Radio frequency-based learning state detection method and device, computer device, and storage medium

Abstract

The application relates to the field of learning behavior analysis, and discloses a radio frequency-based learning state detection method and device, computer equipment and a storage medium, which comprises the following steps: acquiring distance-angle-Doppler three-dimensional data of a target person through a radio frequency sensing module; determining human posture data and breathing and heartbeat data of the target person according to the distance-angle-Doppler three-dimensional data; processing the human posture data according to a preset concentration degree evaluation method to obtain learning concentration degree evaluation data of the target person; processing the breathing and heartbeat data according to a preset anxiety state evaluation method to obtain anxiety state evaluation data of the target person; and generating learning state data of the target person according to the learning concentration degree evaluation data and the anxiety state evaluation data. The application can improve the detection effect of the learning state of the target person.

Description

Technical Field

[0001] This invention relates to the field of learning behavior analysis, and more particularly to a radio frequency-based learning state detection method, apparatus, computer device, and storage medium. Background Technology

[0002] In today's information age, self-directed learning ability has become a key component of the core competencies of teenagers. In promoting self-directed learning, how to scientifically assess learning status and provide effective intervention has become a significant challenge for the field of educational technology.

[0003] Existing learning aids mostly rely on camera image recognition or contact sensors, which suffer from limited detection dimensions. They can only acquire surface behavioral information such as posture, failing to simultaneously capture physiological parameters like breathing and heart rate, as well as deeper cognitive states such as focus and anxiety levels, resulting in incomplete perception of learning status. While wearable devices or seat pressure sensors can collect physiological signals, they require direct contact with the body, easily causing discomfort and restricting movement, affecting user compliance, especially among teenagers. Furthermore, vision-based monitoring methods require continuous collection of facial and limb images, raising privacy concerns and failing to meet the needs of protecting sensitive information of teenagers. More importantly, existing systems mostly employ passive response mechanisms, only alerting after abnormal behavior occurs, lacking the ability to proactively predict and intervene in learning status, resulting in poor timeliness of regulation. Therefore, there is an urgent need to find a new method for detecting learning status to improve its effectiveness. Summary of the Invention

[0004] This invention provides a radio frequency-based learning state detection method, apparatus, computer device, and storage medium to improve the detection effect of the learning state of a target person.

[0005] A radio frequency-based learning state detection method, comprising:

[0006] The distance-angle-Doppler three-dimensional data of the target person are obtained through the radio frequency sensing module;

[0007] The target person's posture data and respiratory and heart rate data are determined based on the distance-angle-Doppler three-dimensional data.

[0008] The human posture data is processed according to a preset attention assessment method to obtain the learning attention assessment data of the target person;

[0009] The breathing and heart rate data are processed according to a preset anxiety state assessment method to obtain the anxiety state assessment data of the target person;

[0010] The learning status data of the target person is generated based on the learning focus assessment data and the anxiety state assessment data.

[0011] Optionally, the frequency range of the radio frequency sensing module includes 60 GHz to 77 GHz;

[0012] The radio frequency sensing module is positioned 0.5m to 1.5m in front of the target person;

[0013] The sampling frequency of the radio frequency sensing module is 100Hz~200Hz.

[0014] Optionally, determining the target person's posture data and respiratory and heart rate data based on the distance-angle-Doppler three-dimensional data includes:

[0015] An electromagnetic feature space is constructed based on the distance-angle-Doppler three-dimensional data;

[0016] The target electromagnetic feature space is segmented from the electromagnetic feature space;

[0017] Locate the coordinate data of the body parts in the target electromagnetic feature space; the coordinate data of the body parts includes the key voxel set of the upper body, the coordinates of the left shoulder, the coordinates of the right shoulder, the coordinates of the waist, and the coordinates of the neck;

[0018] The human posture data is determined based on the coordinate data of the described body parts.

[0019] Optionally, determining the target person's posture data and respiratory and heart rate data based on the distance-angle-Doppler three-dimensional data includes:

[0020] Extract the center point of the thoracic cavity from the aforementioned set of key upper body voxels;

[0021] An empirical mode decomposition model is constructed based on the central point of the thoracic cavity;

[0022] The mode decomposition model is solved to obtain the respiratory and heart rate data; the respiratory and heart rate data includes respiratory waveform data and heart rate waveform data.

[0023] Optionally, the step of processing the human posture data according to a preset attention assessment method to obtain the learning attention assessment data of the target person includes:

[0024] Based on the human posture data, determine the percentage of time spent in place, the percentage of time spent in the correct sitting posture, and the percentage of time spent in the correct body orientation.

[0025] The learning focus assessment data is determined based on the percentage of time spent in the seat, the percentage of time spent in the correct sitting posture, and the percentage of time spent in the body facing the correct direction.

[0026] Optionally, the step of processing the breathing and heart rate data according to a preset anxiety state assessment method to obtain the anxiety state assessment data of the target person includes:

[0027] Respiratory time-domain features, respiratory frequency-domain features, and heartbeat time-domain features are extracted from the respiratory and heartbeat data;

[0028] Anxiety-promoting factors are determined based on the respiratory time-domain characteristics, the respiratory frequency-domain characteristics, and the heart rate frequency-domain characteristics.

