A method and system for detecting fatigue driving of a vehicle

Abstract

The application relates to a kind of automobile fatigue driving detection method and detection system, by obtaining the pressure value F of steering wheel, according to the change of pressure value F in detection period T l , obtain the pressure value variation F l , according to detection period T l And pressure value variation F l , calculate the first fatigue value P1;Collect the heart rate of driver, according to the change of heart rate in detection period, obtain the heart rate variation N l , according to detection period T l And heart rate variation N l , calculate the second fatigue value P2;Identify the blink data of driver, according to the change of blink times in detection period, obtain the blink times variation Z l And blink period variation t l , according to detection period T l , blink times variation Z l And blink period variation t l , calculate the third fatigue value P3;Calculate the total fatigue value P in detection period, P=P1+P2+P 3， If P exceeds threshold value, then fatigue driving alarm is carried out.The application combines hand grip, heart rate, facial visual information, judges the fatigue degree of human body comprehensively, and improves the accuracy of detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of detection systems, and particularly relates to a vehicle fatigue driving detection system. BACKGROUND

[0002] Vehicles have become an essential tool for people, however, in long-distance travel and the logistics industry, the phenomenon of fatigue driving often occurs, which is a bad driving habit that brings great safety hazards to road safety and great safety risks to oneself and others, therefore, fatigue driving behavior needs to be detected and appropriate measures need to be taken to avoid major traffic accidents.

[0003] At present, there are many methods for detecting fatigue driving behavior, some of which recognize the driver's eyes and facial features through video image recognition, such as closing eyes, yawning, etc., thereby generating a fatigue warning to alert the driver, such a method has more misjudgments, and the fatigue state cannot be identified in time based on the statistical method. Some fatigue driving behavior detection is detected by brain waves, but brain wave detection requires the driver to wear a brain wave collection device, and the brain wave collection device with many connections will affect the emergency operation of the driver and the cost of manufacturing the brain wave collection device is relatively high. At the same time, the fatigue degree of the driver is based on multi-dimensional judgment, such as the force of holding the steering wheel, heart rate, etc., and single data collection detection is very easy to misjudge, because everyone's physical condition, driving method, etc. is different.

[0004] With the in-depth research of the detection method, a method for judging whether fatigue driving is through multi-dimensional data collection has appeared, such as the invention patent with publication number CN112590800B discloses an intelligent recognition steering wheel based on multi-sensor fatigue driving and a recognition method, which comprehensively collects data through pressure sensors, blood oxygen sensors and cameras in order to improve the detection accuracy, but this detection method is only multi-sensor data collection and analysis, and fatigue driving is recognized by independent threshold judgment, although this redundant design improves the detection accuracy to a certain extent, it still cannot solve the problem of false positives in single data collection detection. SUMMARY

[0005] The purpose of the present application is to provide a vehicle fatigue driving detection method and a detection system to solve the above problems existing in the current vehicle fatigue driving detection.

[0006] To achieve the above purpose, the present application adopts the following technical solutions:

[0007] A vehicle fatigue driving detection method, by acquiring the pressure value F of the steering wheel, according to the change of the pressure value F in the detection period T l , the pressure value change amount Fl According to the detection cycle T l and pressure value change F l The first fatigue value P1 is calculated; the driver's heart rate is collected, and the heart rate change N is obtained based on the changes in heart rate during the detection period. l According to the detection cycle T l and heart rate variability N l The second fatigue value P2 is calculated; driver blink data is identified, and the change in blink count Z is obtained based on the change in blink count within the detection period. l and the change in blink cycle t l According to the detection cycle T l Change in blink rate Z l and the change in blink cycle t l The third fatigue value P3 is calculated; the total fatigue value P within the detection period is calculated, where P = P1 + P2 + P3; if P exceeds the threshold, a fatigue driving alarm is triggered.

[0008] Further optimization involves using a flexible pressure sensor module mounted on the steering wheel to acquire the steering wheel pressure value F, with a single acquisition cycle of 1 minute, and calculating the pressure value change F. l P1=k1×F l ×T l Where k1 is the influence coefficient, k1=0.03.

[0009] Further optimization involves using a flexible pressure sensor module to acquire 100 pressure values ​​F at a frequency of 10Hz. The top ten maximum values ​​and the bottom ten minimum values ​​are then removed using a bubble sort algorithm, and the average pressure F is calculated. μ Continuous sampling for 1 minute to obtain six average pressure values ​​F μ The average pressure F collected from the six samples μ Calculate the standard deviation σ F If σ F If the value is less than 2, the data is valid. The valid data is then used for calculation, starting from the first valid data point, to calculate the pressure change F. l .

