Abnormal state detection method and device based on physiological signals, equipment and medium

Abstract

The invention provides an abnormal state detection method and device based on physiological signals, electronic equipment and a storage medium. The method comprises the steps that multiple physiological signals of a target object are acquired; determining a physiological feature vector of the current physiological state of the target object based on the multiple physiological signals; determining whether the state of the target object is in an abnormal state or not according to the physiological feature vector; and when the state of the target object is in an abnormal state, triggering early warning prompt information. According to the method, the multiple physiological signals of the target object are comprehensively analyzed, and compared with a single physiological signal index, the actual state of the target object can be reflected more accurately; compared with misjudgment possibly caused by a single data source or redundant equipment such as a sensor needed by analyzing physiological signals such as a complex electrocardiogram, the burden of the target object is relieved, and the detection accuracy and reliability are improved.

Description

Technical Field

[0001] The present invention relates to the field of safe driving technology, and in particular to an abnormal state detection method, device, electronic device and computer-readable storage medium based on physiological signals. Background Art

[0002] With the rapid progress of automobile electrification and intelligence, the automation and ease of operation of small vehicles have been greatly improved. Automatic driving and assisted driving have greatly reduced the difficulty of vehicle operation for drivers. However, the number of traffic accidents caused by drunk driving, fatigue driving, distracted driving, etc. on the road has not decreased significantly. Therefore, the normality of the driver's personal state during the vehicle process is crucial to the overall safety of the vehicle during driving.

[0003] In the related technologies, the abnormal state detection of the driver usually uses physiological or behavioral monitoring of the driver, for example, detection based on driving behavior related information (such as the vehicle's position in the lane, speed, steering wheel usage, braking and shifting, etc.), or computer vision detection based on driver body analysis (such as eye state information and head movement, yawning and facial expressions, etc.), or detection based on biomedical multi-signals (such as electroencephalogram, electrocardiogram, skin galvanic response and breathing, etc.) etc. However, these detection methods are relatively simple and have great limitations, resulting in low detection accuracy.

[0004] Therefore, how to accurately detect the abnormal state of the driver while the vehicle is driving is a problem that needs to be solved. Summary of the invention

[0005] The present invention provides a method, device, electronic device and computer-readable storage medium for detecting abnormal state based on physiological signals, so as to at least solve the problem in the related art that the accuracy of detecting abnormal state of the driver is reduced due to the single detection method of the abnormal state of the driver. The technical solution of the present invention is as follows:

[0006] According to a first aspect of an embodiment of the present invention, a method for detecting an abnormal state based on physiological signals is provided, comprising: acquiring a plurality of physiological signals of a target object; determining a physiological feature vector of a current physiological state of the target object based on the plurality of physiological signals; determining whether the state of the target object is in an abnormal state according to the physiological feature vector; and triggering an early warning prompt message when the state of the target object is in an abnormal state.

[0007] According to a second aspect of an embodiment of the present invention, there is provided an abnormal state detection device based on physiological signals, comprising: an acquisition module for acquiring a plurality of physiological signals of a target object; a first determination module for determining a physiological characteristic vector of a current physiological state of the target object based on the plurality of physiological signals; a second determination module for determining whether the state of the target object is in an abnormal state according to the physiological characteristic vector; and a trigger module for triggering an early warning prompt message when the state of the target object is in an abnormal state.

[0008] According to a third aspect of an embodiment of the present invention, there is provided an abnormal state detection system based on physiological signals, the system comprising: a physiological signal acquisition module, for acquiring a plurality of physiological signals of a target object; a signal filtering and processing module, for performing signal preprocessing on the plurality of acquired physiological signals; a signal analysis and processing module, for determining a physiological characteristic vector of a current physiological state of the target object based on the plurality of preprocessed physiological signals; an abnormal state detection module, for determining whether the state of the target object is in an abnormal state according to the physiological characteristic vector; and an abnormal detection alarm triggering module, for triggering a warning prompt message when the abnormal state detection module detects that the state of the target object is in an abnormal state.

[0009] According to a fourth aspect of an embodiment of the present invention, there is provided an electronic device, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement the abnormal state detection method based on physiological signals as described above.

[0010] According to a fifth aspect of an embodiment of the present invention, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the abnormal state detection method based on physiological signals as described above.

[0011] According to a sixth aspect of an embodiment of the present invention, there is provided a computer program product, comprising a computer program or instructions, which, when executed by a processor of an electronic device, implements the abnormal state detection method based on physiological signals as described above.

[0012] The technical solution provided by the embodiments of the present invention brings at least the following beneficial effects:

[0013] In an embodiment of the present invention, multiple physiological signals of the target object are obtained; based on the multiple physiological signals, the physiological characteristic vector of the current physiological state of the target object is determined; based on the physiological characteristic vector, it is determined whether the state of the target object is in an abnormal state; when the state of the target object is in an abnormal state, an early warning prompt information is triggered. That is to say, in an embodiment of the present invention, by comprehensively analyzing the multiple physiological signals of the target object, the actual state of the target object can be more accurately reflected than a single physiological signal indicator, that is, a more comprehensive perspective is provided to monitor the state of the target object; compared with the misjudgment that may be caused by a single data source or the redundant equipment such as sensors required by analyzing complex physiological signals such as electrocardiograms, the burden on the target object is reduced, thereby improving the accuracy and reliability of detection.

[0014] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present invention, and are used together with the specification to explain the principles of the present invention, and do not constitute an improper limitation of the present invention. In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art description will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 This is a flow chart of an abnormal state detection method based on physiological signals provided by an embodiment of the present invention.

[0017] Figure 2 It is a schematic diagram of sliding window movement provided by an embodiment of the present invention.

[0018] Figure 3 It is a schematic diagram of a smoothing calculation based on Gaussian filtering provided by an embodiment of the present invention.

[0019] Figure 4 It is a block diagram of an abnormal state detection device based on physiological signals provided by an embodiment of the present invention.

[0020] Figure 5 This is another block diagram of an abnormal state detection device based on physiological signals provided by an embodiment of the present invention.

[0021] Figure 6 It is a block diagram of a preprocessing module provided by an embodiment of the present invention.

[0022] Figure 7 It is a block diagram of an abnormal state detection system based on physiological signals provided by an embodiment of the present invention.