[0029] Anxiety inhibition factors were determined based on the described heart rate time-domain characteristics;

[0030] The anxiety state assessment data are determined based on the anxiety promoting factors and the anxiety inhibiting factors.

[0031] Optionally, after generating the learning status data of the target person based on the learning focus assessment data and the anxiety state assessment data, the method further includes:

[0032] Obtain intervention measures that match the learning state data;

[0033] The intervention measures are executed through preset intervention components.

[0034] A radio frequency-based learning state detection device, comprising:

[0035] The data acquisition module is used to acquire distance-angle-Doppler three-dimensional data of the target person through the radio frequency sensing module;

[0036] The physiological feature extraction module is used to determine the human posture data and respiratory and heart rate data of the target person based on the distance-angle-Doppler three-dimensional data;

[0037] The focus assessment module is used to process the human posture data according to a preset focus assessment method to obtain the learning focus assessment data of the target person.

[0038] Anxiety assessment module is used to process the breathing and heart rate data according to a preset anxiety state assessment method to obtain anxiety state assessment data of the target person;

[0039] A learning status data generation module is used to generate learning status data for the target person based on the learning focus assessment data and the anxiety status assessment data.

[0040] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described radio frequency-based learning state detection method.

[0041] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described radio frequency-based learning state detection method.

[0042] The aforementioned radio frequency (RF)-based learning state detection method, device, computer equipment, and storage medium achieve synchronous monitoring of human posture and physiological signals through non-contact RF sensing, avoiding interference from wearable devices and improving user comfort and continuity. By integrating a dual-dimensional assessment of focus and anxiety levels, it can more comprehensively and objectively depict the learner's true psychological and behavioral state. The system possesses real-time dynamic analysis capabilities, which facilitates personalized learning intervention and teaching strategy optimization, improving learning efficiency and mental health management. This invention improves the detection effectiveness of the target individual's learning state. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a flowchart of a radio frequency-based learning state detection method according to an embodiment of the present invention;

[0045] Figure 2 This is a schematic diagram of the structure of an intelligent learning efficiency improver in one embodiment of the present invention;

[0046] Figure 3 This is a schematic diagram of a radio frequency-based learning state detection device according to an embodiment of the present invention;

[0047] Figure 4 This is a schematic diagram of a computer device according to an embodiment of the present invention. Detailed Implementation

[0048] 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, not all, of the embodiments of the present invention. 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.

[0049] In one embodiment, such as Figure 1 As shown, a radio frequency-based learning state detection method is provided, including the following steps S10~S50.

[0050] S10. Obtain distance-angle-Doppler three-dimensional data of the target person through the radio frequency sensing module.

[0051] Understandably, the radio frequency (RF)-based learning state detection method provided in this embodiment can be applied to an intelligent learning efficiency enhancer. The intelligent learning efficiency enhancer includes an RF sensing module and a main control processing module. The RF-based learning state detection method provided in this embodiment is primarily executed by the main control processing module.

[0052] Specifically, radio frequency (RF) sensing modules (such as 60-77 GHz millimeter-wave radar) transmit frequency-modulated continuous waves (FMCW) via an antenna array. When the electromagnetic wave encounters a target, the reflected echo is captured by the receiving antenna. The RF sensing module's internal ADC (analog-to-digital converter) converts the analog signal into a digital signal, which is then processed to obtain range-angle-Doppler three-dimensional data. This range-angle-Doppler data contains information about the target in three dimensions: range (radial position), angle (azimuth), and Doppler (radial velocity). Range-angle-Doppler three-dimensional data is a processed radar echo signal. In this context, the target could be an object requiring monitoring of learning status, such as a teenage student.

[0053] Optionally, the frequency range of the radio frequency sensing module includes 60 GHz to 77 GHz;

[0054] The radio frequency sensing module is positioned 0.5m to 1.5m in front of the target person;

[0055] The sampling frequency of the radio frequency sensing module is 100Hz~200Hz.

[0056] Understandably, the frequency range of the radio frequency (RF) sensing module is 60 GHz to 77 GHz, falling within the millimeter-wave range. This frequency range is characterized by its large bandwidth, high resolution, and strong anti-interference capabilities. The RF sensing module can be positioned 0.5 m to 1.5 m in front of the target. Within this distance range, the beamwidth of the RF sensing module can completely cover the upper body area of ​​the target. The sampling frequency of the RF sensing module can be set to 100 Hz to 200 Hz, which is beneficial for improving the robustness of physiological feature extraction.

[0057] S20. Determine the human posture data and breathing and heart rate data of the target person based on the distance-angle-Doppler three-dimensional data.

[0058] Understandably, an electromagnetic feature space can be constructed based on distance-angle-Doppler three-dimensional data, from which the coordinate data of the target person's body parts can be analyzed, and then the target person's posture data and breathing and heart rate data can be calculated.

[0059] Optionally, step S20, namely determining the human posture data and respiratory and heart rate data of the target person based on the distance-angle-Doppler three-dimensional data, includes:

[0060] S201. Construct an electromagnetic feature space based on the distance-angle-Doppler three-dimensional data;

[0061] S202. Segment the target electromagnetic feature space from the electromagnetic feature space;

[0062] S203. Locate the coordinate data of the body part in the target electromagnetic feature space; the coordinate data of the body part includes the key voxel set of the upper body, the coordinates of the left shoulder, the coordinates of the right shoulder, the coordinates of the waist, and the coordinates of the neck;

[0063] S204. Determine the human body posture data based on the location coordinate data.

