CN-122004817-A - WiFi signal-based heart rate variability fatigue detection method and system

Abstract

The invention discloses a heart rate variability fatigue detection method and system based on WiFi signals, and relates to the technical field of wireless sensing; the method comprises the steps of controlling WiFi receiving and transmitting equipment to collect original CSI signals of an environment where a target individual is located, preprocessing the original CSI signals, dissociating independent source signals from the preprocessed CSI signals based on a blind source separation technology, calculating selected heart signal components through quantification, extracting heart beat interval sequences from the heart signal components, extracting heart rate variability features reflecting states of an autonomous nervous system, inputting the heart rate variability features into a pre-trained fatigue classification model, and outputting fatigue state judging results corresponding to the target individual. The invention utilizes the existing commercial WiFi facilities, and users do not need to wear equipment, so that the early, objective and noninductive high-precision identification of physiological fatigue is realized, and the invention also has certain popularization value while taking non-invasiveness, privacy protection and low cost into consideration.

Inventors

• ZHANG ZAILONG
• WU GE

Assignees

• 南京邮电大学

Dates

Publication Date

20260512

Application Date

20251210

Claims (10)

1. 1. The heart rate variability fatigue detection method based on the WiFi signals is characterized by comprising the following steps of: Step S1, controlling WiFi receiving and transmitting equipment to acquire an original CSI signal of an environment where a target individual is located, and preprocessing the original CSI signal; s2, dissociating an independent source signal from the preprocessed CSI signal based on a blind source separation technology, and calculating a selected heart signal component through quantization; Step S3, extracting a heart beat interval sequence from the heart signal component, and extracting heart rate variability characteristics reflecting the state of an autonomic nervous system based on the heart beat interval sequence; And S4, inputting the heart rate variability characteristics into a pre-trained fatigue classification model, and outputting a fatigue state judging result corresponding to the target individual.
2. 2. The WiFi signal-based heart rate variability fatigue detection method according to claim 1, wherein in step S1, preprocessing the raw CSI signal comprises: The original CSI signals are filtered by a band-pass filter to separate out physiological frequency bands where heartbeats and respiration are located, and low-frequency drift and high-frequency environmental noise caused by body posture change are filtered.
3. 3. The WiFi signal-based heart rate variability fatigue detection method according to claim 2, wherein step S1 further comprises: and carrying out standardization processing on each path of filtered CSI signals, and selecting a preset number of subcarriers with maximum variances by calculating the time sequence variances of all subcarrier signals to form a new CSI signal set.
4. 4. A method of detecting heart rate variability fatigue based on WiFi signals according to claim 3, wherein in step S2, the separation of the independent source signals from the preprocessed CSI signals based on blind source separation techniques comprises: Performing principal component analysis on the preprocessed CSI signals to realize data whitening of the signals, and projecting the signals onto the first m orthogonal principal components with the maximum variance to realize data dimensionality reduction and decorrelation of the signals; Where m is an integer set to preserve critical physiological signals and suppress noise.
5. 5. The WiFi signal-based heart rate variability fatigue detection method according to claim 4, wherein performing independent component analysis on the CSI signal after principal component analysis comprises: And (3) using a fixed point iterative algorithm to separate a statistically independent source signal from the CSI signal subjected to principal component analysis with the aim of maximizing the statistical independence among signal components.
6. 6. The WiFi signal-based heart rate variability fatigue detection method according to claim 5, wherein performing a quantization calculation on each independent source signal comprises: And identifying and selecting the independent source signal with the largest frequency spectrum power or highest power ratio in the specific physiological frequency range as the heart signal component.
7. 7. The method for WiFi signal-based heart rate variability fatigue detection according to claim 6, wherein in step S3, the extracting of the heart beat interval sequence comprises: performing a secondary band pass filtering on the selected cardiac signal component, the passband range of the secondary band pass filtering being narrower than the preprocessing filtering range in step S1, set to 0.8Hz to 2.0Hz; and positioning the most obvious characteristic point in each heartbeat pulse on the heart signal after secondary filtering based on a peak detection algorithm, and calculating a heartbeat interval sequence according to the characteristic point.
8. 8. The WiFi signal based heart rate variability fatigue detection method according to claim 7, wherein heart rate variability characteristics of at least one standard time domain or frequency domain are calculated within a specific time window based on the heart beat interval sequence.
9. 9. The WiFi signal based heart rate variability fatigue detection method according to claim 5, wherein the fixed point iterative algorithm is a fastca algorithm; The objective function that maximizes non-gaussian is a kurtosis-based contrast function or a negative entropy contrast function.
10. 10. A heart rate variability fatigue detection system based on WiFi signals, comprising: The signal acquisition and preprocessing module is configured to control the WiFi receiving and transmitting equipment to acquire an original CSI signal of an environment where a target individual is located and preprocess the original CSI signal; A signal dissociation module configured to dissociate an independent source signal from the preprocessed CSI signal based on a blind source separation technique and calculate a selected cardiac signal component by quantization; a feature extraction module configured to extract a cardiac interval sequence from the cardiac signal component and to extract heart rate variability features reflecting the state of the autonomic nervous system based on the cardiac interval sequence; And the fatigue state decision module is configured to input the heart rate variability characteristics into a pre-trained fatigue classification model and output a fatigue state decision result corresponding to a target individual.