[0010] Further optimization involves acquiring the driver's heart rate using a heart rate sensor module mounted on the steering wheel, with a single data acquisition cycle of 30 seconds, and calculating the heart rate change N. l P2 = k2 × N l ×T l， Where k2 is the influence coefficient, k2=1.

[0011] Further optimization involves collecting 10 valid values ​​in a single acquisition cycle. After removing the maximum and minimum values ​​from the collected data, the standard deviation σ is calculated. N If σN If <3, the data is valid, and the valid heart rate value is obtained by averaging the valid values, and the valid heart rate value is used for calculation, and the heart rate change N is calculated from the first valid data. l .

[0012] Further preferably, the camera module is used to collect the facial image of the driver to obtain the blink data of the driver, and a single collection cycle is one minute, and the blink frequency change Z and the blink cycle change t are calculated. l l P3=k3×Z l +k4×t l Wherein, k3 and k4 are influence coefficients, k3=10, and k4=30.

[0013] Further preferably, the blink frequency data is collected in five single collection cycles, and the standard deviation σ is calculated after removing one maximum value and one minimum value from the obtained data. Z If σ Z <1.5, the data is valid, and the valid data is averaged to obtain the blink frequency change Z from the first valid data. l .

[0014] Further preferably, the blink cycle data is collected in a single collection cycle, and the average value is calculated after removing one maximum value and one minimum value from the obtained data to obtain the blink cycle data in the collection cycle. The standard deviation σ is calculated after removing one maximum value and one minimum value from the five continuous blink cycle data. t If σ t <1.5, the data is valid, and the valid data is averaged to obtain the blink cycle change t from the first valid data. l .

[0015] Further preferably, if P>100, the fatigue driving alarm is performed, and if the detection cycle T l >4, the fatigue driving alarm is directly performed.

[0016] A fatigue driving detection system for a vehicle, comprising a flexible pressure sensor module, a flexible heart rate sensor module, a camera module, a Bluetooth module, a voice module, and a main controller. The flexible pressure sensor module, the flexible heart rate sensor module, the camera module, the Bluetooth module, and the voice module are all connected to the main controller. The flexible pressure sensor module and the flexible heart rate sensor module are arranged on the steering wheel to detect the grip force of the driver on the steering wheel and the heart rate of the driver. The camera is used to collect the facial information of the driver. The voice module is used to perform fatigue driving alarm. The main controller is further provided with a 5G module, and the main controller is connected to a cloud server through the 5G module. ​

[0017] Advantages of the present application:

[0018] The automobile fatigue driving detection method of the present application calculates a first fatigue value according to the pressure value change of the steering wheel by collecting the gripping force of the driver holding the steering wheel, calculates a second fatigue value according to the change of the heart rate by collecting the heart rate of the driver, calculates a third fatigue value according to the change of the blinking frequency and the change of the blinking period by identifying the blinking data of the driver, and superimposes the three fatigue values as a total fatigue value according to the weight of the first fatigue value, the second fatigue value and the third fatigue value, and judges whether the total fatigue value exceeds the threshold value to identify the fatigue driving of the driver. The present application combines the hand gripping force, the heart rate and the facial visual information to comprehensively judge the fatigue degree of the human body, and improves the accuracy of detection. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 is the principle diagram of the automobile fatigue driving detection system of the present application;

[0020] Figure 2 is the principle diagram of the automobile fatigue driving detection method of the present application. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the present application will be clearly and completely described in combination with the drawings in the embodiments of the present application.

[0022] Embodiment 1 of the present application:

[0023] The automobile fatigue driving detection method in the present embodiment is to detect the state of the driver by the flexible pressure sensor module and the flexible heart rate sensor module arranged on the steering wheel in cooperation with the camera module arranged in the vehicle, to realize the real-time detection and reminding of the fatigue driving and ensure the driving safety.

[0024] In the present embodiment, Figure 1 As shown in the figure, the vehicle is provided with a main controller, the main controller is provided with a Bluetooth module, the flexible pressure sensor module and the flexible heart rate sensor module are connected with the main controller through the Bluetooth module, the main controller is further connected with a voice module, the voice module can send sound to remind the driver, so as to avoid the fatigue driving of the driver, the flexible pressure sensor module is used to collect the pressure data of the steering wheel to judge whether the driver holds the steering wheel and the force of holding, the flexible heart rate sensor module collects the heart rate of the driver to judge the physical state of the driver, the camera module collects the facial information of the driver to judge the blinking frequency and the blinking period to evaluate the mental state of the driver, the data collected by the above modules are calculated in the main controller to calculate a total fatigue value P, when the total fatigue value P exceeds the threshold value, the fatigue driving alarm is given through the voice module, so as to ensure the driving safety.