[0064] Understandably, the antenna data received by the RF sensing module can be processed using both range-FFT (Fast Fourier Transform) and angle-FFT. Range-FFT transforms the signal from the time domain to the frequency domain, thus resolving the distance between the target and the antenna; angle-FFT (or beamforming) is used to resolve the azimuth angle of the target relative to the antenna. The range-angle-Doppler three-dimensional data includes the range data obtained after range-FFT processing and the angle data obtained after angle-FFT processing. Combining the range and angle data allows the construction of the electromagnetic feature space E.

[0065] By analyzing the amplitude information at each point in the electromagnetic feature space E, it is possible to determine whether a person is present and to preliminarily determine the approximate location of the target. After locking onto the target person, the data corresponding to the target area in the electromagnetic feature space E is segmented to obtain the target electromagnetic feature space E. T .

[0066] Based on the target electromagnetic feature space The target voxel filtering method can be used to accurately locate key points in multiple parts of the human body and generate part coordinate data. The part coordinate data includes the set of key voxels for the upper body, coordinates of the left shoulder, coordinates of the right shoulder, coordinates of the waist, and coordinates of the neck.

[0067] In one example, the target voxel screening method includes:

[0068]

[0069] in A screening factor for the target voxel; this factor is used to measure the degree of low-frequency motion caused by respiration contained in a voxel in the electromagnetic feature space. The calculation method is as follows: within the frequency range of 0.1-0.4Hz (the human breathing frequency range), find the maximum value of the signal amplitude after phase expansion and Fourier transform, and then calculate the ratio of this maximum value to the total amplitude over the entire frequency range.

[0070] This represents the target electromagnetic feature space segmented from the electromagnetic feature space E;

[0071] Indicates the target space voxel E T Phase unrolling is performed; the phase changes of radar echo signals are extremely sensitive to minute movements. Phase unrolling can solve the phase ambiguity problem, thereby more accurately recovering the true phase change history caused by periodic movements such as breathing.

[0072] This means performing a Fourier transform on the expanded phase signal to convert it from the time domain to the frequency domain, thus obtaining its frequency response. f The amplitude at that location is used to analyze the motion spectrum at that voxel position;

[0073] This indicates the search for the maximum spectral amplitude within the breathing frequency band (0.1-0.4Hz). Breathing is a low-frequency, periodic movement, and its energy is concentrated in this frequency band.

[0074] This represents the sum of the spectral amplitudes across the entire effective frequency range, used for normalization, such that... It becomes a relative ratio, unaffected by absolute signal strength.

[0075] By setting an empirical threshold (e.g., 0.7), voxels larger than the threshold are used as the upper torso voxel set. In one example, It can be represented as:

[0076]

[0077] in, for Voxels within.

[0078] This allows for the location of key points in multiple parts of the human body. For example, the center point of the chest cavity. c Defined as Medium-voxel screening factor The voxel point with the largest value. This is because the amplitude of motion at the center of the chest cavity is usually the most significant during respiration, corresponding to... The value should also be the highest. E torsoThe topmost voxel is defined as the neck keypoint, the leftmost and rightmost voxels are defined as the left shoulder keypoint and the right shoulder keypoint, respectively, and the bottommost voxel is defined as the waist keypoint.

[0079] After determining the coordinate data of the body parts, the human posture data can be determined based on this data. For example, based on the coordinates of the key points of the neck and waist, the degree of tilt in a sitting posture can be calculated; based on the coordinates of the key points of the left and right shoulders and the key point of the neck, the degree of shoulder inward rotation can be calculated, as follows:

[0080]

[0081]

[0082]

[0083]

[0084]

[0085] in Indicates the coordinate positions of key points in the neck in three-dimensional space;

[0086] Indicates the coordinate position of key points in the waist in three-dimensional space;

[0087] This indicates the coordinate position of the key point on the left shoulder in three-dimensional space;

[0088] This indicates the coordinate position of the key point on the right shoulder in three-dimensional space;

[0089] This represents the spatial vector pointing from the waist keypoint to the neck keypoint, with components as follows: (left and right directions) (up and down direction), (Front-back direction), approximately reflects the orientation of the spine;

[0090] Indicates the angle of inclination while sitting; The larger the number, the more the spine deviates from the vertical direction, and the more severe the posture tilt.

[0091] This is a spatial unit vector between the key points of the left and right shoulders, with the direction pointing from the left shoulder to the right shoulder, representing the direction of the shoulder line;

[0092] bThe vector is a spatial unit vector between the key points of the neck and the key points of the left and right shoulders, with the direction pointing from the midpoint of the line connecting the two shoulders to the key points of the neck, which approximately reflects the orientation of the front of the chest cavity.

[0093] Calculated for Euclidean distance;

[0094] The unit vector along the y-axis;

[0095] This refers to the inward rotation angle of the shoulders. This reflects the degree of forward and inward shoulder tilt. The larger the angle, the more pronounced the inward shoulder tilt and hunched posture.

