorithm uses improved constraints to limit the path search range to a time window of ±5 seconds near the diagonal. The heading angle deviation calculation converts the instantaneous difference between the actual heading and the template heading into an angle deviation through the inverse tangent function. The altitude control error is calculated using a relative error method to avoid scale problems caused by absolute values. The accuracy rating is divided into four levels: A (excellent), B (good), C (qualified), and D (unqualified). The grading threshold is dynamically adjusted according to the flight stage. For example, the tolerance range in the landing stage is 30% stricter than that in the cruise stage.

[0040] The comprehensive assessment report generation unit establishes a state-performance correlation model. The judgment of high-risk status adopts a composite condition: the neurophysiological coordination score is less than 60 points for three consecutive sampling points, and the operation accuracy rating during the same period is D. While marking the high-risk period on the timeline, the system automatically retrieves the abnormal pattern of physiological characteristics within the period. The detection of abnormal peaks of respiratory wave frequency amplitude fluctuation rate adopts the local extreme value algorithm, and defines fluctuations exceeding 2 times the standard deviation of the mean within the adjacent 60-second range as valid abnormalities. The report generation process implements automatic summarization technology to extract key event nodes, duration distribution and abnormal pattern characteristics from hours of training data to form a structured assessment conclusion.

[0041] The data visualization subsystem uses layered rendering technology to present the evaluation results. The main view displays a superimposed time series graph of the coordination score curve and the operational accuracy rating, using color coding to distinguish different flight phases. The detailed view supports single-click query of detailed data at any time point, including the EEG spectrum, physiological indicator values, and flight control inputs at that time. Abnormal event annotations are displayed using semi-transparent highlighted areas to maintain the visibility of the underlying data. All charts support dynamic zooming and panning operations, and the time resolution can be adjusted from an overview of the entire training to second-level details. Visualization parameters are stored as style templates, allowing for quick switching of display modes based on different evaluation needs.

[0042] Subsequent analysis of high-risk conditions utilizes pattern mining technology. The system automatically clusters high-risk events from historical training, extracting common temporal characteristics, flight phase distribution, and physiological indicator combination patterns. Each risk pattern is associated with a typical scenario description, such as "instrument scan failure" or "excessive control input." The pattern library supports similarity retrieval, matching emerging risk events with existing patterns to provide possible causal analysis. Risk pattern visualization uses parallel coordinate plots to display multi-dimensional feature relationships, highlighting key distinguishing dimensions.

[0043] The report export function supports multiple structured formats. Standard assessment reports are stored in JSON-LD format and include complete time series data, assessment conclusions, and metadata descriptions. Summary reports generate Word document templates that automatically populate key data tables and trend charts. Interactive reports are output in HTML5 format, with built-in dynamic visualization components supporting offline browsing. All reports include digital signatures and timestamps to ensure the authenticity and immutability of assessment results. The report distribution system implements fine-grained permission control, and users with different roles can access different report content and levels of detail.

[0044] The personalized adaptation of assessment parameters utilizes a progressive calibration approach. The system records each pilot's historical assessment data, gradually establishing an individualized scoring baseline. A new pilot's first three assessment results are used as a reference only; starting with the fourth, they are formally incorporated into their individual competency profile. The calibration process accounts for differences in flight experience, introducing aircraft-specific assessment parameter adjustments for pilots undergoing conversion training. Personalized parameters are stored in a separate configuration file, including customized settings for scoring weights, classification thresholds, and risk assessment rules.

[0045] The system integration test framework verifies the reliability of the evaluation process. The test case library contains hundreds of standard test scenarios, covering a variety of typical cognitive states and operational performance combinations. Automated test scripts simulate the entire evaluation process, verifying every step from raw data input to report generation. Performance testing monitors evaluation latency, ensuring real-time evaluation calculations are completed within 200ms. Compatibility testing verifies the system's support for acquisition devices with different sampling rates and from different manufacturers. Test results generate detailed coverage reports, identifying functional modules requiring enhanced verification.

[0046] Explanatory analysis of evaluation results utilizes explainable AI technology. Feature importance analysis reveals the reasons behind significant score changes, showing which physiological indicator changes drive these changes. Decision path visualizations illustrate the key reasoning steps in score calculation, helping to understand the system's evaluation logic. Counterfactual analysis simulates score changes after improvements in specific indicators, quantifying the impact of various factors. Explanatory analysis results are presented using natural language descriptions combined with visual charts, lowering the technical requirements.