[0023] Figure 8 It is a block diagram of an electronic device provided by an embodiment of the present invention.

[0024] Fig. 9 The invention is a block diagram of a device for abnormal state detection based on physiological signals provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to enable ordinary persons in the art to better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings.

[0026] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0027] Technical terms:

[0028] Support Vector Machine (SVM), a widely used machine learning algorithm, is mainly used for classification and regression problems.

[0029] Random Forest (RF), a machine learning algorithm based on ensemble learning, obtains the final prediction result by constructing multiple decision trees and combining them.

[0030] Fast Fourier Transform (FFT) is an algorithm for quickly calculating discrete Fourier transform. Its main function is to efficiently convert a time domain signal to the frequency domain and analyze the spectral characteristics of the signal more quickly.

[0031] Short-time Fourier transform (STFT) is a commonly used signal processing technology used to analyze the time-frequency characteristics of non-stationary signals, that is, to obtain the local stationary signal of the non-stationary signal, and then use short-time Fourier transform to perform frequency domain analysis.

[0032] Principal Component Analysis (PCA) is a commonly used dimensionality reduction and feature extraction technique.

[0033] Respiration Rate Variability (RRV) is an important physiological indicator for evaluating and detecting respiratory abnormalities by calculating the standard deviation of the periodic variation of the respiratory signal.

[0034] Arterial oxygen saturation (SpO2, Oxygen Saturation) is the percentage of oxygen content in the blood to the total amount of oxygenated hemoglobin in the blood, that is, the degree of oxygen saturation in the blood.

[0035] Heart rate variability (HRV) refers to the changes in the intervals between consecutive heartbeats, which reflects the ability of the autonomic nervous system to regulate the heart.

[0036] The commonly used indicator of heart rate variability (SDNN, Standard Deviation of NN Intervals) is used to measure the variability of the time interval (NN interval) between two adjacent heartbeats.

[0037] The commonly used indicator of heart rate variability (RMSSD, Root Mean Square of Successive Differences) represents the root mean square of the difference between two adjacent heartbeats and is used to evaluate the changes in heart rate in the short term.

[0038] On the basis of understanding the above technical terms, please refer to the following embodiments.

[0039] Figure 1 is a flow chart of an abnormal state detection method based on physiological signals provided by an embodiment of the present invention, such as Figure 1 As shown, the method comprises the following steps:

[0040] Step 101: Acquire multiple physiological signals of a target object.

[0041] Step 102: Determine a physiological feature vector of the current physiological state of the target object based on the multiple physiological signals.

[0042] Step 103: Determine whether the state of the target object is in an abnormal state according to the physiological feature vector.

[0043] Step 104: When the state of the target object is in an abnormal state, trigger an early warning prompt message.

[0044] The abnormal state detection method based on physiological signals described in the present invention can be applied to terminals, servers, etc., without limitation herein. The terminal implementation device can be electronic devices such as smart phones, laptops, tablet computers, desktop computers, personal digital assistants (PDAs) and wearable devices. The server can be an independent server or a server cluster, or a server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, intermediate services, domain name services, security services, content distribution networks, or a big data and artificial intelligence platform, etc., without limitation herein.

[0045] The abnormal state detection method based on physiological signals described in the present invention can also be applied to the vehicle side, cloud side, etc., without limitation here. The vehicle-side implementation equipment can be a vehicle-mounted terminal, a vehicle control platform, an industrial computer and other electronic equipment. The cloud side can be an independent server or a server cluster, or a server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, intermediate services, domain name services, security services, content distribution networks, or a big data and artificial intelligence platform, etc., without limitation here.

[0046] Combine the following Figure 1 , the specific implementation steps of an abnormal state detection method based on physiological signals provided by an embodiment of the present invention are described in detail. For the convenience of description, the target object in the embodiment of the present invention is described by taking a driver as an example, but it is not limited to this in specific applications.

[0047] In step 101, a plurality of physiological signals of a target object are acquired.

[0048] The target object in this embodiment can be any user of abnormality detection, such as a driver, a passenger, a worker based on industrial or safety detection, an athlete optimizing training programs and enhancing athletic performance, or a person undergoing medical testing based on basic health care, etc., and this embodiment does not impose any restrictions.

[0049] Among them, the multiple physiological signals may include but are not limited to: heart rate, respiratory rate, blood oxygen saturation and other key physiological signals. Of course, for different abnormal detection users, the multiple physiological signals obtained are not exactly the same. For example, the multiple physiological information of the driver is obtained. In addition to including: heart rate, respiratory rate, blood oxygen saturation and other key physiological signals, it can also include: driver driving behavior (such as the vehicle's position in the lane, speed, steering wheel usage, braking and shifting, etc.), facial expressions (such as eye closure percentage, eye closure duration and eye closure frequency, or yawning, etc.), which are not limited in this embodiment.

[0050] Optionally, in this step, a variety of physiological signals of the target object can be obtained through a wearable terminal and computer vision. In this embodiment, the wearable terminal is such as a smart watch, which integrates multiple physiological sensing functions such as heart rate monitoring, blood oxygen saturation detection, respiration monitoring, heart rate detection, etc., and can collect various physiological signals required by the driver in one stop. Compared with existing scattered sensors, smart watches are more compact and integrated. Through various physiological sensors on smart watches, such as heart rate sensors, blood pressure sensors, respiration sensors, etc., a variety of physiological signals of the driver can be collected in real time, where physiological signals can also be called physiological signal data.

[0051] In this embodiment, the smart watch is a portable device worn by the driver, which can continuously and imperceptibly collect the driver's

[0052] The required physiological signals. Compared with the various sensors, sensor belts and other complex equipment worn, smart watches can better reduce the burden on drivers.

[0053] In step 102, a physiological feature vector of the current physiological state of the target object is determined based on the multiple physiological signals.

[0054] This step includes: 1) performing signal analysis and processing on the multiple physiological signals to extract the characteristics of the current physiological state of the target object. For example, extracting the breathing, blood oxygen saturation, heart rate and other characteristics of the driver's current physiological state. Specifically including:

[0055] 11) A threshold truncation algorithm is used to smooth the multiple physiological signals, and instantaneous fluctuation points in the multiple physiological signals are removed to obtain smoothed physiological signals.