[0096] For example, by setting thresholds for the tilt angle and shoulder inward rotation angle, it can be determined whether a user's posture is abnormal. Specifically:

[0097]

[0098] 1(⋅) is an indicator function that evaluates to 1 when the condition inside the parentheses is true, and 0 otherwise.

[0099] This indicates that either the absolute value of the sitting posture tilt angle or the shoulder inward rotation angle is greater than or equal to 15 degrees.

[0100] when When the parameter is 1, it means that the sitting posture tilt angle or shoulder inward rotation angle exceeds the threshold of the correct sitting posture, which is considered an abnormal sitting posture.

[0101] This embodiment constructs and segments an electromagnetic feature space to accurately extract the spatial characteristics of the target person, effectively suppressing environmental interference and improving the signal-to-noise ratio and positioning accuracy. Based on key voxels and body part coordinates, it achieves fine-grained modeling of the upper body posture, providing high-precision structured data support for subsequent attention assessment. The method for acquiring human posture data and respiratory and heart rate data provided in this embodiment does not require cameras or wearable devices, ensuring privacy and security while supporting stable perception in complex scenarios such as low light and occlusion. It balances physical interpretability and algorithmic robustness, significantly improving the practicality and deployment flexibility of non-contact learning state monitoring.

[0102] Optionally, step S20, namely determining the human posture data and respiratory and heart rate data of the target person based on the distance-angle-Doppler three-dimensional data, includes:

[0103] S205. Extract the center point of the thoracic cavity from the set of key voxels of the upper body;

[0104] S206. Construct an empirical mode decomposition model based on the center point of the thoracic cavity;

[0105] S207. Solve the modal decomposition model to obtain the respiratory and heartbeat data; the respiratory and heartbeat data includes respiratory waveform data and heartbeat waveform data.

[0106] Understandably, after obtaining the set of key voxels of the upper body, an empirical mode decomposition model can be constructed based on the center point of the chest cavity in the set of key voxels of the upper body, and then the respiratory and heart rate data can be solved.

[0107] Specifically, using the central voxel point of the thoracic cavity c Construct a multi-channel subspace with a size of 3*3 neighborhood voxels. Each channel's radio frequency echo signal is .

[0108] Bandpass filtering is performed on each channel to obtain the bandpass signal. And calculate the quality score of the channel data based on the energy percentage. And normalize to obtain the weights Candidate signals are obtained by weighted fusion of multi-channel subspace data within the respiratory and heart rate frequency bands. , represented as:

[0109]

[0110]

[0111]

[0112]

[0113]

[0114]

[0115] in Indicates a bandpass filter. Let represent the set of samples in a sliding window centered at time t and of length q. Indicates energy within the frequency band. Indicates the total energy of the channel; It is a very small positive number used to prevent the denominator from being zero; for example, it could be 10. -6 .

[0116] Construct an adaptive parameter empirical mode decomposition method for radio frequency echo to extract the target respiratory waveform within a time period. With heartbeat waveform , represented as:

[0117]

[0118]

[0119] in The candidate signal after adding n times Gaussian white noise;

[0120] This indicates that the radar echo data has been averaged.

[0121] , where IMF is the participating component after the first empirical mode decomposition. IMF is a function that satisfies the following two conditions: 1) The number of extreme points is equal to or differs by at most one from the number of zero-crossing points throughout the entire data segment; 2) At any point, the mean of the upper envelope determined by the local maxima and the lower envelope determined by the local minima is zero. It is the original signal x ( t ) and the first residual component r 1( c The difference between the two values ​​contains the highest frequency oscillation component of the signal.

[0122] The residual components obtained from k decompositions are subtracted to obtain the kth empirical mode decomposition component. This process continues until the residual component becomes a monotonic function, at which point the signal decomposition ends. The candidate signal can be represented as:

[0123]

[0124] Two empirical mode components, one within the respiratory frequency band and the other within the heart rate frequency band, were selected as the target respiratory waveforms. With heartbeat waveform The output.

[0125] This embodiment processes radar echo signals using an empirical mode decomposition model. It does not require pre-defined basis functions and can decompose signals based on the characteristics of the signal itself, making it highly suitable for processing radar echo signals generated by non-stationary and non-linear physiological activities such as breathing and heartbeat. The method provided in this embodiment can effectively separate weak vital sign signals from complex radar echoes.

[0126] S30. Process the human posture data according to the preset attention assessment method to obtain the learning attention assessment data of the target person.

[0127] Understandably, preset attention assessment methods can be set according to actual needs. By using preset attention assessment methods, human posture data can be evaluated to generate learning attention assessment data for the target individual.

[0128] Optionally, step S30, namely, processing the human posture data according to a preset attention assessment method to obtain the learning attention assessment data of the target person, includes:

[0129] S301. Determine the percentage of time spent in place, the percentage of time spent in correct sitting posture, and the percentage of time spent in the correct body orientation based on the human posture data.

[0130] S302. Determine the learning focus assessment data based on the percentage of time spent in the position, the percentage of time spent in the correct sitting posture, and the percentage of time spent in the body facing the correct direction.