[0047] The long-term trend analysis module tracks the evolution of pilots' abilities. Evaluation results from each training session are stored in a time series database, allowing review of score changes at weekly, monthly, and quarterly granularity. Trend forecasting utilizes an exponential smoothing algorithm to generate a projected performance curve for the next three training sessions. Abnormal trend detection identifies sudden or sustained declines in scores, triggering alerts for special attention. Trend reports compare individual performance with the average level of pilots on the same aircraft type, identifying relative strengths and areas for improvement.

[0048] A multi-rater consistency verification mechanism enhances the credibility of results. The evaluation process for key training tasks incorporates independent scoring by multiple instructors, and intraclass correlation coefficients are calculated to measure consistency between manual and system evaluations. Cases with discrepancies exceeding a threshold undergo expert review, analyzing the source of the discrepancy and updating the evaluation rules. Consistency analysis results are fed back into the evaluation model optimization process to continuously improve the system's evaluation accuracy. The verification process retains complete annotations, recording each evaluator's judgment and comments.

[0049] Example 4: See Figure 5 The training optimization suggestion module's program customization engine generates targeted training programs by analyzing the correlation pattern between neurophysiological coordination scores and operational accuracy ratings. The system first establishes a mapping table between assessment results and training elements.

[0050] Table 1: Establish a mapping relationship table between evaluation results and training elements.

[0051] The plan generation process utilizes a decision tree algorithm, with input features including trends in neurophysiological coordination scores, distribution patterns in operational accuracy ratings, and anomaly patterns in the cognitive stability quantification matrix. Each branch node in the decision tree corresponds to a key judgment condition, such as "Has the coordination score decreased by more than 15% in the last three training sessions?" Leaf nodes are associated with specific training plan templates, whose parameters are dynamically adjusted based on historical optimization results.

[0052] The visual tracking enhancement program utilizes a dynamic sight mark system. The initial shape of the sight mark is a 2cm diameter circle, colored with high-contrast yellow and black stripes. The motion trajectory algorithm includes three basic modes: linear saccade, sinusoidal tracking, and random jump. The mode ratio is automatically adjusted according to the progress of training. When the deviation between the trainee's gaze point and the center of the sight mark is continuously less than 1° of visual angle, the system gradually increases the sight mark movement speed, with each increase not exceeding 5° / s. During the training process, eye movement data is recorded in real time, and the gaze stability index and saccade accuracy are calculated as a basis for program adjustments.

[0053] The mission priority training program designs a multi-task parallel scenario. The system simultaneously presents primary flight instruments, navigation maps, and system warning information on the head-up display, requiring trainees to process each type of information according to their urgency. The mission generator dynamically adjusts the task density based on the trainee's performance, with the initial task interval set to 12 seconds, and each task lasting 3-5 seconds. The accuracy of priority judgment and response delay serve as core evaluation indicators. When the accuracy of five consecutive tasks reaches above 90%, the system shortens the task interval by 10%. The training scenario includes typical interference factors, such as simulated radio calls and sudden system warnings, to cultivate trainees' task switching capabilities.

[0054] The breathing regulation training program integrates biofeedback technology. The system superimposes a breathing guidance indicator on the flight scene, displaying the target breathing rhythm in the form of a dynamic light band. The trainee's real-time breathing waveform is collected by a chest strap sensor and compared with the target waveform. The feedback interface displays two dimensions: respiratory rate deviation and tidal volume consistency. When the deviation value exceeds the threshold, a sound prompt is triggered. The training program adopts a progressive difficulty design. Initially, only maintaining a stable breathing rate is required. Later, operational tasks such as altitude control are added to create dual-task pressure. After each training session, a breathing parameter improvement curve is generated to show the improvement in control stability at each stage.

[0055] The anti-interference training program constructs a multimodal interference environment. Visual interference sources include a high-light simulator that flashes randomly outside the cockpit, with intensity divided into four levels corresponding to different brightness variations. Auditory interference uses acoustically processed radio noise, with the signal-to-noise ratio gradually reduced from +10dB to -5dB. Tactile interference is achieved through high-frequency micro-vibration of the joystick, with the vibration frequency randomly varying within the range of 20-50Hz. The temporal relationship between interference events and key flight operations is carefully designed to avoid the formation of predictable patterns. The training effect evaluation focuses on the standard deviation of operating parameters under interference conditions, reflecting the improvement of the trainees' anti-interference ability.