[0056] In this step, the multiple physiological signals are smoothed by using a threshold truncation method (or threshold truncation method or threshold truncation processing, etc.), which can reduce false alarms of instantaneous fluctuations of relevant physiological signals caused by the driver's body activities. The threshold truncation method simply truncates and removes abnormal fluctuation points that exceed a certain threshold. Before this, the multiple physiological signals obtained can also be smoothed and high-frequency noise suppressed by a sliding window smoothing filter algorithm. For details, see below. Therefore, the threshold truncation method can effectively reduce the impact of instantaneous fluctuations.

[0057] 12) According to the set physiological signal cycle, determining the physiological signal cycle change amount of the smoothed physiological signal; specifically including:

[0058] According to the set physiological signal cycle, the physiological signal in each physiological signal cycle after smoothing is extracted; the difference between adjacent physiological signal cycles of each physiological signal is calculated as the corresponding physiological signal cycle variation.

[0059] That is to say, in this step, a certain amount of time is selected from the filtered physiological signal as a physiological signal cycle, and based on each physiological signal cycle, physiological signals such as respiration, blood oxygen saturation, heart rate, etc. in each cycle are extracted respectively. Then, the difference between adjacent physiological signal cycles of each physiological signal is calculated, and the difference is used as the change in the physiological signal cycle.

[0060] 13) Calculating the standard deviation of the physiological signal periodic variation, and extracting the physiological index variation value corresponding to the current physiological state of the target object.

[0061] In this step, the periodic variation of the physiological signal is calculated. This embodiment takes the respiratory signal, the blood oxygen saturation (SpO2) signal and the heart rate signal as examples, but in practical applications, it is not limited to these.

[0062] For respiratory signals: the standard deviation of the periodic variation of these smoothed respiratory signals is calculated to obtain the RRV value. The standard deviation calculation formula is:

[0063] RRV=std(RR_interval[i]-RR_interval[i-1])

[0064] The RRV is expressed as respiratory rhythm variability. The RRV index value reflects the degree of fluctuation of the respiratory cycle in a short period of time. It can be used to detect changes in the breathing pattern when the driver tries to stay awake. The symptoms of fatigue driving can be identified through this index. RR interval represents the breathing interval. std represents the standard deviation calculation.

[0065] For the arterial oxygen saturation (SpO2) signal: the periodic variation of the smoothed arterial oxygen saturation signals is calculated by standard deviation to obtain the SpO2 Variability (SpO2V) value. The calculation formula is:

[0066] SpO2V=std(SpO2max-SpO2min)

[0067] Among them, SpO2V is arterial blood oxygen saturation, and std standard deviation is calculated. The larger the SpO2V, the greater the fluctuation of arterial blood oxygen saturation during breathing. It can be used to detect abnormal changes in blood oxygen saturation after drinking, so as to identify the state of drunk driving.

[0068] For heart rate signals: the standard deviation of the periodic variation of these smoothed heart rate signals is calculated to obtain the heart rate variability (HRV) value. The calculation formula is:

[0069] HRV=std(RR[i]-RR[i-1])

[0070] Among them, HRV is heart rate variability, which refers to the change of normal continuous heartbeat intervals (RR intervals), which reflects the ability of the autonomic nervous system to regulate the heart. In this embodiment, two time domain indicators in HRV are used for comprehensive detection to ensure more accurate state identification. That is, the commonly used indicators of heart rate variability (SDNN, Standard Deviation of NNIntervals) and the commonly used indicators of heart rate variability (RMSSD, Root Mean Square of Successive Differences).

[0071] Among them, SDNN represents the overall HRV level, and the standard deviation of the RR interval (i.e., normal heartbeat interval) is calculated according to the following formula:

[0072]

[0073] Among them, SDNN represents the commonly used index of heart rate variability, RR i represents the interval between consecutive heart beats, It represents the average value of the heart beat interval. That is, the time domain characteristics in the heart rate variability (HRV) analysis can be obtained through the above formula. The higher the calculated SDNN, the stronger the HRV, and the better the autonomic nervous system regulation function.

[0074] RMSSD represents the short-term HRV level by calculating the average of the square roots of the differences between adjacent RR intervals. The higher the RMSSD, the stronger the parasympathetic nerve activity and the greater the heart rate change.

[0075] Therefore, the embodiment of the present invention can use SDNN and RMSSD to determine the change of HRV in the drunken state. When the time domain indicators such as SDNN and / or RMSSD decrease, it indicates that the overall HRV and parasympathetic nerve activity are reduced.

[0076] 2) All extracted features are integrated to obtain the physiological feature vector of the target object.

[0077] This embodiment is to splice the extracted features of different physiological signals into a feature vector. Specifically, it can be completed by simple feature connection, or splicing using complex methods such as principal component analysis (PCA) to reduce the dimension while maintaining key information.

[0078] In step 103, it is determined whether the state of the target object is in an abnormal state according to the physiological feature vector.

[0079] In this step, various physiological signals of the target object in a normal state are obtained; based on the various physiological signals in the normal state, the fluctuation range of each physiological index of the target object in a normal state is determined as the corresponding physiological index benchmark; the physiological feature vector is compared with the corresponding physiological index benchmark; if the comparison result is greater than a set threshold, the state of the target object is determined to be an abnormal state; otherwise, if the comparison result is less than the set threshold, the state of the target object is determined to be a normal state.

[0080] That is to say, since the driver is in a normal state in most cases when driving a vehicle, in the embodiment of the present invention, a smart watch can be used to continuously obtain various physiological signals of the driver in a normal state, and by extracting the characteristics of the indicators of the physiological signals, the fluctuation range of each physiological indicator in the normal state is obtained, and the fluctuation range of these physiological indicators is used as a physiological indicator benchmark for comparison.

[0081] Therefore, after determining the physiological indicator baseline, to judge whether the driver is in an abnormal state, the physiological feature vector of the physiological signal extracted within the period can be compared with the physiological indicator baseline of the physiological signal under normal state through threshold judgment or joint judgment, and the comparison results of multiple physiological signals can be analyzed. When the comparison result is greater than the set threshold, it can be determined that the driver's state is abnormal. Usually, this embodiment can determine whether the driver is in an abnormal state in a more comprehensive and objective way.