[0131] Understandably, after obtaining human posture data, the percentage of time spent in the correct sitting position, the percentage of time spent in the correct sitting position, and the percentage of time spent in the correct body orientation can be calculated separately. Then, an attention assessment score can be calculated, specifically:

[0132]

[0133]

[0134]

[0135]

[0136]

[0137]

[0138] in, L Set the window length to 60 seconds.

[0139] The anomaly index is based on Doppler energy;

[0140] The energy ratio of the Doppler information in the anomaly;

[0141] Energy is lost due to momentary anomalies;

[0142] This represents the total energy in the low-frequency band.

[0143] As a balance coefficient, when the radio frequency module detects a target, ;

[0144] Indicates the time point at which the timing begins;

[0145] This represents the percentage of time a person is in possession, when the radio frequency module detects the target person. ; This indicates the probability that the radio frequency module detects the target person.

[0146] For correct sitting posture, the dwell time ratio;

[0147] e y and e z They are respectively y shaft and z Unit vector along the axial direction;

[0148] Indicates the spatial coordinates of the key points on the right shoulder;

[0149] Represents the spatial coordinates of the key points on the left shoulder;

[0150] This represents a spatial vector pointing from the left shoulder keypoint to the right shoulder keypoint, which defines the direction of the shoulder line in the human body.

[0151] Indicates the angle of the body's orientation;

[0152] The ratio of body facing forward to dwell time;

[0153] and The weights for calculating the correct sitting posture dwell ratio and the body facing forward dwell ratio are respectively.

[0154] This represents the focus score. In one example, when... If the score is less than 50, it is considered as insufficient focus. The intelligent learning efficiency enhancer will remind you and provide sound, light, and wind interaction.

[0155] This embodiment constructs an interpretable, multi-dimensional attention assessment model by quantifying three behavioral indicators: attention duration, posture conformity, and body orientation stability, closely aligning with the needs of real-world learning scenarios. The calculation of these three indicators is based on objective posture data, avoiding subjective questionnaire bias and improving the objectivity and continuity of the assessment results. Furthermore, this embodiment does not rely on facial or eye tracking, reducing privacy risks, and adapts to common learning behaviors such as occlusion and head-down posture, enhancing system robustness. It provides educators and learners with real-time feedback, facilitating personalized learning management and attention training interventions.

[0156] S40. Process the breathing and heart rate data according to the preset anxiety state assessment method to obtain the anxiety state assessment data of the target person.

[0157] Understandably, after obtaining respiratory and heart rate data, the data is evaluated according to a pre-defined anxiety assessment method to obtain the target person's anxiety assessment data.

[0158] Optionally, step S40, namely, processing the breathing and heart rate data according to a preset anxiety state assessment method to obtain the anxiety state assessment data of the target person, includes:

[0159] S401. Extract respiratory time-domain features, respiratory frequency-domain features, heartbeat time-domain features, and heartbeat frequency-domain features from the respiratory and heartbeat data;

[0160] S402. Determine anxiety-promoting factors based on the respiratory time-domain characteristics, the respiratory frequency-domain characteristics, and the heart rate frequency-domain characteristics;

[0161] S403. Determine the anxiety inhibition factor based on the heartbeat time-domain characteristics;

[0162] S404. Determine the anxiety state assessment data based on the anxiety promoting factor and the anxiety inhibiting factor.

[0163] Understandably, anxiety-promoting and anxiety-inhibiting factors can be analyzed based on the time-frequency domain characteristics of breathing and heart rate information, thereby determining anxiety state assessment data. Anxiety state assessment data can be represented by anxiety level scores.

[0164] Respiratory time-domain features, respiratory frequency-domain features, heart rate time-domain features, and heart rate frequency-domain features can be extracted from respiratory and heart rate data. Among these, the respiratory time-domain features include the median dominant respiratory frequency. Coefficient of variation of respiratory rate Respiratory frequency domain characteristics include the normalized power spectrum within the respiratory frequency band. The time-domain features of heartbeat information include RMSSD, SDNN, and PNN50; the frequency-domain features of heartbeat information include low-frequency power. LF Compared to low-frequency high-frequency power ratio LF / HF The time window L can be set to 60 seconds. The specific calculation process is as follows:

[0165]

[0166]

[0167]

[0168]

[0169]

[0170]

[0171] in This represents the cycle of each breath in the respiratory waveform;

[0172] and These represent calculations using the median and standard deviation, respectively.

[0173] The normalized spectrum within the respiratory frequency band;

[0174] Indicates anxiety-promoting factors;

[0175] Indicates anxiety-inhibiting factors;

[0176] and These represent the calculations of the mean and standard deviation of the features, respectively.

[0177] This represents an anxiety level score. In one example, when... When the value is greater than 50, it is considered to be in an anxious state, and the intelligent learning efficiency enhancer will provide reminders and interactive sound, light, and wind.

[0178] This embodiment constructs a facilitator-inhibitor dual-factor assessment mechanism by integrating multi-dimensional time-frequency features of respiration and heartbeat, which more scientifically reflects the physiological dynamic balance of anxiety and improves recognition accuracy. Simultaneously, it utilizes non-contact radio frequency signals to extract physiological parameters, avoiding the discomfort and interference of traditional electrodes or wearable devices, and supporting long-term, imperceptible monitoring. Furthermore, by distinguishing between facilitators and inhibitors, the model's adaptability to individual differences and situational fluctuations is enhanced, reducing the misjudgment rate and providing quantitative evidence for emotional intervention during the learning process, thus contributing to learning status monitoring and the construction of adaptive learning environments.

