Driving monitoring method and vehicle driving risk assessment system

Abstract

The invention provides a driving monitoring method and a vehicle driving risk assessment system. In the method, the safe response time relative to the driving state of the vehicle is determined. And evaluating a risk value by inputting the safety response time to a cut-off function. The larger the safety response time is, the smaller the risk value corresponding to the cut-off function is, and the shorter the safety response time is, the larger the risk value corresponding to the cut-off function is. Therefore, a driver can be timely reminded.

Description

Technical Field

[0001] The present invention relates to a monitoring technology, and particularly to a driving monitoring method and a vehicle driving risk assessment system. Background Art

[0002] During the driving of a vehicle, accidents may occur due to the poor physiological state of the driver. For example, when the driver is fatigued and the pupils of the eyes become smaller and blink more frequently, it will be difficult to concentrate, and the reaction time may slow down, thus causing a car accident.

[0003] On the other hand, existing driving anomaly detection systems may frequently send alarms due to minor anomalies, which are not only annoying but also affect the driving mood. For example, during high-speed driving, the anomaly detection system may trigger an anomaly alarm due to detecting minor fatigue of the driver. However, if the current vehicle position is far from the ramp or it is impossible to change to a low-speed lane, the continuous alarm will only make the driver feel helpless and bored. Even if the alarm system can be turned off, the abnormal state of the driver still persists, increasing the danger. Summary of the Invention

[0004] The present invention provides a driving monitoring method and a vehicle driving risk assessment system, which can timely remind the driver to reduce the occurrence of accidents and will not overly interfere with the driver.

[0005] The driving monitoring method according to an embodiment of the present invention includes (but is not limited to) the following steps: determining a safety reaction time corresponding to the vehicle driving state; and evaluating a risk value by inputting the safety reaction time into a cut-off function, wherein the larger the safety reaction time, the smaller the risk value corresponding to the cut-off function, and the smaller the safety reaction time, the larger the risk value corresponding to the cut-off function.

[0006] The vehicle driving risk assessment system according to an embodiment of the present invention includes (but is not limited to) a sensor, a memory, and a processor. The sensor detects the vehicle driving state. The memory stores program codes. The processor is coupled to the sensor and the memory, loads the program codes and executes: determining a safety reaction time corresponding to the vehicle driving state; and evaluating a risk value by inputting the safety reaction time into a cut-off function, wherein the larger the safety reaction time, the smaller the risk value corresponding to the cut-off function, and the smaller the safety reaction time, the larger the risk value corresponding to the cut-off function.

[0007] Based on the above, the driving monitoring method and the vehicle driving risk assessment system according to an embodiment of the present invention evaluate a risk value corresponding to the critical degree based on the safety reaction time, based on the characteristic that the cut-off function causes a larger increase in the output value as the input value exceeding the corresponding critical value decreases. The magnitude of this risk value can be used to determine whether to issue an alarm. Thus, when the critical degree of the vehicle is higher than a certain level, an alarm will be issued in a timely manner.

[0008] In order to make the above - mentioned features and advantages of the present invention more obvious and understandable, specific embodiments are given below and will be described in detail in conjunction with the accompanying drawings of the specification as follows. Description of the Drawings

[0009] Figure 1 It is a block diagram of components of a vehicle driving risk assessment system according to an embodiment of the present invention.

[0010] Figure 2 It is a flowchart of a driving monitoring method according to an embodiment of the present invention.

[0011] Figures 3A to 3F It is a schematic diagram showing exponential functions of multiple exponential rates according to an embodiment of the present invention.

[0012] Figure 4A and Figure 4B It is a schematic diagram showing exponential functions of different upper - limit times according to an embodiment of the present invention.

[0013] Figure 5 It is a flowchart of an alarm mechanism of an alarm processing module according to an embodiment of the present invention.

[0014] Figure 6 It is a flowchart of an alarm detection process according to the first embodiment of the present invention.

[0015] Figure 7 It is a schematic diagram showing a queue according to an embodiment of the present invention.

