Automatic driving health and safety emergency early warning control method based on vital sign monitoring

Abstract

The invention belongs to the technical field of vehicle control, and particularly relates to an automatic driving health and safety emergency early warning control method based on vital sign monitoring, which comprises the following steps: acquiring vital sign index data; driving state data and ambient environment data; calculating a vital sign index change rate and a deviation change rate; predicting the change trend of the vital sign indexes; comparing the vital sign index change rate with the predicted change trend, and outputting abnormal signal intensity according to the current vital sign index deviation change rate; comprehensive abnormal signal intensity is calculated in a weighted summation mode; based on the comprehensive abnormal signal intensity, the health risk level of the driver is judged, multiple groups of virtual emergency strategy scenes are generated in combination with the health risk level and the driving state data and the surrounding environment data of the vehicle, and an optimal emergency strategy is screened out and triggered to be executed. According to the system, tiny physiological changes of vital signs can be caught acutely, abnormal signals can be recognized and evaluated accurately, early warning can be conducted in time, and emergency measures can be taken.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle control, and particularly relates to an autonomous driving health and safety emergency warning control method based on vital sign monitoring. Background Art

[0002] With the continuous development of autonomous driving technology, the intelligence level of vehicles is getting higher and higher, but the health status of drivers is still an important factor affecting driving safety. During driving, a driver may suddenly get sick, such as having a heart attack or sudden syncope. These situations may cause the driver to lose the ability to control the vehicle, thus leading to serious traffic accidents. Traditional safety systems (such as seat belt pre-tensioning, airbag triggering) can only respond passively after a collision, cannot avoid accidents, and have limited real-time monitoring and emergency handling capabilities for the driver's vital signs.

[0003] Existing ADAS (Advanced Driving Assistance System) relies on the driver's operation behaviors (such as steering wheel angle, accelerator pedal force) to indirectly infer the health status. However, in the initial stage of health abnormalities, there may only be minor physiological changes (such as sudden increase in heart rate, rapid breathing), which are easily misjudged as fatigue driving or operation errors. It cannot actively intervene in driving behaviors, resulting in a significant increase in accident risks. Summary of the Invention

[0004] In order to solve the problem that when the driver's life is abnormal during driving, in the initial stage of health abnormalities, there may only be minor physiological changes, and the driving system can give a timely warning, the present invention provides the following technical solutions:

[0005] An autonomous driving health and safety emergency warning control method based on vital sign monitoring, comprising the following steps:

[0006] S1. Real-time collect the vital sign index data of the driver through distributed on-vehicle biosensors; meanwhile, obtain the driving state data and surrounding environment data of the vehicle;

[0007] S2. According to the collected vital sign index data, calculate the change rate of each vital sign index per unit time; and calculate the median deviation change rate of the vital sign index data relative to the normal range of the preset vital sign index;

[0008] S3. Through the LSTM network, train to obtain a vital sign change trend prediction model, and predict the vital sign index change trend within a future time window based on the vital sign change trend prediction model; compare the vital sign index change rate with the predicted change trend, and at the same time, according to the degree of deviation of the current vital sign index deviation change rate from the normal range, identify the abnormal signals that may have potential disease risks in the vital signs, and output the abnormal signal intensity;

[0009] S4. Assign corresponding dynamic weight coefficients to different types of abnormal signal strength at the same time point, and calculate the comprehensive abnormal signal strength by weighted summation;

[0010] S5. Based on the comprehensive abnormal signal strength, the driver's health risk level is determined. Combined with the health risk level and the vehicle's driving status data and surrounding environment data, multiple sets of virtual emergency strategy scenarios are generated. The optimal emergency strategy is selected from the multiple sets of virtual emergency strategy scenarios and triggered for execution.

[0011] Furthermore, in step S1, the distributed vehicle-mounted biosensor includes at least one of a steering wheel pressure-sensitive respiratory sensor, a seat-embedded photoplethysmography sensor, and an intelligent rearview mirror infrared thermal imager.

[0012] Furthermore, in step S1, the driving status data includes vehicle speed, acceleration, steering angle, braking status, gear information, driving direction and driving duration.

[0013] Furthermore, in step S1, the surrounding environment data includes road type, traffic flow, traffic signs, traffic light status, obstacle distance and lane line status.

[0014] Furthermore, in step S3, the process of constructing the vital signs change trend prediction model includes: arranging the driver's historical vital signs indicator data, as well as the change rate and deviation change rate of their vital signs indicator data in chronological order to form time series data; extracting features from the time series data; inputting the extracted features into the LSTM network for training, and adjusting the network parameters so that the model can accurately predict the change trend of vital signs indicators in the future time window.