[0082] Among them, the threshold judgment is that when the comparison result (ie, the output probability) exceeds a certain threshold (such as 70%, etc.), the driver's state is determined to be an abnormal state, otherwise, the driver's state is determined to be a normal state.

[0083] The combined judgment is to consider the comprehensive status of all physiological signals according to the comparison results. For example, if the heart rate is too fast and the blood oxygen saturation is reduced, the driver's status is determined to be abnormal. Otherwise, the driver's status is determined to be normal.

[0084] The following is an example to illustrate:

[0085] If the driver's breathing physiological signals are more rapid than normal, the driver may have a respiratory disease such as asthma; if the driver's breathing physiological signals are slower than normal, it may be that drinking alcohol has reduced the sensitivity of the respiratory center; or it may be that the driver is in a drowsy state, which reduces the sensitivity of the respiratory center and causes a slower breathing rate.

[0086] As for blood oxygen saturation, when the driver is in a state of fatigue or sleepiness, the oxygen content in the blood will decrease and the blood oxygen saturation will decrease; when the driver drinks alcohol, the alcohol will inhibit the sensitivity of the respiratory center, causing a decrease in breathing frequency and depth, which will reduce the oxygen content in the blood and cause the blood oxygen saturation to decrease.

[0087] As for the heart rate signal, alcohol will stimulate the sympathetic nerves and increase the heart rate. Therefore, the heart rate signal of the driver after drinking alcohol will be significantly accelerated. Alcohol will inhibit the regulatory function of the autonomic nervous system. SDNN will decrease by 10-30%, indicating that the overall heart rate variability is weakened, and RMSSD will decrease by 20-50%, reflecting that the vagus nerve activity is inhibited; when the driver is in a fatigued state, the heart rate will usually decrease, which may decrease by 5-10 times / minute, and the heart rate will have a larger fluctuation range, reflecting that the regulatory function of the autonomic nervous system is reduced. SDNN will usually decrease by 10-20%, and RMSSD will decrease significantly, usually by 20-40. It should be noted that in actual applications, the percentage of decrease in this embodiment is not limited to this, and it can also be appropriately enlarged or reduced on this basis, etc., and this embodiment is not limited.

[0088] Therefore, by comparing multiple physiological signals, this embodiment can more comprehensively and objectively assess whether the driver is in an abnormal state and prevent potential safety hazards. Once abnormal symptoms such as rapid or slow breathing, increased heart rate, and decreased blood oxygen saturation appear, it means that the driver is in an abnormal state and urgently needs to rest or stop driving to reduce safety hazards.

[0089] Optionally, for the physiological feature vector of the normal state, the present embodiment can also use the support vector machine SVM and the random forest RF to determine whether the state of the target object is an abnormal state. The process of SVM or RF for judgment is already a well-known technology for those skilled in the art and will not be repeated here.

[0090] It should be noted that the detection of abnormal conditions may involve statistical analysis or machine model learning methods to identify patterns or anomalies. If logistic regression is used for classification, it can be implemented according to the following formula:

[0091]

[0092] Among them, x n is a feature, β n are model parameters and y is the class label.

[0093] In step 104, when the state of the target object is in an abnormal state, an early warning prompt information is triggered.

[0094] In this step, when the state of the target object is in an abnormal state, the smart cockpit audio is controlled to issue a voice reminder to prompt that the target object is currently in an abnormal state; or when the state of the target object is in an abnormal state, the instrument panel is controlled to light up a conspicuous warning light to prompt that the target object is currently in an abnormal state; or when the state of the target object is in an abnormal state, the display screen is controlled to display safety monitoring warning information to prompt that the target object is currently in an abnormal state.

[0095] In this embodiment, when it is confirmed that the driver's physiological state is in an abnormal state, a warning prompt message is triggered, specifically including:

[0096] The driver can be reminded through the smart cockpit audio system. Specifically, the vehicle audio system can send a voice reminder to the driver, such as "It is detected that you are currently in a state of fatigue. Please find a safe area to stop and rest as soon as possible." Further, the audio system can also send reminders to other passengers to let them pay attention to the driver's status and take safety precautions.

[0097] The driver can also be warned by the instrument lights: a conspicuous warning light will light up on the driving instrument panel to remind the driver that he is in an abnormal state or fatigue state and needs to take a rest or stop to rest. The color and flashing frequency of the warning light can be differentiated according to the degree of abnormality to attract the driver's attention.

[0098] Of course, in addition to being applicable to the cockpit driver monitoring field, the embodiments of the present invention can also be applied to the following fields:

[0099] For example, basic medical testing (health care, not diagnostics), other industrial / safety testing, and status monitoring of workers in sensitive environments; in sports science, optimizing training programs and enhancing athletic performance by analyzing athletes' physiological signals (such as heart rate and muscle activity).

[0100] For these fields, when the state of the target object is in an abnormal state, the display screen can be controlled to display safety monitoring warning information or emit an alarm sound to prompt that the target object is currently in an abnormal state.

[0101] In an embodiment of the present invention, multiple physiological signals of the target object are obtained; based on the multiple physiological signals, the physiological characteristic vector of the current physiological state of the target object is determined; based on the physiological characteristic vector, it is determined whether the state of the target object is in an abnormal state; when the state of the target object is in an abnormal state, an early warning prompt information is triggered. That is to say, in an embodiment of the present invention, by comprehensively analyzing the multiple physiological signals of the target object, the actual state of the target object can be more accurately reflected than a single physiological signal indicator, that is, a more comprehensive perspective is provided to monitor the state of the target object; compared with the misjudgment that may be caused by a single data source or the redundant equipment such as sensors required by analyzing complex physiological signals such as electrocardiograms, the burden on the target object is reduced, thereby improving the accuracy and reliability of detection.

[0102] Optionally, in another embodiment, based on the above embodiment, the method may further include: performing signal preprocessing on the acquired multiple physiological signals by using a sliding window smoothing filter combined with frequency domain analysis.

[0103] Among them, the sliding window smoothing filter method (algorithm) is a signal preprocessing technology, which is usually used to remove noise interference in various physiological signals. This method is simple to implement and has high computational efficiency; it can effectively remove high-frequency noise and improve the signal-to-noise ratio; and the length of the sliding window can be flexibly selected to meet different needs. The embodiment of the present invention uses a sliding window smoothing filter algorithm combined with a short-time Fourier transform to perform frequency domain analysis and filter the various physiological signals obtained from the target object, which can effectively reduce false alarms caused by changes in related physiological signals caused by physical activities, thereby obtaining a variety of high-quality physiological signals.

