Search

CN-122004898-A - Multi-scene fatigue state detection system and method based on electroencephalogram signals and wearable cap body

CN122004898ACN 122004898 ACN122004898 ACN 122004898ACN-122004898-A

Abstract

The invention provides a multi-scene fatigue state detection system and method based on an electroencephalogram signal and a wearable cap body, which relate to the technical field of biological signal detection and comprise a processing module, a processing module and a control module, wherein the processing module is used for preprocessing an original electroencephalogram signal of a scalp area of a user to obtain a conditioned electroencephalogram signal; the device comprises a flexible dry electrode, an electrode channel, an extraction module, a touch state analysis module and an electrode distribution configuration entropy calculation module, wherein the touch state analysis module is used for carrying out touch state analysis on the conditioned brain electrical signals, the touch stability index corresponding to each flexible dry electrode is calculated to determine effective brain electrical signals after screening, and the extraction module is used for extracting a first phase lock point, a second phase lock point and a third phase lock point from electrode channels corresponding to the effective brain electrical signals after screening, constructing an elliptic space potential field curved surface and calculating the electrode distribution configuration entropy. The invention realizes detection, comprehensive evaluation and hierarchical alarm of the fatigue state of the user under multiple scenes, can synchronously transmit the judgment result, improves the comprehensiveness and accuracy of fatigue monitoring, and ensures the operation safety of the user.

Inventors

  • ZHANG HONGZHAO
  • ZHANG HONGZHAO
  • GAO JIANQUAN
  • ZHENG JINGYING

Assignees

  • 广东云脑智能科技有限公司

Dates

Publication Date
20260512
Application Date
20260408

Claims (10)