[0015] Furthermore, in step S4, the entropy weight-hierarchy analysis method is used in combination with the correlation analysis of historical abnormal signal data to assign corresponding dynamic weight coefficients to the abnormal signal strengths of different types of vital signs.

[0016] Furthermore, the entropy weight-hierarchy analysis method specifically includes: using the entropy weight method to calculate the objective weight of the change rate of each vital sign indicator, and assigning different weights to each abnormal signal strength in different time periods; and using the hierarchical analysis method, combined with the correlation analysis of historical data and the deviation change rate of vital sign indicator data, to determine the subjective weight of each abnormal signal strength; linearly combining the objective weight and the subjective weight to obtain the final dynamic weight coefficient of each abnormal signal strength.

[0017] Furthermore, in step S5, each set of virtual emergency strategy scenarios is simulated and deduced to evaluate the risk reduction degree and driver safety protection effect in each scenario, thereby screening out the optimal emergency strategy from multiple sets of virtual emergency strategy scenarios.

[0018] Further, in step S5, the emergency strategy includes vehicle deceleration and stop, adjustment of in-vehicle environmental parameters, sending distress messages and vehicle position information to emergency contacts or the emergency center through the vehicle-mounted communication system.

[0019] Further, when 0 < comprehensive abnormal signal intensity ≤ 1, the health risk level is mildly abnormal; when 1 < comprehensive abnormal signal intensity ≤ 3, the health risk level is moderately abnormal; when comprehensive abnormal signal intensity > 3, the health risk level is severely abnormal.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] By collecting various vital sign index data through distributed vehicle-mounted biosensors, not only the driver's vital sign data is collected, but also the vehicle's driving state data and surrounding environment data are obtained at the same time; by the change rate of vital sign indexes and the deviation change rate relative to the median value of the preset normal range, comparing the actual change rate with the predicted trend, and considering the degree of deviation from the normal range at the same time to identify abnormal signals, it can sensitively capture the minute physiological changes of vital signs, accurately identify and evaluate abnormal signals, and can give early warnings and take emergency measures.

[0022] The entropy weight - analytic hierarchy process is used to assign dynamic weight coefficients to different types of abnormal signal intensities at the same time point, and the comprehensive abnormal signal intensity is calculated by weighted summation, which fully considers the relative importance of different vital sign indexes and the correlation between them, and more accurately evaluates the overall health risk status of the driver. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flowchart of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] The technical solutions of the present invention will be clearly described below in conjunction with the drawings. Obviously, the described embodiments are not all embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0025] As Figure 1 shown, the present invention provides an autonomous driving health and safety emergency warning and control method based on vital sign monitoring, including the following steps:

[0026] S1. Real-time collect the vital sign index data of the driver through distributed vehicle-mounted biosensors; at the same time, obtain the vehicle's driving state data and surrounding environment data;

[0027] The vital sign index data includes cardiovascular indexes, respiratory indexes and galvanic skin response indexes;

[0028] The distributed vehicle-mounted biosensor includes at least one of a seat-embedded photoplethysmogram sensor, a steering-wheel integrated piezoresistive respiration sensor, a seatbelt integrated galvanic skin response sensor, an intelligent rearview mirror infrared thermal imager, and a vehicle-mounted respiratory gas analyzer.

[0029] The photoplethysmogram sensor is used to collect the heart rate and blood oxygen saturation of the driver; the piezoresistive respiration sensor detects the respiration rate and indirectly measures the respiration rate by detecting the change in the hand pressure of the driver; the galvanic skin response sensor measures the skin conductivity; the intelligent rearview mirror infrared thermal imager is used to monitor the facial temperature distribution of the driver in real time.

[0030] The driving state data includes: 1) basic driving parameters: vehicle speed, acceleration, steering angle, braking state, gear information, driving direction, and driving duration; 2) power system parameters: engine speed, torque output curve, and transmission oil temperature; and 3) chassis system parameters: tire ground contact pressure distribution and suspension system damping coefficient change rate.

[0031] The surrounding environment data includes: 1) road environment data: road type, slope, curvature, lane line state, and road surface friction coefficient; 2) traffic environment data: traffic flow, average vehicle speed, traffic signs, and traffic signal states; 3) obstacle data: obstacle type, distance, moving speed, and direction; and 4) meteorological data: weather conditions, light intensity, visibility, and road surface dryness and wetness states.