[0016] Figure 8 It is a schematic diagram showing the sorting of risk values in a queue in an application scenario.

[0017] Figure 9 It is a flowchart of an alarm elimination process according to the first embodiment of the present invention.

[0018] Figure 10 It is a schematic diagram showing the sorting of risk values in a queue in an application scenario.

[0019] Figure 11 It is a flowchart of an alarm detection process according to the second embodiment of the present invention.

[0020] Figure 12 It is a flowchart of an alarm elimination process according to the second embodiment of the present invention.

[0021] Description of the Reference Numerals:

[0022] 10: Vehicle driving risk assessment system

[0023] 11: Sensor

[0024] 111: Distance sensor

[0025] 112: Vehicle speed sensor

[0026] 113: Physiological sensor

[0027] 14: Memory

[0028] 141: Personnel detection module

[0029] 142: Vehicle detection module

[0030] 143: Alarm processing module

[0031] S210 - S220, S510 - S530, S605 - S623, S905 - S934: Steps

[0032] CF1 - CF8: Exponential functions

[0033] T REF : Reference time

[0034] QE: Queue

[0035] QL: Maximum capacity Detailed implementation manners

[0036] Figure 1 is a block diagram of the components of a vehicle driving risk assessment system 10 according to an embodiment of the present invention. Please refer to Figure 1 , the vehicle driving risk assessment system 10 includes (but is not limited to) a sensor 11, a memory 14, and a processor 15. The vehicle driving risk assessment system 10 can be an in-vehicle system, a mobile device, a wearable device, a tablet computer, a smart assistant device, or other electronic devices. In one embodiment, the vehicle driving risk assessment system 10 is installed on a vehicle. The vehicle can be various types of mobile vehicles. For example, a car, a truck, or a bus. In another embodiment, the vehicle driving risk assessment system 10 is externally connected to the vehicle.

[0037] In one embodiment, the sensor 11 includes a distance sensor 111.

[0038] The distance sensor 111 can be a radar, a lidar, a depth sensor, a time-of-flight (ToF) sensor, or a stereo camera. In one embodiment, the distance sensor 111 detects the distance between the vehicle (equipped with the distance sensor 111) and another vehicle or multiple vehicles.

[0039] In one embodiment, the sensor 11 includes a vehicle speed sensor 112.

[0040] The vehicle speed sensor 112 can be a position sensor (e.g., by inductive current, potentiometer, or optical encoder), a motion sensor (e.g., inertial measurement unit, accelerometer, and / or gyroscope), a pressure sensor (e.g., microelectromechanical or capacitive), a speed sensor (e.g., Hall effect sensor, magnetoresistive effect sensor, or passive sensor), or an electronic control unit (ECU) of the vehicle. In one embodiment, the vehicle speed sensor 112 detects the (moving) speed of the vehicle (equipped with the vehicle speed sensor 112).

[0041] In one embodiment, the sensor 11 includes a physiological sensor 113.

[0042] The physiological sensor 113 can be an image acquisition device or an infrared sensor. In one embodiment, the physiological sensor 113 takes pictures of the person (e.g., the driver) on the vehicle (equipped with the physiological sensor 113) to generate an acquired image.

[0043] The memory 14 can be any form of fixed or removable random access memory (RAM), read only memory (ROM), flash memory, traditional hard disk drive (HDD), solid state drive (SSD), or similar components. In one embodiment, the memory 14 is used to store program codes, software modules (e.g., the person detection module 141, the vehicle detection module 142, and the alarm processing module 143), configuration settings, data (e.g., speed, distance, time, risk value, or function), or files, and their embodiments will be described in detail later.

[0044] The processor 15 is coupled to the sensor 11 and the memory 14. The processor 15 can be a Central Processing Unit (CPU), a Graphic Processing Unit (GPU), or other programmable general-purpose or special-purpose microprocessors, Digital Signal Processors (DSPs), programmable controllers, Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), neural network accelerators, or other similar components or combinations of the above components. In one embodiment, one or more processors 15 are used to execute all or part of the operations of the vehicle driving risk assessment system 10, and can load and execute each program code, software module, file, and data stored in the memory 14. In some embodiments, the functions of the processor 15 can be implemented by software or chips.