[0056] Closed-loop optimization is implemented during the execution of training plans. Before each training session, the system loads a basic plan template and fine-tunes the initial parameters based on the trainee's recent performance. Key physiological indicators and operational performance are monitored in real time during training. When an expected improvement trend is detected, the training difficulty level is automatically increased; if maladaptation occurs, the training mode is downgraded or switched. Plan adjustments are recorded in a detailed log, including the adjustment time point, decision basis, and parameter changes, forming a complete optimization trajectory. After training, the system compares the actual results with the expected goals and calculates the plan effectiveness index, which is used to update the parameter mapping relationship of the decision tree nodes.

[0057] Personalized training plan recommendations utilize a collaborative filtering algorithm. The system maintains a database containing training records for hundreds of pilots. When a new trainee's assessment characteristics are similar to those of a specific group in the database, training plans that respond well to that group are prioritized. Similarity calculations consider the distribution pattern of neurophysiological characteristics, years of flight experience, and aircraft adaptation curves. The recommended results present three alternative plans, each focusing on a different optimization aspect, for the instructor's final selection. The actual results of the implemented plans are fed back to the recommendation system to continuously optimize the similarity calculation model.

[0058] Example 5: The data persistence storage module uses a hierarchical storage architecture to manage the multimodal data generated by the flight training evaluation system. The original neurophysiological data is stored in binary format, and each data packet contains an acquisition timestamp, a device identifier, and a signal quality tag. The EEG signal is processed using a lossy compression algorithm to reduce the data volume to 30% of the original size while maintaining the valid frequency band information. The ECG and respiratory signals store RR interval sequences and respiratory wave envelope data rather than complete waveform records. The galvanic skin response data records baseline values ​​and mutation event parameters to avoid storing continuously changing raw signals. All physiological data are indexed at three levels by pilot ID, training date, and mission phase, supporting millisecond-level time accuracy retrieval.

[0059] Cognitive state modeling data is persistently stored in a matrix format. The cognitive stability quantification matrix is ​​compressed and stored using sparse matrix technology, recording only non-zero elements and their positional indices. The cognitive load index curve stores key feature points rather than the complete sampling sequence, and an interpolation algorithm is used to reconstruct the original curve upon reading. The storage process implements an incremental update strategy, appending only newly added data segments during each training session to avoid duplicate storage of unmodified historical data. Data versioning utilizes snapshot technology to preserve the complete state of important evaluation nodes, while routine modifications only record the differences. Metadata description files detail the physical meaning and numerical units of matrix dimensions to ensure long-term readability.

[0060] Assessment report data is stored in a hybrid format of structured documents and relational databases. The text content of the comprehensive training assessment report is stored in Markdown format, preserving formatting while reducing storage space. Chart data is stored separately, with vector graphics saved in SVG format and bitmaps using WebP lossy compression. Raw comparison data for operational accuracy ratings is stored in a time series database format, supporting efficient range queries and similarity retrieval. High-risk status markers are associated with binary event description files, which contain trigger conditions, duration, and associated evidence chains. Reference relationships between reports are maintained in a graph database, forming a network structure of assessment knowledge.

[0061] Execution records for specialized training plans are stored using an event tracing model. The execution process of each training plan is broken down into a series of immutable event objects, which are appended to a read-only log in chronological order. Event objects contain complete context, including execution time, parameter settings, operational behavior, and result feedback. When reading, the training state at any point in time is reconstructed by replaying the event sequence. Plan adjustment records store the reasoning chain of the decision-making process, including considered alternative plans and reasons for rejection. Execution effect data and evaluation reports establish a bidirectional reference, supporting tracing back from training plans to evaluation results, or retrieving related training records from evaluation defects.

[0062] The Long-Term Evolution Tracking Database implements spatiotemporal multi-dimensional indexing. The time dimension supports three query methods: natural time, training cycle, and physiological rhythm. Data can be analyzed by calendar date, training batch, or biological clock cycle. The spatial dimension records the characteristics of the training environment, including scene parameters such as simulator model, cockpit layout, and ambient noise level. The physiological adaptation dimension tracks neuroplasticity indicators, such as the long-term drift trend of EEG frequency band energy distribution. The index structure uses a hybrid layout of B+ tree and R tree to optimize query efficiency in different dimensions. The data aging strategy automatically converts detailed raw data over five years into aggregated statistical form for storage, retaining trend information while freeing up storage space.