  1. 1. Multi-scene fatigue state detection system based on brain electrical signals and wearable cap body, which is characterized by comprising: the processing module is used for preprocessing the original electroencephalogram signal of the scalp area of the user to obtain a conditioned electroencephalogram signal, analyzing the contact state of the conditioned electroencephalogram signal, and calculating the contact stability index corresponding to each flexible dry electrode so as to determine the effective electroencephalogram signal after screening; the extraction module is used for extracting the first phase lock point, the second phase lock point and the third phase lock point from the electrode channel corresponding to the screened effective electroencephalogram signal, constructing an elliptic space potential field curved surface and calculating an electrode distribution configuration entropy value; The fusion module is used for dynamically reconstructing the fusion weight of the screened effective electroencephalogram signals according to the entropy value of the electrode distribution configuration to obtain the reconstructed fusion weight, and carrying out weighted fusion on the screened effective electroencephalogram signals to generate a comprehensive electroencephalogram signal sequence; The analysis module is used for carrying out time-frequency and time-domain waveform analysis on the comprehensive electroencephalogram signal sequence, calculating a first fatigue characterization index and determining peripheral behavior fatigue symptom characteristics; The evaluation module is used for inputting the first fatigue characterization index, the peripheral behavior fatigue symptom feature, the multidimensional environment and the operation time sequence feature set into a pre-trained evaluation model to obtain a comprehensive fatigue grade judgment result and trigger corresponding grading alarm measures, and transmitting the comprehensive fatigue grade judgment result to an external monitoring platform or a local monitoring terminal.
  2. 2. The multi-scene fatigue state detection system based on the electroencephalogram signal and the wearable cap according to claim 1, wherein the original electroencephalogram signal of the scalp region of the user is acquired through a flexible dry electrode integrated at the forehead middle position, a flexible dry electrode at the left temporal region position and a flexible dry electrode at the right temporal region position in the wearable cap.
  3. 3. The multi-scene fatigue state detection system based on electroencephalogram signals and wearable cap according to claim 2, wherein the contact state analysis is performed on the conditioned electroencephalogram signals, and the contact stability index corresponding to each flexible dry electrode is calculated to determine the effective electroencephalogram signals after screening, and the system comprises: respectively carrying out contact state analysis on the conditioned electroencephalogram signals according to the electrode channels corresponding to each flexible dry electrode, continuously collecting a contact impedance value sequence of each electrode channel in a preset time window, calculating the mean value and variance of the contact impedance value sequence, and calculating the contact stability index of each electrode channel according to the mean value and variance; Comparing the contact stability index of each electrode channel with a preset threshold value, and reserving conditioned electroencephalogram signals corresponding to the electrode channels with the contact stability index not lower than the preset threshold value as effective electroencephalogram signals after screening; meanwhile, a corresponding relation table of electrode channel identifiers and contact stability indexes is established, and the electrode channel identifiers and the corresponding contact stability indexes of each reserved electrode channel are recorded.
  4. 4. The multi-scenario fatigue state detection system based on electroencephalogram signals and a wearable cap according to claim 3, wherein the first phase lock point is located at a geometric center of a flexible dry electrode at a forehead median position in the wearable cap, the second phase lock point is located at a geometric center of a flexible dry electrode at a left temporal region position in the wearable cap, and the third phase lock point is located at a geometric center of a flexible dry electrode at a right temporal region position in the wearable cap.
  5. 5. The system for detecting the fatigue state of multiple scenes based on the electroencephalogram signals and the wearable cap according to claim 4, wherein the steps of constructing an elliptic space potential field curved surface and calculating an entropy value of an electrode distribution configuration include: When detecting that the contact stability index of the electrode channel corresponding to any one of the first phase lock point, the second phase lock point and the third phase lock point is lower than a preset dynamic calibration threshold, taking the physical position of the first phase lock point on the wearable cap body as a spatial reference origin, wherein the selection of the spatial reference origin only depends on the physical structure of the wearable cap body and does not depend on the contact stability index of the electrode channel corresponding to the corresponding phase lock point; taking the physical position of the first phase lock point on the wearable cap as an origin, taking the connecting line from the first phase lock point to the second phase lock point as a first direction vector, and taking the connecting line from the first phase lock point to the third phase lock point as a second direction vector, wherein the geometric center positions of the first phase lock point, the second phase lock point and the third phase lock point are fixed by the physical structure of the wearable cap; constructing an elliptic space potential field curved surface based on a plane formed by the first direction vector and the second direction vector; Taking a connecting line of the first phase lock point and the second phase lock point as a first equipotential segmentation datum line, taking a connecting line of the first phase lock point and the third phase lock point as a second equipotential segmentation datum line, and carrying out equipotential segmentation on the elliptic space potential field curved surface along the first equipotential segmentation datum line and the second equipotential segmentation datum line to form a plurality of solid angle unit cells; When the number of the flexible dry electrodes integrated in the wearable cap body reaches a preset spatial distribution analysis threshold, counting the number of the flexible dry electrodes which are contained in each solid angle unit cell and correspond to the electrode channels which are reserved after screening, and calculating the entropy of the electrode distribution configuration according to the variation coefficient of the contact stability index of each corresponding flexible dry electrode.
  6. 6. The multi-scene fatigue state detection system based on electroencephalogram signals and wearable caps according to claim 5, wherein constructing an elliptical spatial potential field curved surface based on a plane formed by the first direction vector and the second direction vector comprises: taking the starting point of the first direction vector as a reference end point, taking the end point of the first direction vector as a first reference end point, extracting a line segment expression of the first direction vector, and converting the line segment expression of the first direction vector into a parameterized linear equation to obtain a first parameterized linear equation; taking the starting point of the second direction vector as the reference end point, taking the end point of the second direction vector as a second reference end point, extracting a line segment expression of the second direction vector, and converting the line segment expression of the second direction vector into a parameterized linear equation to obtain a second parameterized linear equation; According to the first parameterized linear equation and the second parameterized linear equation, respectively calculating an inclination angle of the first direction vector relative to a preset reference coordinate system and an inclination angle of the second direction vector relative to the preset reference coordinate system to obtain a first inclination angle and a second inclination angle; Judging whether the first direction vector and the second direction vector are in a collinear state or not based on the difference value of the first inclination angle and the second inclination angle, when the difference value of the first inclination angle and the second inclination angle is in a preset collinear threshold range, taking the intersection point of the perpendicular bisector of the first direction vector and the perpendicular bisector of the second direction vector as the center point of the curved surface of the elliptical space potential field, and taking the average value of the lengths of the first direction vector and the second direction vector as a reference radius to construct the curved surface of the circular space potential field; When the difference between the first inclination angle and the second inclination angle exceeds a preset collineation threshold range, taking the end point of the first direction vector as a first projection point, calculating the shortest distance from the first projection point to a straight line corresponding to the second parameterized straight line equation to obtain a first projection distance; And determining a first boundary control point and a second boundary control point of the elliptic space potential field curved surface according to the first projection distance and the second projection distance, taking the first boundary control point and the second boundary control point as two focuses of an ellipse, taking the distance between the first boundary control point and the second boundary control point as an elliptic focal length, and taking the sum of the lengths of the first direction vector and the second direction vector as the length of a major axis of the ellipse to construct the elliptic space potential field curved surface.