[0045] In one embodiment, the processor 15 executes the personnel detection module 141, the vehicle detection module 142, and / or the alarm processing module 143. The functions of each module 141-143 will be described in detail in subsequent embodiments.

[0046] Hereinafter, the method described in the embodiments of the present invention will be described in conjunction with various devices, components, and modules in the vehicle driving risk assessment system 10. Each process of this method can be adjusted according to the implementation situation.

[0047] Figure 2 is a flowchart of a driving monitoring method according to an embodiment of the present invention. Please refer to Figure 2 , the processor 15 determines the relative safe reaction time of the vehicle driving state through the vehicle detection module 142 (step S210). In one embodiment, the safe reaction time includes measuring the speed of the vehicle and the distance between two vehicles (i.e., this vehicle and another vehicle). For example, the vehicle detection module 142 detects the speed of the vehicle through the vehicle speed sensor 112 and detects the distance from other vehicles through the distance sensor 111.

[0048] In one embodiment, the safe reaction time includes the physiological state of the driver. The processor 15 detects the physiological state through the personnel detection module 141. For example, the processor 15 captures an image of the driver (of the face or other body parts) through the physiological sensor 113 and identifies the physiological state based on image recognition technology (such as deep learning inference or image feature comparison). Features such as eye color, pupil size, pupil position, yawning, or blink count are all used for the physiological state.

[0049] This physiological state is related to the driver's ability to drive a vehicle. In one embodiment, the physiological state is lack of concentration in the eyes. For example, looking around left and right within a short period of time (e.g., 1, 3, or 5 seconds). In one embodiment, the physiological state is that the yawning frequency is higher than the corresponding threshold value. For example, more than 5 yawns in 1 minute, or 2 yawns within 10 seconds. In one embodiment, the physiological state is that the eyes are red or the pupils are constricted. For example, the area ratio of red in the three primary colors in the eyeball is greater than the corresponding threshold value; or the pupils are constricted by more than 5%. In one embodiment, the physiological state is that the blink frequency is higher than the corresponding threshold value. For example, the number of blinks within 5 seconds exceeds 2; or the number of blinks within 10 seconds exceeds 3. The above physiological states can be regarded as slightly abnormal states.

[0050] In one embodiment, the physiological state belongs to a severely abnormal state. A severely abnormal state is usually a situation where the driver has difficulty or is unable to drive. For example, continuously closing the eyes for three seconds, the head falling on the steering wheel, or the head continuously shaking.

[0051] It should be noted that there are other changes in the physiological state, and its content and definition can be adjusted according to actual needs.

[0052] In one embodiment, the processor 15 determines the safety reaction time based on the speed of the vehicle and its distance from another vehicle or more vehicles through the vehicle detection module 142. In one embodiment, the safety reaction time is the ratio of the distance to the speed. For example, T Cur is the safety reaction time, LF is the distance (e.g., in meters, but not limited thereto), and V car is the speed (e.g., in kilometers per hour, but not limited thereto).

[0053] The safety reaction time calculated from the current speed and the current distance (i.e., the vehicle distance) can represent the time required to collide with other vehicles. Therefore, the smaller the safety reaction time, the shorter the time to collide with other vehicles; the larger the safety reaction time, the longer the time to collide with other vehicles.

[0054] In other embodiments, the safety reaction time can be a function based on the ratio of the distance to the speed. For example, this ratio is multiplied by a weight value, or this ratio is added with a bias value.