[0104] Among them, the signal preprocessing of the multiple physiological signals obtained by using a sliding window smoothing filter combined with a frequency domain analysis specifically includes: performing smoothing calculations on the multiple physiological signals obtained according to a set sliding window to obtain waveform graphs of the multiple physiological signals; converting the waveform graphs of the multiple physiological signals into corresponding amplitude matrices; assigning different weights to the multiple physiological signals to obtain weight matrices of the multiple physiological signals; performing convolution operations on the amplitude matrix and the weight matrix to obtain an amplitude matrix after noise is filtered out; converting the amplitude matrix after noise is filtered out into a corresponding waveform graph; performing frequency domain analysis on the waveform graph using a short-time Fourier transform, and performing signal filtering processing on the multiple physiological signals to remove noise interference in the multiple physiological signals.

[0105] The sliding window smoothing filter combined with frequency domain analysis provided in this embodiment can remove noise interference in various physiological signals by setting a sliding window, smoothing calculation, and frequency domain analysis processing through short-time Fourier transform. Specifically, it includes:

[0106] 1) Sliding Window

[0107] This embodiment sets a sliding window, that is, defines a time window of a fixed length (such as 2 seconds, 5 seconds or 10 seconds, etc.). The sliding window will continue to slide forward over time, moving one sampling point or one time unit each time. Figure 2 As shown, Figure 2 A schematic diagram of a sliding window movement provided by an embodiment of the present invention. For each time point, the smoothing calculation uses multiple sampling points in the sliding window to perform average calculation to obtain a smoothing value at the time point.

[0108] 2) Smooth calculation

[0109] The smoothing calculation method provided in this embodiment can adopt a Gaussian filtering calculation method (pre-assigning different weights to multiple physiological signals such as breathing, heart rate, blood pressure, etc. in the sliding window), as follows: Figure 3 As shown, Figure 3 A schematic diagram of a smoothing calculation based on Gaussian filtering provided by an embodiment of the present invention. The Gaussian filtering calculation method was first applied in images, and different weights were assigned to pixel values ​​at different positions of the image. The closer the distance, the greater the weight, and the farther the distance, the smaller the weight. The Gaussian filtering algorithm is a process of weighted averaging of the entire image. The value of each pixel point is obtained by weighted averaging itself and other pixel values ​​in the neighborhood (with different weights). Based on this, this embodiment applies this idea to the scenario of this embodiment. For the various physiological signals collected by this embodiment, the waveform diagrams of the various physiological signals obtained are converted into corresponding matrix waveform diagrams by the method of amplitude matrix waveform diagrams, and the amplitude values ​​in the matrix waveform diagrams are stored in the matrix, so that the amplitude matrix corresponding to the various physiological signals can be obtained. For the weight matrix, different weights are assigned to different physiological signals in advance, so as to distinguish the importance of each different physiological signal for abnormal state detection in the system. In this embodiment, the amplitude matrix of the physiological signal is convolved with the weight matrix according to the sliding window to obtain the amplitude matrix after noise is filtered out, and then the amplitude matrix after noise is filtered out is converted into a corresponding waveform diagram.

[0110] 3) Short-time Fourier transform (STFT)

[0111] In the embodiment of the present invention, by short-time Fourier transform, the signal of the converted waveform is divided into shorter time segments, and each time segment is subjected to frequency domain analysis processing of short-time Fourier transform, so as to retain certain time information. This method can provide the frequency component changes of the signal at different time points, and is suitable for analyzing time-varying signals. Dual information of time and frequency (that is, time domain and frequency domain) can be provided. And it is relatively easy to understand and implement. However, there is an inherent trade-off between the resolution of time and frequency (that is, time domain and frequency domain) (the sliding window size is inversely proportional to the frequency resolution). It may not be sensitive enough for rapidly changing frequency components. Sliding window and smoothing calculation, as well as how to use STFT specifically, can be appropriately selected according to the hardware situation and application conditions. The present embodiment realizes time-frequency (that is, time domain and frequency domain) analysis by applying a series of localized short-time Fourier transforms to the signal. Mathematically, STFT can be expressed by the following formula:

[0112]

[0113] Where x(t) is the signal in the time domain; w(t-τ) is the creation function, which localizes the signal at time τ, and f is the frequency; e -j2πft It is the core of Fourier transform.

[0114] In the embodiments of the present invention, a method is provided for improving signal quality by implementing time domain and frequency domain analysis through a sliding window smoothing filtering algorithm, thereby reducing false alarms caused by external environmental interference, thereby improving detection accuracy and reliability.

[0115] In the embodiment of the present invention, a sliding window smoothing filter technique is first used to remove physiological signals with certain noise, and then a threshold truncation method is used to reduce false positives of instantaneous fluctuations of relevant physiological signals caused by body activities, thereby improving the accuracy and reliability of detection.

[0116] In the embodiment of the present invention, multiple signal sources are integrated: it is not limited to analyzing a single physiological or behavioral signal, but integrates driving behavior data, computer vision analysis, and biomedical signal processing to provide a more comprehensive perspective to monitor the driver's status. This comprehensive method helps to reduce the misjudgment that may be caused by a single data source and improves the accuracy and reliability of detection.

[0117] Real-time data processing and feedback mechanism: Provides a clear real-time data processing and feedback mechanism, which is critical in improving driving safety and immediate response. Through real-time analysis and timely feedback, the system can quickly identify potential abnormal behavior and take action, which is especially important when the driver is tired or distracted.

[0118] Advanced signal processing technology: Using advanced signal processing technologies such as STFT (Short-time Fourier Transform), time series data can be analyzed more accurately, overcoming the limitations of traditional FFT in processing non-stationary signals. This is especially important for the changing physiological and behavioral signals in complex driving environments.

[0119] Intelligent learning and model adaptability: By using machine learning techniques such as SVM (support vector machine) and RF (random forest), it can not only process complex data sets, but also learn and adapt to new behavior patterns from continuous data streams. This learning and adaptability enables the system to remain efficient under changing road and driving conditions.