[0025] In this embodiment, Figure 2 As shown, the main controller calculates the fatigue value of the driver according to the data of the flexible pressure sensor, denoted as the first fatigue value, denoted as P1, calculates the fatigue value of the driver according to the data of the flexible heart rate sensor module, denoted as the second fatigue value, denoted as P2, and calculates the fatigue value of the driver according to the data obtained by the camera module, denoted as the third fatigue value, denoted as P3, in this embodiment:

[0026] P = P1 + P2 + P3

[0027] When the P value exceeds the threshold value, an alarm is issued through the voice module.

[0028] The calculation method of P1, P2 and P3 will be described in detail below.

[0029] The flexible pressure sensor module arranged on the steering wheel obtains the pressure value F of the steering wheel, that is, the holding force of the driver on the steering wheel. When the value obtained by the pressure sensor is less than 200N and the duration is greater than 5s, it indicates that the driver does not manipulate the steering wheel. At this time, the main controller issues a reminder through the voice module, and the camera module records the video picture of this time and uploads it to the server.

[0030] Otherwise, the pressure value is continuously collected and calculated in the main controller.

[0031] Specifically, the collection frequency of the flexible pressure sensor module is 10Hz, that is, 10 pressure values are collected per second. First, 100 pressure values F are collected in 10s. Through the bubble algorithm, the first ten maximum values and the last ten minimum values are removed, and the remaining values are averaged to obtain the pressure average value F μ In this embodiment, the single collection period is 1 minute. After continuous collection for 1 minute, six pressure average values F μ are obtained. μ The standard deviation is calculated. If the standard deviation σ F is less than 2, the data is judged to be valid, otherwise the average value F μ is not included in the statistical range. Finally, the valid data is calculated. From the first valid data, the pressure value change F l in the detection period T l is calculated. The pressure value change is calculated by taking the absolute value.

[0032] Through the above data collection, the first fatigue value in the detection period T l can be calculated:

[0033] P1 = k1 × F l × T lWhere k1 is the influence coefficient, k1=0.03.

[0034] The driver's heart rate is obtained by a heart rate sensor module installed on the steering wheel, and a second fatigue value P2 is calculated.

[0035] Specifically, the heart rate sensor module collects heart rate data in 30-second intervals. If no heart rate is detected within 30 seconds, it indicates that the driver's hands have left the steering wheel, and the heart rate data collected in this interval is invalid. Otherwise, 10 valid values ​​are collected in a single interval. After removing the maximum and minimum values ​​from the collected data, the standard deviation σ is calculated. N If σ N If the heart rate is less than 3, the data is valid. The valid heart rate values ​​are then averaged to obtain the effective heart rate value, which is used in the detection period T. l In this process, the heart rate change N is calculated starting from the first effective heart rate value. l， The absolute value of the heart rate change is used for calculation.

[0036] Based on the above data collection, the detection period T can be calculated. l The second fatigue value within:

[0037] P2=k2×N l ×T l , where k2 is the influence coefficient, k2=1.

[0038] The driver's blink data is collected by a camera module installed inside the vehicle to calculate the third fatigue value P3.

[0039] In this embodiment, the camera module determines the fatigue value by recognizing the change in the number of blinks and the change in the blink cycle. The recognition algorithm is an existing algorithm and belongs to the prior art, so it will not be described in detail here.

[0040] Specifically, the blink data is collected in one-minute intervals. If the eye leaves the camera module's collection area within one minute, the collection is invalid. After collecting data for five valid intervals, the maximum and minimum values ​​are removed, and the standard deviation σ is calculated. Z If σ Z If the average value is less than 1.5, the data is valid. The average value of the valid data is selected, and the calculation is performed starting from the average value of the first valid data over the detection period T. l Change in blink rate Z l The change in blink rate is calculated using absolute values.

[0041] Similarly, the blink cycle data is collected at the same time as the blink frequency data, and the single collection period of the blink cycle data is also one minute. If the personnel leave the camera module collection area within one minute, this collection is invalid. In a single collection period, the blink cycle data is collected, and the average value is calculated after removing a maximum value and a minimum value. The obtained value is the blink cycle data in a single collection period. The blink cycle data is continuously collected five times, and the standard deviation sigma is calculated after removing a maximum value and a minimum value. t If sigma t < 1.5, the data is valid, and the average value of the valid data is selected. The detection period T l is calculated from the average value of the first valid data. l The blink cycle change amount t

[0042] Through the above data collection, the third fatigue value in the detection period T l can be calculated:

[0043] P3 = k3 * Z l + k4 * t l , wherein k3 and k4 are influence coefficients, k3 = 10, and k4 = 30.

[0044] In this embodiment, the unit of the detection period T l is hour, the unit of the blink frequency change amount Z l is piece, and the unit of the blink cycle change amount t l is second.