[0055] Please refer to Figure 2, the processor 15 uses the alarm processing module 143 to evaluate the risk value by inputting the safety response time into the cut-off function (step S220). Specifically, one of the characteristics of this cut-off function is that if the input value is higher than the corresponding critical value, it will cause little difference in the output value of the cut-off function (for example, the difference is less than one or two percent). That is to say, the difference in the output values corresponding to any two input values greater than the corresponding critical value is small. On the other hand, if the input value is lower than the corresponding critical value, the output value of the cut-off function will increase significantly as the input value decreases. That is to say, the difference in the output values corresponding to any two input values less than the corresponding critical value is greater than the difference in the output values corresponding to any two input values greater than the corresponding critical value.

[0056] The value of the safety response time is used as the input value of the cut-off function. The processor 15 can set that the greater the safety response time, the smaller the corresponding risk value with respect to the cut-off function; the smaller the safety response time, the greater the corresponding risk value with respect to the cut-off function. That is to say, if the safety response time (representing the time required to collide with other vehicles) calculated from the current speed and the current distance (i.e., the vehicle distance) is smaller, it will cause the risk value (i.e., the output value of the cut-off function) to be greater (a higher chance of triggering an alarm); conversely, the greater the safety response time, the smaller the risk value and even remain unchanged after exceeding the corresponding critical value (a lower chance of triggering an alarm).

[0057] In one embodiment, the cut-off function is an exponential function. The processor 15 can set the upper limit time and the lower limit time of the safety response time. The upper limit time can be based on the upper limit of the detectable vehicle speed and / or the detectable distance. For example, if the maximum supported vehicle speed of the vehicle speed sensor 112 is 200 km / h, it can travel 200×1000 / 3600 = 55.55 meters per second. If the farthest detectable vehicle distance is 200 meters, it takes 200 / 55.55 = 3.6 seconds for the vehicle to travel 200 meters (which can be used as the upper limit time). The lower limit time can refer to the definition of the safety vehicle distance by research institutions. For example, the safety vehicle distance (in meters) is half of the vehicle speed (in km / h). If the vehicle speed is 60 km / h and the safety vehicle distance is 30 meters, then by conversion, 30 / (60*1000 / 3600) = 1.8 seconds. This 1.8 seconds can be set as the starting point of the significant change of the cut-off function (i.e., the corresponding threshold value of the above cut-off function), and reaches the highest point after 0.2 seconds, and the difference is 1.8 - 0.2 = 1.6 seconds (which can be used as the lower limit time).

[0058] The processor 15 can set the range interval of the risk value. For example, the range interval is 1 to 100, 1 to 50, or 1 to 500.

[0059] Next, the processor 15 can define an exponential function based on the upper limit time, lower limit time, and range interval. The upper limit time and lower limit time are used in the exponential function to normalize the safety response time. For example, the safety response time is normalized between 0 and 1, but not limited thereto. The range interval is related to the frequency of detecting the driving state of the vehicle (e.g., the speed of the vehicle, the distance between this vehicle and another vehicle, and / or the physiological state). That is, the frequency of monitoring the vehicle state or physiological state. For example, monitoring the state within the most recent 10 seconds or 20 seconds. Alternatively, the range interval is related to the execution frequency of subsequent risk value-based alarm triggering. For example, the frequency of calculating the risk value per second. If the frequency is higher, it is recommended that the range interval be a smaller range value to accumulate the risk values at more time points. For example, assume that the alarm processing module calculates the risk value only once per second and accumulates ten risk values within the most recent 10 seconds. When the sum of the accumulated risk values is greater than the trigger threshold value (e.g., 60, but can be defined according to the actual application) and an alarm is sent, the range interval can be set to cover the trigger threshold value. For example, the range interval is from 1 to 100.

[0060] In one embodiment, the exponential function is:

[0061]

[0062] , T Cur is the safety response time, T MIN is the lower limit time, T MAX is the upper limit time, ER is the Exponential Rate, O MAX is the upper limit of the range interval, O MIN is the lower limit of the range interval, and I Safe is the risk value.

[0063] The exponential rate ER should be set such that when the safety response time T Cur decreases downward and exceeds the corresponding threshold value (e.g., 1.8), the risk value I Safe can change rapidly, but when the value of the safety response time T Cur does not reach the corresponding risk value I before 1.8 Safe the change cannot be too large.