[0032] S2. Calculate the change rate of each vital sign index per unit time based on the collected vital sign index data; and calculate the deviation change rate of the vital sign index data relative to the median of the normal range of the preset vital sign index.

[0033] S3. Through the LSTM network, train to obtain a vital sign change trend prediction model, and predict the change trend of the vital sign index within the future time window based on the vital sign change trend prediction model; compare the vital sign index change rate with the predicted change trend, and at the same time, according to the degree to which the current vital sign index deviation change rate deviates from the normal range, identify the abnormal signal that may have a potential disease risk in the vital signs, and output the abnormal signal intensity.

[0034] The construction process of the vital sign change trend prediction model includes: arranging the historical vital sign index data of the driver, as well as the change rate and deviation change rate of its vital sign index data in chronological order to form time series data; extracting features from the time series data, such as extracting the periodic features of heart rate and the fluctuation features of blood pressure, etc.; inputting the extracted features into the LSTM network for training and adjusting the network parameters to enable the model to accurately predict the change trend of vital sign indexes within a future time window.

[0035] Specifically, the method for identifying abnormal signals: for example, when there is a large deviation between the predicted heart rate change trend and the actual change rate, and the deviation change rate exceeds its preset threshold, it is determined as an abnormal signal, that is, the vital sign index appears abnormal.

[0036] For example, if it is predicted that the heart rate will change steadily within the next 5 minutes, but the actual heart rate change rate suddenly increases, and at the same time the deviation change rate of the current heart rate exceeds the upper limit of the preset normal range (such as deviating from the normal range by 5%), the system determines that a heart rate abnormal signal appears.

[0037] The abnormal signal intensity is equal to the standardized value of the abnormal signal multiplied by its deviation change rate. The larger the deviation change rate, the higher the abnormal signal intensity.

[0038] S4. Assign corresponding dynamic weight coefficients to the abnormal signal intensities of different types at the same time point, and calculate the comprehensive abnormal signal intensity by means of weighted summation;

[0039] Calculate the abnormal signal intensities of indicators such as heart rate, respiratory rate, and blood oxygen saturation respectively. The entropy weight - analytic hierarchy process is used to assign corresponding dynamic weight coefficients to the abnormal signal intensities of different types in combination with the correlation analysis of historical abnormal signal data.

[0040] The entropy weight - analytic hierarchy process specifically includes: using the entropy weight method to calculate the objective weights of each abnormal signal intensity, and assigning different weights to the abnormal signal intensities at different time periods; and using the analytic hierarchy process, combining the correlation analysis of historical data and the deviation change rate of vital sign index data, to determine the subjective weights of each abnormal signal intensity; linearly combining the objective weights and subjective weights to obtain the final dynamic weight coefficient of each abnormal signal intensity.

[0041] According to the final dynamic weight coefficients, for multiple abnormal signal intensities (such as heart rate, respiratory rate, blood oxygen saturation) at the same time point, calculate the comprehensive abnormal signal intensity through weighted fusion.

[0042] S5. Based on the comprehensive abnormal signal strength, the driver's health risk level is determined. Combined with the health risk level and the vehicle's driving status data and surrounding environment data, multiple sets of virtual emergency strategy scenarios are generated. The optimal emergency strategy is selected from the multiple sets of virtual emergency strategy scenarios and triggered for execution.

[0043] By simulating and deducing each set of virtual emergency strategy scenarios, the risk reduction level and driver safety protection effect under each scenario are evaluated, so as to screen out the optimal emergency strategy from multiple sets of virtual emergency strategy scenarios.

[0044] Emergency strategy scenarios include slowing down and stopping the vehicle, adjusting the in-vehicle environmental parameters, turning on the hazard lights, sending a distress message and vehicle location information to the emergency center through the in-vehicle communication system, and contacting emergency contacts.

[0045] When 0<comprehensive abnormal signal strength≤1, the health risk level is mild abnormality; adjust the seat posture to a comfortable position and increase lumbar support; adjust the temperature in the car to 24±2℃ and the humidity to 50±10%; turn on the air purification system to increase the oxygen concentration; voice reminder: "It is recommended that you rest in the service area ahead."

[0046] When 1<comprehensive abnormal signal strength≤3, the health risk level is moderately abnormal; the car automatically decelerates and seeks a safe parking location; activates the hazard warning lights and alerts surrounding vehicles through the on-board display and voice; evaluates destinations such as service areas, hospitals, and parking lots, and selects the shortest route with the estimated arrival time calculated using the Dijkstra algorithm based on real-time road conditions.