[0120] In addition, this embodiment can also reduce the interference of external environmental factors by integrating computer vision and biomedical signals, and can continue to work effectively when the driver wears sunglasses, etc. by analyzing multiple physiological signals. The interference to the driver can be reduced by using non-invasive or less invasive monitoring technology. And compared with the driver abnormality judgment system based on multimodal large models, if the system design allows running on resource-limited vehicle hardware, or adopts more effective model compression and optimization technology, it may be more efficient in terms of computing resources and energy consumption.

[0121] It should be noted that, for the method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that this implementation disclosure is not limited by the order of the actions described, because according to the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the present invention.

[0122] Figure 4 4 is a block diagram of an abnormal state detection device based on physiological signals provided by an embodiment of the present invention. The device includes: an acquisition module 401, a first determination module 402, a second determination module 403 and a trigger module 404, wherein:

[0123] An acquisition module 401 is used to acquire multiple physiological signals of a target object;

[0124] A first determination module 402, configured to determine a physiological feature vector of a current physiological state of the target object based on the multiple physiological signals;

[0125] A second determination module 403 is used to determine whether the state of the target object is in an abnormal state according to the physiological feature vector;

[0126] The trigger module 404 is used to trigger an early warning prompt message when the state of the target object is in an abnormal state.

[0127] Optionally, in another embodiment, based on the above embodiment, the device may further include: a preprocessing module 501, whose structural block diagram is as follows: Figure 5 As shown,

[0128] The preprocessing module 501 is used to perform signal preprocessing on the multiple physiological signals acquired by the acquisition module 401 by using a sliding window smoothing filter combined with frequency domain analysis.

[0129] Optionally, in another embodiment, based on the above embodiment, the preprocessing module 501 includes: a smoothing calculation module 601, a first conversion module 602, a distribution module 603, a convolution operation module 604, a second conversion module 605 and a filtering module 606, and its structural block diagram is as follows: Figure 6 As shown,

[0130] A smoothing calculation module 601 is used to perform smoothing calculation on the acquired multiple physiological signals according to a set sliding window to obtain waveform diagrams of the multiple physiological signals;

[0131] A first conversion module 602, configured to convert the waveform diagrams of the plurality of physiological signals into corresponding amplitude matrices;

[0132] An allocation module 603, configured to allocate different weights to the plurality of physiological signals to obtain a weight matrix of the plurality of physiological signals;

[0133] A convolution operation module 604 is used to perform a convolution operation on the amplitude matrix and the weight matrix to obtain an amplitude matrix after noise is filtered out;

[0134] A second conversion module 605, used to convert the amplitude matrix after noise is filtered out into a corresponding waveform diagram;

[0135] The filtering module 606 is used to perform frequency domain analysis on the waveform using short-term Fourier transform, and perform signal filtering processing on the multiple physiological signals to remove noise interference in the multiple physiological signals.

[0136] Optionally, in another embodiment, based on the above embodiment, the acquisition module is specifically used to acquire multiple physiological signals of the target object through a wearable terminal and computer vision.

[0137] Optionally, in another embodiment, based on the above embodiment, the first determining module includes:

[0138] An analysis and processing module, used to perform signal analysis and processing on the multiple physiological signals to extract characteristics of the current physiological state of the target object;

[0139] The fusion module is used to fuse all the extracted features to obtain the physiological feature vector of the target object.

[0140] Optionally, in another embodiment, based on the above embodiment, the analysis and processing module includes:

[0141] A smoothing processing module, used for smoothing the multiple physiological signals by using a threshold truncation method, removing instantaneous fluctuation points in the multiple physiological signals, and obtaining smoothed physiological signals;

[0142] A variation determination module, used to determine the physiological signal cycle variation of the smoothed physiological signal according to the set physiological signal cycle;

[0143] The change value calculation module is used to calculate the standard deviation of the periodic change of the physiological signal and extract the physiological indicator change value corresponding to the current physiological state of the target object.

[0144] Optionally, in another embodiment, based on the above embodiment, the variation determination module includes:

[0145] An extraction module, used for extracting the physiological signal in each physiological signal cycle after smoothing according to the set physiological signal cycle;

[0146] The difference calculation module is used to calculate the difference between adjacent physiological signal cycles of each physiological signal as the corresponding physiological signal cycle variation.

[0147] Optionally, in another embodiment, based on the above embodiment, the second determining module includes:

[0148] A signal acquisition module, used to acquire various physiological signals of the target object in a normal state;

[0149] A reference determination module, used to determine the fluctuation range of each physiological index of the target object in the normal state based on various physiological signals in the normal state as a corresponding physiological index reference;

[0150] A comparison module, used for comparing the physiological characteristic vector with a corresponding physiological index benchmark;

[0151] The abnormality determination module is used to determine that the state of the target object is an abnormal state when the comparison result is greater than a set threshold.

[0152] Optionally, in another embodiment, based on the above embodiment, the device further includes:

[0153] The judgment module is used to determine that the state of the target object is a normal state when the comparison result of the first comparison module is less than a set threshold value, and use support vector machine SVM and random forest RF for the physiological feature vector of the normal state to determine whether the state of the target object is an abnormal state.

[0154] Optionally, in another embodiment, based on the above embodiment, the trigger module includes at least one of the following:

[0155] A first control module, configured to control the smart cockpit audio system to issue a voice reminder to prompt that the target object is currently in an abnormal state when the second determination module determines that the state of the target object is in an abnormal state;

[0156] A second control module, configured to control a striking warning light on the instrument panel to light up when the second determination module determines that the state of the target object is in an abnormal state, so as to prompt that the target object is currently in an abnormal state;

[0157] The third control module is used to control the display screen to display safety monitoring warning information or emit an alarm sound to prompt that the target object is currently in an abnormal state when the second determination module determines that the state of the target object is in an abnormal state.

[0158] Optionally, an embodiment of the present invention further provides an electronic device, including:

[0159] processor;

[0160] a memory for storing instructions executable by the processor;

[0161] The processor is configured to execute the instructions to implement the abnormal state detection method based on physiological signals as described above.

[0162] Optionally, an embodiment of the present invention further provides a computer-readable storage medium, when instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the abnormal state detection method based on physiological signals as described above.