Description

WiFi signal-based heart rate variability fatigue detection method and system Technical Field The invention relates to a heart rate variability fatigue detection method and system based on WiFi signals, and belongs to the technical field of wireless sensing. Background Long-time and high-strength activities can lead to organism fatigue, and especially under the scenes of fatigue driving, high-strength mental work and the like, continuous physiological fatigue is one of main reasons for reducing productivity and frequently occurring safety accidents. Therefore, the fatigue state of the personnel is detected timely, accurately and without interference, and the method has important practical significance for guaranteeing the life and property safety and improving the working efficiency. At present, schemes based on wearable devices, such as Electrocardiogram (ECG) or photoplethysmogram (PPG) devices, can provide high-precision physiological indexes, but have high device cost, are required to be worn by users, have obvious invasiveness, and severely limit the application of the wearable devices in daily and unprepared scenes. Based on the scheme of computer vision, the fatigue is judged by capturing facial features (such as blink frequency and yawning) through the camera, but the risk of invading personal privacy exists though the camera is non-contact, and the camera is easily interfered by environmental factors such as illumination conditions, facial shielding and the like, so that the camera is not robust. Schemes based on wireless signal sensing, such as detection with WiFi Channel State Information (CSI), are of great interest because of their low cost, ubiquitous nature. However, the prior related technology still has the technical core of identifying external macroscopic behaviors, such as "yawning", "rubbing eyes", "nodding" and the like as the fatigue criteria. The method has the fundamental defects that firstly, the macroscopic actions are usually hysteresis performance of deep fatigue and cannot realize early warning, and secondly, the external performance of fatigue is different from person to person and is easy to subjectively inhibit or disguise (such as strong tolerance) by a user, so that the monitoring is unreliable. In addition, there are some methods that judge by quantifying the overall activity level of the user, but the granularity is too coarse to distinguish between "high concentration rest" and "tired drowsiness rest", resulting in ambiguity and uncertainty of judgment. Although heart rate variability is an important indicator of human fatigue. However, in wireless sensing, the intensity of the signal generated by the chest cavity fluctuation caused by respiratory motion is much greater than the weak pulse signal caused by the heartbeat. This results in CSI data acquired at the receiving end being a highly mixed signal, wherein weak heartbeat signals are completely submerged by strong respiration signals. The existing WiFi sensing method lacks an effective signal decoupling means, and is difficult to separate an accurate heartbeat signal sequence from the strong interference background. Disclosure of Invention The invention aims to provide a heart rate variability fatigue detection method and system based on WiFi signals, which are used for extracting pure heart signals from CSI through a blind source separation technology, further calculating Heart Rate Variability (HRV) characteristics and realizing early accurate and objective judgment on physiological fatigue. In order to achieve the above purpose, the invention is realized by adopting the following technical scheme. In one aspect, the invention provides a heart rate variability fatigue detection method based on WiFi signals, which comprises the following steps: Step S1, controlling WiFi receiving and transmitting equipment to acquire an original CSI signal of an environment where a target individual is located, and preprocessing the original CSI signal; s2, dissociating an independent source signal from the preprocessed CSI signal based on a blind source separation technology, and calculating a selected heart signal component through quantization; Step S3, extracting a heart beat interval sequence from the heart signal component, and extracting heart rate variability characteristics reflecting the state of an autonomic nervous system based on the heart beat interval sequence; And S4, inputting the heart rate variability characteristics into a pre-trained fatigue classification model, and outputting a fatigue state judging result corresponding to the target individual. Optionally, in step S1, preprocessing the original CSI signal includes: The original CSI signals are filtered by a band-pass filter to separate out physiological frequency bands where heartbeats and respiration are located, and low-frequency drift and high-frequency environmental noise caused by body posture change are filtered. Optionally, step S1 further includes: an