[0063] Data security is protected through a multi-layer encryption mechanism. Static data is encrypted and stored using the AES-256 algorithm, and the key management system separates the encryption keys from the encrypted data. The transmission process is protected using the TLS 1.3 protocol, and certificates are issued by an internal CA. Access control is based on attribute-based encryption technology, dynamically generating fine-grained data access permissions. Audit logs record all data access activities, including query content, result set size, and access time. Data de-identification removes direct personal information, and pseudonymization technology is used to maintain data relevance. The secure erase function ensures that data on retired storage media cannot be recovered.

[0064] Data consistency is maintained using a distributed transaction protocol. Cross-module data updates are atomically verified using a two-phase commit. All relevant storage nodes must either be updated successfully or rolled back. Concurrency control utilizes multi-version concurrency technology, ensuring that read operations do not block write operations, and vice versa. A conflict detection mechanism identifies concurrent modifications and employs an operation conversion algorithm to resolve editing conflicts for complex objects such as assessment reports. Regular consistency checks scan the entire database, comparing indexes with actual data and repairing any inconsistencies. Data repair tools support selective recovery, allowing data within a specific timeframe to be rolled back to a known good state.

[0065] A tiered caching strategy is employed to optimize storage system performance. Hot data, including recent training records and frequently used assessment templates, is stored in an in-memory database. Warm data, covering detailed training data from the past three months, is stored on high-speed solid-state drives. Cold data is migrated to a mechanical hard disk array to archive historical information. The cache replacement algorithm considers factors such as data access frequency, associated query patterns, and future training plans. A prefetch mechanism pre-loads historical data for relevant pilots based on training schedules. Batch operations are scheduled during system downtime, such as nighttime data compression and index rebuilding.

[0066] Data migration and compatibility are ensured through standardized intermediate formats. Cross-version migrations use intermediate dump files in JSON format, which contain complete data schema and value range descriptions. Format converters handle storage structure changes, such as field splitting or encoding changes. A backward compatibility layer allows new systems to read data in older formats and perform real-time conversions as needed. Data validation tools check the completeness and accuracy of migration results, sampling and comparing original and migrated data for consistency. Migration reports detail processing records, conversion rules, and exceptions for subsequent analysis.

[0067] Data lifecycle management implements automated policies. Original physiological data retains full accuracy for three months before being downsampled and stored, retaining only statistical features per channel per second. Intermediate evaluation results become read-only after one year, prohibiting further modification. Detailed training records older than five years automatically generate aggregate summaries, replacing the original detailed data. The destruction process adheres to secure erase standards, overwriting storage blocks multiple times. Lifecycle policies can be temporarily adjusted for specific research projects, extending the retention period of critical datasets. The policy engine supports conditional rules, such as "raw data related to high-risk events is permanently retained."

[0068] The data retrieval interface offers multilingual support. The core query API supports a subset of SQL, compatible with common data analysis tools. Domain-specific queries are encapsulated as high-level APIs, such as "Get the cognitive load fluctuations for a pilot during landing." A natural language query parser understands colloquial requests such as "Show all time periods with a coordination score below 60 over the past week." A streaming interface allows subscriptions to data updates, delivering new assessment results in real time. A batch export function generates data packages that meet aviation medical research standards and include complete metadata description documents.

[0069] The data visualization subsystem supports historical trend analysis. A timeline view overlays cognitive state parameters and operational performance over time. A heatmap matrix displays the distribution of performance characteristics at different training stages. A radar chart compares current assessment results against historical best results. Visualization configurations can be saved as templates, allowing for quick re-creation of specific analytical perspectives. Interactive exploration allows drill-down to view detailed context for any data point. Export generates interactive HTML reports with dynamic filtering and detailed prompts.

[0070] Data quality control implements full-process monitoring. The acquisition end verifies signal quality indicators and flags low-confidence data segments. The storage process verifies data integrity hash values ​​and detects transmission corruption. Timestamp continuity and numerical plausibility are checked during reading. Periodic data cleansing identifies and corrects outliers, such as drift caused by sensor failure. Quality reports summarize data loss rates and anomaly occurrence rates at each stage, guiding system improvements. Repair tools allow for the selective recalculation of derived data, such as regenerating evaluation metrics from the original signal.