[0179] S50. Generate the learning status data of the target person based on the learning focus assessment data and the anxiety state assessment data.

[0180] Understandably, learning attention assessment data and anxiety state assessment data can be combined to generate learning status data for a target individual. Learning status data includes both learning attention assessment data and anxiety state assessment data. In some examples, learning status data can be represented by a learning status score, which can be a weighted sum of the learning attention assessment data and the anxiety state assessment data. The weights of the learning attention assessment data and the anxiety state assessment data can be determined based on the actual experiment.

[0181] In steps S10-S50, non-contact radio frequency sensing enables simultaneous monitoring of human posture and physiological signals, avoiding interference from wearable devices and improving comfort and continuity of use. The integration of attention and anxiety state assessments provides a more comprehensive and objective portrayal of the learner's true psychological and behavioral state. The system possesses real-time dynamic analysis capabilities, facilitating personalized learning intervention and teaching strategy optimization, thereby improving learning efficiency and mental health management.

[0182] Optionally, after step S50, that is, after generating the learning status data of the target person based on the learning focus assessment data and the anxiety state assessment data, the method further includes:

[0183] S60. Obtain intervention measures that match the learning state data;

[0184] S70. The intervention measures are executed through the preset intervention components.

[0185] Understandably, preset intervention components refer to device components that can intervene in the learning state of a target person, such as airflow adjustment components, lighting adjustment components, and sound adjustment components. Airflow adjustment components include, but are not limited to, fans and air conditioners; lighting adjustment components include, but are not limited to, multi-colored light strips and ambient light sources; sound adjustment components include, but are not limited to, multi-channel speakers. In some examples, preset intervention components also include a display screen to provide the target person with instructional information, such as time information and learning state data.

[0186] The system can pre-configure the correlation between learning status data and intervention measures. After acquiring the learning status data of the target individual, the system matches the corresponding intervention measures from the correlation and then executes the intervention measures in the preset intervention component. Intervention measures can include intervention methods and intervention parameters. For example, when the learning status data indicates abnormal sitting posture, the intervention measures include: activating intermittent gentle breezes (every 3-5 seconds), flashing a 2Hz red light, and providing soft prompts to remind the learner; when the learning status data indicates insufficient focus, the intervention measures include: outputting a continuous, soft gust of wind at 0.5-1 m / s and playing 30-50 dB ambient sounds to improve attention; when the learning status data indicates anxiety, the intervention measures include: outputting a continuous, soft gust of wind at 0.5-1 m / s and playing 30-50 dB ambient sounds to improve attention; when the learning status data indicates abnormal presence, the system intervenes through voice reminders and a 5-second screen pop-up, repeating every 10 seconds to ensure that the learner returns to the learning state in a timely manner, achieving a closed-loop regulation of "perception-judgment-intervention".

[0187] This embodiment achieves closed-loop management of "detection-analysis-intervention," transforming learning status data into personalized intervention strategies in real time, significantly improving the intelligence and responsiveness of education. By automatically executing measures through preset intervention components, it reduces the cost of manual intervention and enhances the adaptability of the learning experience. Simultaneously, it addresses the dual goals of improving concentration and soothing emotions, promoting the synergistic optimization of cognitive efficiency and mental health, driving the evolution of the educational scenario from passive observation to proactive adjustment, and constructing a more scientific and humanistic intelligent learning support system.

[0188] In one example, such as Figure 2As shown, the intelligent learning efficiency enhancer has a humanoid structure, including a head 1, a torso and limbs 2, and a base 3. The head 1 is equipped with radio frequency (RF) sensing components, a multi-colored light strip, a fan, and a speaker. The multi-colored light strip is arranged around the RF module and intersects with the front air vent to form ambient lighting, used for indication and interactive flashing when the learning plan is completed. The RF sensing module operates at a frequency between 24-81 GHz. The intelligent learning efficiency enhancer has a main control processing module inside, connected to the RF sensing module. The main control processing module can be a microcontroller.

[0189] The torso and limbs section 2 is equipped with a display screen and control buttons. The display screen can be set to two display areas. One display area is a clock section, displaying the date, time, prompts, etc. The date and time are continuously displayed by a long-life battery in the base when no external power is available. The other display area displays information such as breathing rate, distance between the learner and the enhancement device, concentration status, and duration.

[0190] The base 2 is equipped with a quick-set rotary knob, an ambient light sensor module, and a built-in long-life battery. The quick-set rotary knob is mainly used to quickly adjust digital settings such as the Pomodoro Technique; the ambient light sensor module's light sensor is located on the learner's side above the base shell, slightly tilted towards the learner to reduce dust and interference, and to alert the learner to changes in ambient light.

[0191] The radio frequency sensing module can detect the posture and physiological information of a target person (such as a student), thereby detecting incorrect posture, concentration, and anxiety. It interacts with the target person using sound, light, and wind, and includes various built-in soft background sounds, flashing modes, and wind models to regulate and motivate the target person through physical stimulation, improving efficiency and concentration.