[0163] Optionally, an embodiment of the present invention further provides a computer program product, including a computer program or instructions, which, when executed by a processor of an electronic device, implements the abnormal state detection method based on physiological signals as described above.

[0164] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0165] See also Figure 7A schematic diagram of an abnormal state detection system based on physiological signals provided by an embodiment of the present invention is shown in FIG. Figure 7 As shown, it includes: a physiological signal acquisition module 701, a signal filtering and processing module 702, a signal analysis and processing module 703, an abnormality detection and determination module 704 and an abnormality detection alarm triggering module 705, wherein,

[0166] The physiological signal acquisition module 701 is used to obtain various physiological signals of the target object; specifically, various physiological signals of the target object can be obtained through a wearable terminal (or wearable instrument) and computer vision. Among them, the various physiological signals may include: breathing, blood oxygen and heart rate and other physiological signals. Of course, it also includes: eye state information, head movement, yawning and facial expressions, etc., which are not limited in this embodiment.

[0167] The signal filtering processing module 702 is used to perform signal preprocessing on the acquired multiple physiological signals by using a sliding window smoothing filter combined with frequency domain analysis; wherein the preprocessing process uses a sliding window smoothing filter combined with frequency domain analysis, which may specifically include: setting a sliding window, smoothing calculation (such as Gaussian filtering, etc.) and frequency domain analysis of short-time Fourier transform (STFT), etc. For details, please refer to the corresponding implementation process in the above embodiment, which will not be repeated here.

[0168] The signal analysis and processing module 703 is used to determine the physiological characteristic vector of the current physiological state of the target object based on the multiple physiological signals after preprocessing, specifically including: smoothing processing (reducing the impact of instantaneous fluctuations), physiological signal period rate estimation, physiological signal feature extraction, analysis and calculation, etc. The implementation process is detailed in the corresponding implementation process in the above embodiment and will not be repeated here.

[0169] The abnormality detection determination module 704 is used to determine whether the state of the target object is in an abnormal state according to the physiological feature vector; the abnormality detection determination module can also be called an abnormal mode decision model, including: abnormality detection decision processing, usually using threshold judgment or joint judgment to perform abnormality detection. For threshold judgment, if the output probability of the model exceeds a certain threshold (such as 70%), the state is considered abnormal; otherwise, the state is considered normal. For joint judgment: consider the comprehensive state of all physiological signals, such as too fast heart rate and reduced blood oxygen saturation at the same time, the state is considered abnormal, otherwise it is considered normal.

[0170] Furthermore, for normal status, statistical analysis or machine learning methods can be used to identify patterns to determine abnormal status. Specifically, SVM or RF can be used for abnormal monitoring, and the specific process is not described here.

[0171] The abnormal detection alarm triggering module 705 is used to trigger early warning prompt information when the state of the target object is in an abnormal state.

[0172] In the system, the specific implementation process of each module is detailed in the corresponding implementation process in the above embodiment, which will not be repeated here.

[0173] When the state of the target object is in an abnormal state, a warning prompt message is triggered. Specifically, when the state of the target object is in an abnormal state, the intelligent cockpit audio system can be controlled to issue a voice reminder, or a conspicuous warning light can be controlled to light up on the instrument panel, or the display screen can be controlled to display safety monitoring warning information or issue an alarm sound to prompt that the target object is currently in an abnormal state.

[0174] In an embodiment of the present invention, a monitoring system for abnormal states of target objects (such as drivers, etc.) based on physiological signals is proposed, which is designed for real-time monitoring and analysis of key physiological signals such as the driver's heart rate, respiratory rate, and blood oxygen saturation. By integrating these physiological signals, abnormal states that may affect driving safety can be identified and warned in real time, thereby enhancing road safety and preventing traffic accidents. The system continuously collects the driver's physiological signals through wearable terminals and sensor devices installed in the cockpit. The collected raw data undergoes preprocessing steps such as noise filtering and signal normalization to ensure the accuracy and reliability of subsequent analysis. The techniques used in the preprocessing process, such as smoothing, convolution calculation, and short-time Fourier transform (STFT), effectively remove signal interference from complex driving environments.

[0175] Feature extraction is performed on multiple physiological signals after preprocessing, and key features reflecting the driver's current physiological state can be extracted from each physiological signal. These features include but are not limited to heart rate variability, breathing regularity, and blood oxygen stability. These features are vector fused and integrated through algorithms such as principal component analysis (PCA) to form a comprehensive description of the physiological state. The fused feature vector is input into the abnormal state determination module, which includes: a threshold judgment module or a joint judgment module, and is also equipped with advanced machine learning algorithms such as support vector machine (SVM) or random forest (RF) classification module. These SVM or RF algorithms learn to identify patterns of normal and abnormal physiological states based on historical training data. When a potential abnormal state is detected, the system will automatically trigger an alarm to notify the driver to take timely measures or notify the traffic management center to intervene.

[0176] Among them, support vector machine (SVM) provides a powerful classification method that can be used to process segmented raster input or standard image. Random forest (RF) refers to a classifier that uses multiple trees to train and predict samples.

[0177] In the embodiment of the present invention, the detection system provided takes into account high accuracy and low interference, and is particularly suitable for driving environments. The real-time response capability of the system can issue early warnings in time when the driver is fatigued or in an abnormal physiological state, significantly improving driving safety. In addition, the flexibility and scalability of the system make it adaptable to various vehicles and driving conditions. At the same time, the algorithm is based on individual characteristics and avoids the complex calculations of the current common LLM model, which has a very high cost performance.

[0178] The embodiment of the present invention proposes a novel abnormal state detection system for a target object (such as a driver) based on physiological signals. The system first uses a smart watch to obtain important physiological signals such as the driver's breathing, blood sample saturation, and heart rate; in order to improve the reliability of detection, an innovative method of sliding window smoothing filtering combined with frequency domain analysis is proposed to filter the acquired relevant physiological signals to reduce false alarms caused by changes in relevant physiological signals caused by physical activities. After these two steps of filtering, high-quality physiological signal data is obtained. Subsequently, these filtered physiological signals are analyzed and processed in real time, and compared with the fluctuation state under normal conditions, so as to detect whether there are signs of abnormal state of the driver. Therefore, the method provided by the embodiment of the present invention can more accurately evaluate the driver's physiological state and timely warn of possible safety hazards.