[0071] Data sharing and collaboration features enable secure data exchange. Sharing policies precisely control which data fields are accessible to which roles, such as instructors seeing performance scores but not raw physiological signals. Collaborative annotation allows multiple experts to annotate the same assessment results, creating a thread of discussion. Version tracking records the modification history of shared data and supports rollback to any shared version. Watermarking technology embeds invisible identifying information in sensitive data, tracing the source of leaks. Auditing also records all data sharing activities, including recipients and access times.

[0072] Backup and disaster recovery systems ensure persistent data availability. New data is replicated to an offsite backup center in real time, with latency under 5 seconds. Daily full backups are saved as read-only snapshots, with a 30-day rolling history. Disaster recovery drills are conducted quarterly to test the ability to restore a complete system from backups. Storage hardware uses a RAID-6 configuration, which can tolerate the simultaneous failure of two disks. A power protection system ensures sufficient time for write operations to complete in the event of a power outage. Critical data is backed up in an air-gapped format on physically isolated storage media to protect against cyberattacks.

[0073] Storage system monitoring and maintenance enable automated operations. A health dashboard displays key metrics such as storage utilization, I / O latency, and error rates in real time. Predictive maintenance uses device SMART data to proactively replace disks that may fail. Performance tuning automatically adjusts database parameters to adapt to varying load patterns. Capacity planning tools predict future storage needs and recommend the timing and scale of capacity expansion. An alarm system provides a tiered response to abnormalities, ranging from email notifications to automatic switching to a backup system. A maintenance log records all configuration changes and system interventions, providing a complete operational history.

[0074] Data storage formats are designed for long-term readability. Core data uses open standard formats, such as HDF5 for scientific data and Parquet for tabular data. These self-describing formats embed metadata describing field meanings and encoding methods. Documents are stored in PDF / A format to ensure consistent rendering over time. A deprecated format detection tool regularly scans the repository, flagging data in outdated formats that require conversion. A format conversion service automatically migrates old data to currently supported formats, preserving the integrity of the original information.

[0075] The storage system architecture supports horizontal scalability. The distributed file system stores data in shards across multiple nodes, locating data blocks using a consistent hashing algorithm. The compute-storage separation architecture allows for independent scaling of compute resources and storage capacity. The stateless service design enables components to be replaced or upgraded at any time. The service discovery mechanism automatically identifies newly added storage nodes and incorporates them into the resource pool. Elastic scaling dynamically adjusts the number of active nodes based on load changes, balancing performance and energy consumption. Data rebalancing automatically adjusts data distribution as nodes are added or removed to maintain load balance.

[0076] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0077] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A flight training assessment system integrating EEG characteristics and physiological indicators, characterized by: include: A physiological signal acquisition module is used to synchronously capture the pilot's multi-channel EEG raw signals and body surface physiological index data during the simulated flight mission. The body surface physiological index data at least includes the ECG RR interval sequence, respiratory wave frequency amplitude, and skin galvanic response amplitude; A multimodal fusion module is used to perform time-frequency decomposition processing on the original EEG signal to extract energy entropy values ​​of different frequency bands, and simultaneously parse the ECG RR interval sequence to generate heart rate variability feature vectors, and establish a dynamic correlation mapping between EEG entropy values ​​and physiological feature vectors through a cross-modal attention mechanism; a cognitive state modeling module, configured to generate a cognitive load index reflecting the pilot's concentration based on the dynamic association map, and to construct a cognitive stability quantification matrix based on the respiratory wave frequency amplitude fluctuation rate; An adaptive feedback module, configured to receive the cognitive load index and the cognitive stability quantification matrix and dynamically adjust the visual stimulation intensity and joystick force feedback coefficient of the simulated flight scene according to preset key operation nodes of the flight phase; The evaluation output module is used to integrate the cognitive load index, cognitive stability quantification matrix and simulated flight operation trajectory data to generate a comprehensive training evaluation report that includes neurophysiological coordination score and operation accuracy rating.

2. The flight training assessment system integrating EEG characteristics and physiological indicators according to claim 1 is characterized by: The physiological signal acquisition module includes an EEG signal preprocessing unit for performing power frequency noise filtering and eye movement artifact correction on the multi-channel EEG raw signals, and separating effective EEG components using independent component analysis; The physiological signal acquisition module also includes a physiological signal synchronization unit, which binds the corrected EEG components, ECG RR interval series, and skin electrodermal response amplitude to the same time coordinate system through a timestamp alignment mechanism to form a time-domain aligned neurophysiological synchronization data set.