  7. 7. The multi-scene fatigue state detection system based on electroencephalogram signals and wearable cap according to claim 6, wherein dynamically reconstructing the fusion weight of the screened effective electroencephalogram signals according to the entropy of the electrode distribution configuration to obtain the reconstructed fusion weight, and performing weighted fusion on the screened effective electroencephalogram signals to generate a comprehensive electroencephalogram signal sequence, and comprising: taking the entropy value of the electrode distribution configuration as a weight adjustment factor, carrying out nonlinear mapping transformation on the contact stability index of each electrode channel in a corresponding relation table of the electrode channel identification and the contact stability index to obtain an initial fusion weight of each electrode channel; normalizing the initial fusion weights of all the electrode channels in the same solid angle unit cell according to the number of the electrode channels contained in each solid angle unit cell to obtain the reconstructed fusion weights of all the electrode channels; And according to the reconstructed fusion weights, multiplying the screened effective electroencephalogram of each electrode channel with the reconstructed fusion weights of the corresponding electrode channels to obtain weighted electroencephalogram of each electrode channel, and accumulating and summing the weighted electroencephalogram of each electrode channel to obtain a comprehensive electroencephalogram sequence.
  8. 8. The system for detecting the fatigue state of the multiple scenes based on the electroencephalogram signals and the wearable cap body according to claim 7 is characterized in that the time-frequency and time-domain waveform analysis is carried out on the comprehensive electroencephalogram signal sequence, a first fatigue characterization index is calculated, and peripheral behavior fatigue symptom characteristics are determined, the duration of operation from the last rest to the current moment of a user is calculated according to the peripheral behavior fatigue symptom characteristics, and a multidimensional environment and operation time sequence characteristic set are obtained, and the system comprises: performing time-frequency analysis on the comprehensive electroencephalogram signal sequence, extracting the power spectral density of a delta frequency band, the power spectral density of a theta frequency band, the power spectral density of an alpha frequency band and the power spectral density of a beta frequency band in the comprehensive electroencephalogram signal sequence, and calculating a first fatigue characterization index according to the power spectral density of the delta frequency band, the power spectral density of the theta frequency band, the power spectral density of the alpha frequency band and the power spectral density of the beta frequency band; Performing time domain waveform analysis on the comprehensive electroencephalogram sequence, and identifying and separating a first type of ocular artifacts waveform induced by blink motion and a second type of ocular artifacts waveform induced by yawning motion from the comprehensive electroencephalogram sequence; Calculating the frequency of occurrence of the yawning characteristic waveform in unit time according to the second type of eye electric artifact waveform, and taking the frequency of occurrence of the yawning characteristic waveform in unit time, the single time of wink average duration and the frequency of occurrence of the yawning characteristic waveform in unit time as the peripheral behavior fatigue symptom characteristic.
  9. 9. The multi-scenario fatigue state detection system based on electroencephalogram signals and wearable caps according to claim 8, wherein the input of the first fatigue characterization index, the peripheral behavioral fatigue symptom feature, and the multi-dimensional environment and operation time sequence feature set to the pre-trained evaluation model, the obtaining of the comprehensive fatigue level determination result, and the triggering of the corresponding hierarchical alarm measures, comprises: The first fatigue characterization index, the peripheral behavior fatigue symptom feature, the multidimensional environment and the operation time sequence feature set are input into a pre-trained evaluation model together, and the interaction weight among the first fatigue characterization index, the peripheral behavior fatigue symptom feature, the multidimensional environment and the operation time sequence feature set is calculated dynamically through a cross attention mechanism in the evaluation model; Performing self-adaptive deep fusion on the first fatigue characterization index, the peripheral behavior fatigue symptom feature, the multi-dimensional environment and the operation time sequence feature set according to the interaction weight to obtain a comprehensive fatigue grade judgment result, wherein the comprehensive fatigue grade judgment result comprises a waking grade, a mild fatigue grade, a moderate fatigue grade and a severe fatigue grade; According to the comprehensive fatigue grade judging result, matching a grading alarm strategy corresponding to the comprehensive fatigue grade judging result from a preset alarm strategy library, wherein the grading alarm strategy comprises the steps of triggering a vibration motor integrated in a wearable cap body to generate a touch vibration alarm with a first preset frequency and a first preset duration when the comprehensive fatigue grade judging result is a mild fatigue grade, triggering a bone conduction earphone integrated in the wearable cap body to play a preset voice prompt when the comprehensive fatigue grade judging result is a moderate fatigue grade, and synchronously sending first-stage text alarm information to an associated mobile terminal, and triggering the touch vibration alarm and the bone conduction voice prompt when the comprehensive fatigue grade judging result is a severe fatigue grade.
  10. 10. A multi-scene fatigue state detection method based on electroencephalogram signals and a wearable cap body, the method realizing the system according to any one of claims 1 to 9, comprising: Step 1, preprocessing an original electroencephalogram signal of a scalp area of a user to obtain a conditioned electroencephalogram signal, analyzing the contact state of the conditioned electroencephalogram signal, and calculating a contact stability index corresponding to each flexible dry electrode to determine a screened effective electroencephalogram signal; Step 2, extracting a first phase lock point, a second phase lock point and a third phase lock point from an electrode channel corresponding to the screened effective electroencephalogram signal, constructing an elliptic space potential field curved surface, and calculating an electrode distribution configuration entropy value; Step 3, dynamically reconstructing the fusion weight of the screened effective electroencephalogram according to the entropy value of the electrode distribution configuration to obtain the reconstructed fusion weight, and carrying out weighted fusion on the screened effective electroencephalogram to generate a comprehensive electroencephalogram sequence; Step 4, performing time-frequency and time-domain waveform analysis on the comprehensive electroencephalogram signal sequence, calculating a first fatigue characterization index and determining peripheral behavior fatigue symptom characteristics; And 5, inputting the first fatigue characterization index, the peripheral behavior fatigue symptom feature, the multidimensional environment and the operation time sequence feature set into a pre-trained evaluation model to obtain a comprehensive fatigue grade judging result, triggering corresponding grading alarm measures, and transmitting the comprehensive fatigue grade judging result to an external monitoring platform or a local monitoring terminal.