[0192] In one application example, the smart learning efficiency enhancer is placed directly in front of a teenager's study desk, 0.6 meters away. The teenager sets parameters via touch control buttons or by connecting to an app via Bluetooth / WiFi. After setting the parameters, keeping the enhancer in the same position, its radio frequency sensing module outputs encoded distance, breathing, and posture characteristics to the main control processing module, which displays the data on the screen and compares it with the set parameters. If poor posture is detected, the fan blows intermittently, and the speaker plays a soft tapping sound as a reminder, with an interval of 3-5 seconds. The number of reminders can be set, and the reminder can be stopped by touching the control buttons.

[0193] When the radio frequency sensing module detects a decrease in learning efficiency through changes in breathing rhythm, the fan outputs a gentle breeze for a period of time to create a dynamic microclimate environment, increasing the feeling of airflow and improving learning efficiency. The speaker plays soft, simulated forest and quiet morning ambient sounds to avoid distractions caused by sudden noises in the learning environment. This music can be selected according to the preferences of teenagers and supports downloading and volume adjustment. This function can be stopped by double-tapping the designated control button.

[0194] In addition, the focus time training function can help learners efficiently complete their study plans and conduct focus training. In wake-up mode, tap the touch control button, then turn the quick-set rotary knob to set the study plan time. In the planning mode, you can use the control button to select the appropriate mode according to the on-screen prompts. Tap the touch control button again to start the countdown for the planned time. When the countdown ends, the sound and light interaction module plays the corresponding music and the item flashes, and the task ends.

[0195] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0196] In one embodiment, a radio frequency (RF)-based learning state detection device is provided, which corresponds one-to-one with the RF-based learning state detection method described in the above embodiments. For example... Figure 3 As shown, the radio frequency-based learning state detection device includes:

[0197] The data acquisition module 10 is used to acquire distance-angle-Doppler three-dimensional data of the target person through the radio frequency sensing module;

[0198] The physiological feature extraction module 20 is used to determine the human posture data and respiratory and heart rate data of the target person based on the distance-angle-Doppler three-dimensional data.

[0199] The focus assessment module 30 is used to process the human posture data according to a preset focus assessment method to obtain the learning focus assessment data of the target person.

[0200] Anxiety assessment module 40 is used to process the breathing and heart rate data according to a preset anxiety state assessment method to obtain anxiety state assessment data of the target person.

[0201] The learning status data generation module 50 is used to generate learning status data of the target person based on the learning focus assessment data and the anxiety status assessment data.

[0202] Optionally, the frequency range of the radio frequency sensing module includes 60 GHz to 77 GHz;

[0203] The radio frequency sensing module is positioned 0.5m to 1.5m in front of the target person;

[0204] The sampling frequency of the radio frequency sensing module is 100Hz~200Hz.

[0205] Optionally, the physiological feature extraction module 20 includes:

[0206] Construct feature space units to build an electromagnetic feature space based on the distance-angle-Doppler three-dimensional data;

[0207] The target space unit is used to segment the target electromagnetic feature space from the electromagnetic feature space;

[0208] A body part coordinate positioning unit is used to locate body part coordinate data in the target electromagnetic feature space; the body part coordinate data includes the upper body key voxel set, left shoulder coordinate, right shoulder coordinate, waist coordinate, and neck coordinate;

[0209] A posture data unit is used to determine the human posture data based on the location coordinate data.

[0210] Optionally, the physiological feature extraction module 20 includes:

[0211] The chest cavity center point unit is used to extract the chest cavity center point from the set of key upper body voxels;

[0212] A modal decomposition model is constructed to build an empirical modal decomposition model based on the central point of the thoracic cavity;

[0213] A respiratory and heart rate data acquisition unit is used to solve the mode decomposition model to obtain the respiratory and heart rate data; the respiratory and heart rate data includes respiratory waveform data and heart rate waveform data.

[0214] Optionally, the focus assessment module 30 includes:

[0215] The indicator calculation unit is used to determine the percentage of time spent in place, the percentage of correct sitting posture, and the percentage of time spent facing the correct body orientation based on the human posture data.

[0216] The focus calculation unit is used to determine the learning focus assessment data based on the percentage of time spent in the position, the percentage of time spent in the correct sitting posture, and the percentage of time spent in the body facing the correct direction.

[0217] Optionally, the anxiety assessment module 40 includes:

[0218] The respiratory and heartbeat feature extraction unit is used to extract respiratory time-domain features, respiratory frequency-domain features, heartbeat time-domain features, and heartbeat frequency-domain features from the respiratory and heartbeat data.

[0219] An anxiety-promoting factor calculation unit is used to determine anxiety-promoting factors based on the respiratory time-domain features, the respiratory frequency-domain features, and the heart rate frequency-domain features.

[0220] An inhibition factor calculation unit is used to determine an anxiety inhibition factor based on the heartbeat time-domain characteristics.

[0221] An anxiety state assessment unit is used to determine the anxiety state assessment data based on the anxiety promoting factor and the anxiety inhibiting factor.