3. The flight training assessment system integrating EEG characteristics and physiological indicators according to claim 2 is characterized by: The multimodal fusion module includes a frequency domain decomposition engine that performs wavelet packet transform on the EEG components to divide the frequency bands into Delta wave, Theta wave, Alpha wave and Beta wave, and calculates the energy entropy value sequence of each frequency band; The multimodal fusion module also includes a physiological feature analysis unit, which analyzes the sample entropy value and fractal dimension of the ECG RR interval sequence through a nonlinear dynamics algorithm to generate a heart rate variability feature vector.

4. The flight training assessment system integrating EEG characteristics and physiological indicators according to claim 3 is characterized by: The cognitive state modeling module includes a cognitive load index generation unit, which receives a frequency band energy entropy value sequence and a heart rate variability feature vector, uses a convolutional neural network model to learn the nonlinear association between the energy entropy value and the heart rate variability feature vector, and outputs a cognitive load index curve that changes with flight mission time; The cognitive state modeling module also includes a stability quantification unit that integrates the respiratory wave frequency amplitude fluctuation rate and the number of sudden changes in the galvanic skin response amplitude to construct a cognitive stability quantification matrix that reflects the autonomic nervous system activation state.

5. The flight training assessment system integrating EEG characteristics and physiological indicators according to claim 4 is characterized by: The adaptive feedback module includes a scene control decision maker, which triggers visual stimulation intensity adjustment instructions and joystick force feedback coefficient adjustment instructions according to the numerical threshold of the cognitive load index curve at the key operation node of the preset flight phase; The scene control decision maker also receives data on the number of mutations in the galvanic skin response amplitude in the cognitive stability quantization matrix, and when the number of mutations exceeds a baseline threshold, increases the damping coefficient of the joystick force feedback coefficient.

6. The flight training assessment system integrating EEG characteristics and physiological indicators according to claim 5 is characterized by: The evaluation output module includes a neurophysiological coordination calculation unit, which performs normalized weighted fusion on the cognitive load index curve and the cognitive stability quantification matrix to generate a neurophysiological coordination score that reflects the degree of coordination between EEG characteristics and physiological indicators; The evaluation output module also includes an operation accuracy analysis unit, which dynamically time-regularizes the simulated flight operation trajectory data with the standard flight trajectory template, calculates the mean heading angle deviation and the altitude control error rate, and outputs the operation accuracy rating.

7. The flight training assessment system integrating EEG characteristics and physiological indicators according to claim 6 is characterized by: The evaluation output module further includes a comprehensive evaluation report generating unit that associates the neurophysiological coordination score with the operational accuracy rating, and marks the flight training phase as a high-risk state when the neurophysiological coordination score is below a critical value and the operational accuracy rating is unqualified; The comprehensive assessment report generating unit further extracts abnormal peak values ​​of respiratory wave frequency amplitude fluctuation rate in the cognitive stability quantification matrix and associates them with time nodes when high-risk states occur.

8. The flight training assessment system integrating EEG characteristics and physiological indicators according to claim 1 is characterized by: The system also includes a training optimization suggestion module that receives the comprehensive training evaluation report generated by the evaluation output module and analyzes the correlation pattern between the neurophysiological coordination score and the operation accuracy rating; The training optimization suggestion module generates a special training program for visual attention allocation or physiological stress regulation according to the correlation pattern, and feeds the special training program back to the adaptive feedback module to update the visual stimulation intensity adjustment rule.

9. The flight training assessment system integrating EEG characteristics and physiological indicators according to claim 8 is characterized by: The training optimization suggestion module includes a program customization engine that generates a visual tracking training program to enhance situational awareness when it detects that the neurophysiological coordination score is continuously below the baseline; The solution customization engine also receives steep drop point data of the cognitive load index curve, and generates anti-interference physiological adjustment training instructions based on the flight operation type corresponding to the steep drop point.

10. The flight training assessment system integrating EEG characteristics and physiological indicators according to claim 9 is characterized by: The system is deployed with a data persistence storage module, which archives the neurophysiological synchronization data set of the physiological signal acquisition module, the cognitive stability quantification matrix of the cognitive state modeling module, and the comprehensive training evaluation report of the evaluation output module in time series; The data persistence storage module is also associated with the execution records of the special training program generated by the training optimization suggestion module to form an evolution tracking database of the pilot's long-term neurophysiological adaptability.

Images