[0045] When the value of P exceeds 100, it indicates that the driver has fatigue driving phenomenon, and the main controller controls the voice module to issue a sound to remind the driver.

[0046] At present, according to the regulations of the traffic management department, it is necessary to rest after continuous driving for more than 4 hours to avoid fatigue driving. In this embodiment, when the detection period T l is greater than four hours, the main controller controls the voice module to issue a sound to remind the driver.

[0047] When the voice module reminds, the camera module synchronously records the current video information.

[0048] In this embodiment, the main controller is also provided with a 5G module, and the main controller is connected with the cloud server through the 5G module. When the fatigue state is detected, the relevant video and sensor data information can be remotely sent to the cloud server, which is convenient for accident analysis.

Claims

1. An automobile fatigue driving detection method, characterized in that: it comprises a flexible pressure sensor module, a flexible heart rate sensor module, a camera module, a Bluetooth module, a voice module and a main controller, the flexible pressure sensor module, the flexible heart rate sensor module, the camera module, the Bluetooth module and the voice module are connected with the main controller, the flexible pressure sensor module and the flexible heart rate sensor module are arranged on a steering wheel, which are used for detecting the grip of the driver on the steering wheel and the heart rate of the driver, the camera module is used for collecting the face information of the driver, the voice module is used for fatigue driving alarm, and the main controller is further provided with a 5G module, and the main controller is connected with a cloud server through the 5G module. Obtain the pressure value F of the steering wheel, according to the change of the pressure value F in the detection period T l Obtain the pressure value change F l , according to the detection period T l And the pressure value change F l , calculate the first fatigue value P1; the unit of detection period T l Is hour; the pressure value F of the steering wheel is obtained through the flexible pressure sensor module arranged on the steering wheel, the single collection period is 1 minute, the pressure value change F l Is calculated, P1=k1xF l XT l , wherein, k1 is the influence coefficient, k1=0.03; the flexible pressure sensor module collects 100 pressure values F at the frequency of 10Hz, removes the first ten maximum values and the last ten minimum values through the bubble algorithm, and takes the pressure average value F μ , collects continuously for 1 minute, obtains six pressure average values F μ , calculates the standard deviation σ μ Of the six collected pressure average values F F , if σ F <2, the data is valid, the last valid data is taken to calculate, the pressure value change F l Is calculated from the first valid data. Collecting the heart rate of the driver, according to the change of the heart rate in the detection period T l , obtaining the heart rate change amount N l , according to the detection period T l and the heart rate change amount N l , calculating the second fatigue value P2; obtaining the heart rate of the driver through the heart rate sensor module arranged on the steering wheel, the single collection period is 30 seconds, calculating the heart rate change amount N l , P2=k2×N l ×T l， , wherein k2 is the influence coefficient, k2=1; in the single collection period, collecting 10 effective values, after removing one maximum value and one minimum value from the collected data, calculating the standard deviation σ N , if σ N <3, the data is valid, finally taking the average value of the effective values to obtain the effective heart rate value, taking the effective heart rate value to calculate, from the first effective data, calculating the heart rate change amount N l ; The driver's blink data is identified, and the blink frequency change Z and the blink period change t are obtained according to the change in the detection period T l l l The third fatigue value P3 is calculated according to the detection period T l , the blink frequency change Z l and the blink period change t l The driver's face image is collected by the camera module to obtain the driver's blink data, and the blink frequency change Z l and the blink period change t l are calculated, P3=k3×Z l +k4×t l , wherein k3 and k4 are influence coefficients, k3=10, and k4=30; the blink frequency data is collected in five single collection periods, the data obtained is removed from a maximum value and a minimum value, the standard deviation σ Z is calculated, if σ Z <1.5, the data is valid, and the average value of the valid data is finally taken, the blink frequency change Z l is calculated from the average value of the first valid data; the blink period data is collected in a single collection period, the average value is calculated after the data obtained is removed from a maximum value and a minimum value, the blink period data in the collection period is obtained, the blink period data is continuously collected for five times, the standard deviation σ t is calculated after a maximum value and a minimum value are removed, if σ t <1.5, the data is valid, and the average value of the valid data is finally selected, the blink period change t l is calculated from the average value of the first valid data;​​ The detection period T is calculated l The total fatigue value P in the inside is calculated, wherein P=P1+P2+P3; if P exceeds the threshold value, the fatigue driving alarm is performed.

2. The automobile drowsy driving detection method according to claim 1, characterized by: If P > 100, fatigue driving alarm is given, and if detection period T l > 4, fatigue driving alarm is given directly.

3. A system for implementing the method for detecting fatigue driving of a vehicle according to claim 1, characterized in that: ​

Images