[0222] Optionally, the learning state detection device further includes:

[0223] An intervention matching module is used to obtain intervention measures that match the learning state data;

[0224] An intervention execution module is used to execute the intervention measures through preset intervention components.

[0225] Specific limitations regarding the radio frequency (RF)-based learning state detection device can be found in the limitations of the RF-based learning state detection method described above, and will not be repeated here. Each module in the aforementioned RF-based learning state detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0226] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and the database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores data related to the radio frequency-based learning state detection method. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements a radio frequency-based learning state detection method.

[0227] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the radio frequency-based learning state detection method described in the above embodiment; to avoid repetition, this will not be repeated here. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the embodiment of the radio frequency-based learning state detection device; to avoid repetition, this will not be repeated here.

[0228] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When executed by a processor, the computer program implements the radio frequency-based learning state detection method described in the above embodiment. To avoid repetition, this will not be described again here. Alternatively, when executed by a processor, the computer program implements the functions of each module / unit in the above embodiment of the radio frequency-based learning state detection device. To avoid repetition, this will not be described again here.

[0229] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0230] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0231] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A radio frequency-based learning state detection method, characterized by, The method comprises the steps of: acquiring distance-angle-Doppler three-dimensional data of a target person through a radio frequency sensing module; determining body posture data and breathing and heartbeat data of the target person according to the distance-angle-Doppler three-dimensional data; processing the body posture data according to a preset concentration evaluation method to obtain learning concentration evaluation data of the target person; processing the breathing and heartbeat data according to a preset anxiety state evaluation method to obtain anxiety state evaluation data of the target person; generating learning state data of the target person according to the learning concentration evaluation data and the anxiety state evaluation data.

2. The radio frequency-based learning state detection method of claim 1, wherein, The frequency range of the radio frequency sensing module comprises 60GHz-77GHz; The radio frequency sensing module is arranged at a position 0.5m-1.5m in front of the target person; The sampling frequency of the radio frequency sensing module is 100Hz-200Hz.

3. The radio frequency-based learning state detection method of claim 1, wherein, The method of determining the body posture data and the breathing and heartbeat data of the target person according to the distance-angle-Doppler three-dimensional data comprises the steps of: constructing an electromagnetic feature space according to the distance-angle-Doppler three-dimensional data; segmenting a target electromagnetic feature space from the electromagnetic feature space; locating part coordinate data in the target electromagnetic feature space; the part coordinate data comprises an upper body key voxel set, a left shoulder coordinate, a right shoulder coordinate, a waist coordinate and a neck coordinate; determining the body posture data according to the part coordinate data.

4. The radio frequency-based learning state detection method of claim 3, wherein, The method of determining the body posture data and the breathing and heartbeat data of the target person according to the distance-angle-Doppler three-dimensional data comprises the steps of: extracting a chest center point from the upper body key voxel set; constructing an empirical mode decomposition model based on the chest center point; solving the mode decomposition model to obtain the breathing and heartbeat data; the breathing and heartbeat data comprises breathing waveform data and heartbeat waveform data.

5. The radio frequency based learning state detection method of claim 1, wherein, The method of processing the body posture data according to the preset concentration evaluation method to obtain the learning concentration evaluation data of the target person comprises the steps of: determining an in-position time length proportion, a correct sitting posture residence ratio and a body forward orientation residence ratio according to the body posture data; determining the learning concentration evaluation data according to the in-position time length proportion, the correct sitting posture residence ratio and the body forward orientation residence ratio.

6. The radio frequency-based learning state detection method of claim 1, wherein, The method of processing the breathing and heartbeat data according to the preset anxiety state evaluation method to obtain the anxiety state evaluation data of the target person comprises the steps of: extracting breathing time domain features, breathing frequency domain features, heartbeat time domain features and heartbeat frequency domain features from the breathing and heartbeat data; determining an anxiety promoting factor according to the breathing time domain features, the breathing frequency domain features and the heartbeat frequency domain features; determining an anxiety inhibiting factor according to the heartbeat time domain features; determining the anxiety state evaluation data according to the anxiety promoting factor and the anxiety inhibiting factor.

7. The radio frequency-based learning state detection method of claim 1, wherein, After the learning state data of the target person is generated according to the learning concentration evaluation data and the anxiety state evaluation data, the method further comprises the steps of: obtaining an intervention measure matched with the learning state data; executing the intervention measure through a preset intervention component.

8. A radio frequency-based learning state detection apparatus, characterized by comprising: The method comprises the steps of: An acquisition data module is configured to acquire distance-angle-Doppler three-dimensional data of a target person through the radio frequency sensing module; A physiological feature extraction module is configured to determine human posture data and breathing and heartbeat data of the target person according to the distance-angle-Doppler three-dimensional data; A concentration evaluation module is configured to process the human posture data according to a preset concentration evaluation method to obtain learning concentration evaluation data of the target person; An anxiety evaluation module is configured to process the breathing and heartbeat data according to a preset anxiety state evaluation method to obtain anxiety state evaluation data of the target person; A learning state data generation module is configured to generate learning state data of the target person according to the learning concentration evaluation data and the anxiety state evaluation data.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the radio frequency-based learning state detection method according to any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the radio frequency-based learning state detection method according to any one of claims 1-7.

Images