[0179] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0180] Figure 8 800 is a block diagram of an electronic device 800 provided by an embodiment of the present invention. As shown in the figure, it includes a processor 801, a communication interface 802, a memory 803 and a communication bus 804, wherein the processor 801, the communication interface 802, and the memory 803 communicate with each other through the communication bus 804;

[0181] A memory 803, used to store the processor executable instructions;

[0182] The processor 801 is used to implement the above method when executing the executable instructions in the memory 803.

[0183] The communication bus in this embodiment may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0184] The communication interface is used for communication between the above electronic device and other devices.

[0185] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0186] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0187] In another embodiment of the present invention, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method described above.

[0188] In another embodiment of the present invention, a computer program product including a computer program or instructions is provided. When the computer program or instructions are executed by a processor of an electronic device, the abnormal state detection method based on physiological signals as described above is implemented.

[0189] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented by software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk Solid State Disk (SSD)), etc.

[0190] In an embodiment, a computer-readable storage medium is also provided, and when the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device can perform the abnormal state detection method based on physiological signals shown above. For example, the computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0191] In an embodiment, a computer program product is also provided, including a computer program or instructions. When the computer program or instructions are executed by a processor of an electronic device, the electronic device executes the abnormal state detection method based on physiological signals described above.

[0192] Fig. 9 is a block diagram of an apparatus 900 for abnormal state detection based on physiological signals provided by an embodiment of the present invention. For example, the apparatus 900 may be provided as a server. Fig. 9 , the apparatus 900 includes a processing component 922, which further includes one or more processors, and a memory resource represented by a memory 932 for storing instructions, such as an application, that can be executed by the processing component 922. The application stored in the memory 932 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 922 is configured to execute the instructions to perform the above method.

[0193] The device 900 may also include a power supply component 926 configured to perform power management of the device 900, a wired or wireless network interface 950 configured to connect the device 900 to a network, and an input / output (I / O) interface 958. The device 900 may operate based on an operating system stored in the memory 932, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, or the like.

[0194] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed by the present invention. The specification and examples are to be considered exemplary only, and the true scope and spirit of the present invention are indicated by the following claims.

[0195] It should be understood that the present invention is not limited to the exact construction that has been described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A method for detecting abnormal state based on physiological signals, characterized in that: include: Acquire multiple physiological signals of the target object; Based on the multiple physiological signals, determining a physiological feature vector of the current physiological state of the target object; Determining whether the state of the target object is in an abnormal state according to the physiological feature vector; When the state of the target object is in an abnormal state, an early warning prompt message is triggered.

2. The abnormal state detection method based on physiological signals according to claim 1, characterized in that: The method further comprises: The acquired multiple physiological signals are preprocessed by using a sliding window smoothing filter combined with frequency domain analysis.

3. The abnormal state detection method based on physiological signals according to claim 2 is characterized in that: The method of performing signal preprocessing on the acquired multiple physiological signals by using a sliding window smoothing filter combined with a frequency domain analysis comprises: Performing smoothing calculation on the acquired multiple physiological signals according to a set sliding window to obtain waveform diagrams of the multiple physiological signals; Converting the waveform diagrams of the plurality of physiological signals into corresponding amplitude matrices; assigning different weights to the multiple physiological signals to obtain a weight matrix of the multiple physiological signals; Performing a convolution operation on the amplitude matrix and the weight matrix to obtain an amplitude matrix after noise is filtered out; Converting the amplitude matrix after noise filtering into a corresponding waveform diagram; The waveform is subjected to frequency domain analysis using short-time Fourier transform, and the multiple physiological signals are subjected to signal filtering processing to remove noise interference in the multiple physiological signals.

4. The abnormal state detection method based on physiological signals according to any one of claims 1 to 3, characterized in that: The step of determining the physiological feature vector of the current physiological state of the target object based on the multiple physiological signals includes: Performing signal analysis and processing on the multiple physiological signals to extract characteristics of the current physiological state of the target object; All extracted features are fused to obtain a physiological feature vector of the target object.

5. The abnormal state detection method based on physiological signals according to claim 4 is characterized in that: The performing signal analysis and processing on the multiple physiological signals to extract the characteristics of the current physiological state of the target object includes: Using a threshold truncation method to smooth the multiple physiological signals, removing instantaneous fluctuation points in the multiple physiological signals, and obtaining smoothed physiological signals; According to the set physiological signal cycle, determining the physiological signal cycle change amount of the smoothed physiological signal; The standard deviation of the physiological signal periodic variation is calculated to extract the physiological index variation value corresponding to the current physiological state of the target object.

6. The abnormal state detection method based on physiological signals according to claim 1 or 2, characterized in that: The determining whether the state of the target object is in an abnormal state according to the physiological feature vector includes: Acquiring various physiological signals of the target object in a normal state; Based on the various physiological signals in the normal state, determining the fluctuation range of each physiological index of the target object in the normal state as a corresponding physiological index benchmark; Comparing the physiological characteristic vector with a corresponding physiological index benchmark; When the comparison result is greater than a set threshold, it is determined that the state of the target object is an abnormal state.

7. The abnormal state detection method based on physiological signals according to claim 1 or 2, characterized in that: When the state of the target object is in an abnormal state, triggering early warning prompt information includes: When the state of the target object is in an abnormal state, controlling the smart cockpit audio system to issue a voice reminder to prompt that the target object is currently in an abnormal state; or When the state of the target object is in an abnormal state, a conspicuous warning light on the control panel is turned on to prompt that the target object is currently in an abnormal state; or When the state of the target object is in an abnormal state, the control display screen displays safety monitoring warning information or emits an alarm sound to prompt that the target object is currently in an abnormal state.

8. An abnormal state detection device based on physiological signals, characterized in that: include: An acquisition module, used to acquire various physiological signals of a target object; A first determination module, configured to determine a physiological feature vector of a current physiological state of the target object based on the multiple physiological signals; A second determination module, configured to determine whether the state of the target object is in an abnormal state according to the physiological feature vector; The trigger module is used to trigger early warning information when the state of the target object is in an abnormal state.

9. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the abnormal state detection method based on physiological signals as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the abnormal state detection method based on physiological signals as described in any one of claims 1 to 7.

Images