[0047] When the comprehensive abnormal signal strength is >3, the health risk level is severe; a distress message containing the driver's real-time vital signs data is sent to the preset emergency contact through the in-vehicle communication system; the emergency center is automatically dialed and the vehicle's precise location and driving trajectory are simultaneously sent; the vehicle positioning tracking function is turned on and the location information is continuously updated; the door is automatically unlocked, the window is opened, and the seat is adjusted to a lying position.

Claims

1. An autonomous driving health and safety emergency warning control method based on vital sign monitoring, characterized in that The steps are as follows: S1. Real-time collect the vital sign index data of the driver through a distributed in-vehicle biosensor; meanwhile, obtain the vehicle driving state data and the surrounding environment data; S2. According to the collected vital sign index data, calculate the change rate of each vital sign index per unit time; and calculate the median deviation change rate of the vital sign index data relative to the normal range of the preset vital sign index; S3. Through the LSTM network, train to obtain a vital sign change trend prediction model, and predict the vital sign index change trend within a future time window based on the vital sign change trend prediction model; compare the vital sign index change rate with the predicted change trend, and at the same time, according to the degree to which the current vital sign index deviation change rate deviates from the normal range, identify the abnormal signals that may have potential disease risks in the vital signs, and output the abnormal signal intensity; S4. Assign corresponding dynamic weight coefficients to the abnormal signal intensities of different types at the same time point, and calculate the comprehensive abnormal signal intensity by means of weighted summation; S5. Based on the comprehensive abnormal signal intensity, judge the health risk level of the driver, combine the health risk level and the vehicle driving state data and the surrounding environment data, generate multiple groups of virtual emergency strategy scenarios, screen out the optimal emergency strategy from the multiple groups of virtual emergency strategy scenarios, and trigger the execution.

2. The method according to claim 1, wherein In step S1, the distributed in-vehicle biosensor includes at least one of a steering wheel pressure-sensitive respiration sensor, a seat-embedded photoplethysmogram sensor, and an intelligent rearview mirror infrared thermal imager.

3. The method according to claim 1, wherein In step S1, the driving state data includes vehicle speed, acceleration, steering angle, braking state, gear information, driving direction, and driving duration.

4. The method according to claim 1, characterized in that, In step S1, the surrounding environment data includes road type, traffic flow, traffic signs, traffic signal light state, obstacle distance, and lane line state.

5. The method according to claim 1, characterized in that In step S3, the construction process of the vital sign change trend prediction model includes: arranging the historical vital sign index data of the driver, as well as the change rate and deviation change rate of its vital sign index data in chronological order to form time series data; extracting features from the time series data; inputting the extracted features into the LSTM network for training, and adjusting the network parameters so that the model can accurately predict the vital sign index change trend within a future time window.

6. The method according to claim 1, wherein In step S4, the entropy weight - analytic hierarchy process is adopted, and in combination with the correlation analysis of historical abnormal signal data, corresponding dynamic weight coefficients are assigned to the abnormal signal intensities of different types of vital signs.

7. The method according to claim 6, wherein The entropy weight - analytic hierarchy process specifically includes: calculating the objective weights of the change rates of each vital sign index by using the entropy weight method, and assigning different weights to the abnormal signal intensities at different time periods; and using the analytic hierarchy process, in combination with the correlation analysis of historical data and the deviation change rate of the vital sign index data, to determine the subjective weights of each abnormal signal intensity; linearly combining the objective weights and the subjective weights to obtain the final dynamic weight coefficient of each abnormal signal intensity.

8. The method according to claim 7, characterized in that, In step S5, each set of virtual emergency strategy scenarios is simulated and deduced to evaluate the risk reduction degree and driver safety protection effect in each scenario, so as to select the optimal emergency strategy from multiple sets of virtual emergency strategy scenarios.

9. The method according to claim 8, wherein In step S5, the emergency strategy includes slowing down and stopping the vehicle, adjusting the in-vehicle environmental parameters, and sending a distress message and vehicle location information to an emergency contact or emergency center via the vehicle communication system.

10. The method according to claim 9, wherein When 0 < comprehensive abnormal signal strength ≤ 1, the health risk level is mild abnormality; when 1 < comprehensive abnormal signal strength ≤ 3, the health risk level is moderate abnormality; When the comprehensive abnormal signal strength is >3, the health risk level is severe abnormality.

Images