Description

Multi-scene fatigue state detection system and method based on electroencephalogram signals and wearable cap body Technical Field The invention relates to the technical field of biological signal detection, in particular to a multi-scene fatigue state detection system and method based on an electroencephalogram signal and a wearable cap body. Background In the scenes of long-time driving, high-intensity operation or physical training, the fatigue state of a person is one of important factors causing efficiency reduction, misoperation and even safety accidents, and at present, partial technical schemes for detecting the fatigue state are proposed. Some existing fatigue detection methods rely on visual analysis of facial expressions (such as blinking and nodding) or operation behaviors (such as steering wheel holding) of people, however, such methods may be interfered by factors such as ambient light changes, shielding of wearing objects (such as glasses, masks and safety hats) and the like in some cases, a certain improvement space exists in detection stability, meanwhile, visual characteristics mostly show obvious changes when fatigue is obvious, early warning may have certain hysteresis, and a sufficient time is difficult to reserve for intervention. Other technologies try to perform fatigue evaluation by using an electroencephalogram signal, wherein the electroencephalogram signal can directly reflect the neural activity state of the brain, physiological information related to alertness, cognitive load and fatigue depth is stored in the electroencephalogram signal, the traditional electroencephalogram acquisition equipment is generally high in electrode number, relatively complicated in wearing process, and needs to be coated with conductive paste, and the appearance of the traditional electroencephalogram acquisition equipment is possibly not coordinated with a daily operation scene, the applicability of the traditional electroencephalogram acquisition equipment is still to be further verified in a scene needing to be continuously worn for a long time or having certain requirements on comfort and concealment, and at present, part of consumer-level electroencephalogram equipment is improved in portability, but still has exploration space in the aspect of deep integration with a carrier (such as a safety helmet and a working cap) needing to be worn for daily operation, and further optimization possibility exists in the aspects of comfort of long-time wearing, long-term stability of electrode contact, continuous usability under multiple scenes and the like. For example, in a long distance freight scenario, a driver needs to wear a safety helmet or a working helmet for a long time, if a traditional multi-electrode electroencephalogram acquisition device is adopted, the situation that the wearing process is relatively complex, the appearance is more prominent and the like may exist, the subjective acceptance of the driver needs to be further improved, and the detection effect of the fatigue monitoring system based on facial vision alone may be affected when the driver drives at night or wears a sunglasses and a mask. Disclosure of Invention The invention provides a multi-scene fatigue state detection system and method based on an electroencephalogram signal and a wearable cap body, and strong adaptability of wearable equipment in actual use. In order to solve the technical problems, the technical scheme of the invention is as follows: in a first aspect, a multi-scenario fatigue state detection system based on electroencephalogram signals and a wearable cap, comprising: the processing module is used for preprocessing the original electroencephalogram signal of the scalp area of the user to obtain a conditioned electroencephalogram signal, analyzing the contact state of the conditioned electroencephalogram signal, and calculating the contact stability index corresponding to each flexible dry electrode so as to determine the effective electroencephalogram signal after screening; the extraction module is used for extracting the first phase lock point, the second phase lock point and the third phase lock point from the electrode channel corresponding to the screened effective electroencephalogram signal, constructing an elliptic space potential field curved surface and calculating an electrode distribution configuration entropy value; The fusion module is used for dynamically reconstructing the fusion weight of the screened effective electroencephalogram signals according to the entropy value of the electrode distribution configuration to obtain the reconstructed fusion weight, and carrying out weighted fusion on the screened effective electroencephalogram signals to generate a comprehensive electroencephalogram signal sequence; The analysis module is used for carrying out time-frequency and time-domain waveform analysis on the comprehensive electroencephalogram signal sequence, calculating a first fatigue charact