[0064] For example, Figures 3A to 3F is a schematic diagram illustrating exponential functions CF1 to CF6 with multiple exponential rates according to an embodiment of the present invention. Please refer to Figures 3A to 3F , assuming that the lower limit time T MIN is 1.6, the upper limit time T MAX is 3.6, the upper limit O MAX of the range interval is 100, and the lower limit O MIN of the range interval is 1. Figures 3A to 3FExponential functions CF1 to CF6 with exponential rates ER of 5, 10, 15, 20, 25, and 30 respectively.

[0065] Please refer to Figure 3A and Figure 3B , assuming that the input safety response time is the reference time T REF (for example, its value is 1.8), the corresponding risk value of the exponential function CF1 is approximately 60, but the corresponding risk value of the exponential function CF2 is approximately 35. That is to say, after the input value of the exponential function CF2 is lower than the reference time T REF , its corresponding risk value will rise from 35 to 100. Compared with the exponential function CF1, the increase amplitude of the corresponding risk value of the exponential function CF2 after the input value is lower than the reference time T REF is larger. However, the risk value of the exponential function CF1 can increase significantly after the input value is less than approximately 2.75.

[0066] Please refer to Figure 3C and Figure 3D , assuming that the input safety response time is the reference time T REF (for example, its value is 1.8), the corresponding risk value of the exponential function CF3 is approximately 20, but the corresponding risk value of the exponential function CF4 is approximately 15. That is to say, after the input value of the exponential function CF4 is lower than the reference time T REF , its corresponding risk value will rise from 15 to 100. Compared with the exponential function CF3, the increase amplitude of the corresponding risk value of the exponential function CF4 after the input value is lower than the reference time T REF is larger. However, the risk value of the exponential function CF3 can increase significantly after the input value is less than approximately 2.

[0067] Please refer to Figure 3E and Figure 3F , assuming that the input safety response time is the reference time T REF (for example, its value is 1.8), the corresponding risk value of the exponential function CF5 is approximately 10, but the corresponding risk value of the exponential function CF6 is approximately 5. That is to say, after the input value of the exponential function CF6 is lower than the reference time T REF , its corresponding risk value will rise from 5 to 100. Compared with the exponential function CF5, the increase amplitude of the corresponding risk value of the exponential function CF6 after the input value is lower than the reference time T REF is larger. However, the risk value of the exponential function CF5 can increase significantly after the input value is less than approximately 1.9.

[0068] In one embodiment, the processor 15 can set the exponential rate according to the physiological state. There are many types of physiological states, and different types can correspond to different exponential rates.

[0069] In one embodiment, when the physiological state is that the eyes are not focused, the processor 15 may set the exponential rate to the first rate; when the physiological state is that the yawning frequency is higher than the corresponding threshold value, the processor 15 may set the exponential rate to a second rate, and the second rate is greater than or equal to the first rate; when the physiological state is that the eyes are red or the pupils are constricted, the processor 15 may set the exponential rate to a third rate, and the third rate is greater than or equal to the second rate; when the physiological state is that the number of blinks is higher than the corresponding threshold value, the processor 15 may set the exponential rate to a fourth rate, and the fourth rate is greater than or equal to the third rate; when the physiological state belongs to a serious abnormal state, the processor 15 may set the safety reaction time to the lower limit time.

[0070] For example, if the physiological state is that the eyes are not focused, the exponential rate is 5. As Figure 3A shown, the chance of triggering an alarm increases when the safety reaction time is less than 2.75 seconds. If the physiological state is that the yawning frequency is higher than the corresponding threshold value, the exponential rate is 10. As Figure 3B shown, the chance of triggering an alarm increases when the safety reaction time is less than 2.25 seconds. If the physiological state is that the eyes are red or the pupils are constricted or the number of blinks is higher than the corresponding threshold value, the exponential rate is 20. As Figure 3D shown, the chance of triggering an alarm increases when the safety reaction time is less than 2 seconds. If the physiological state belongs to a serious abnormal state, the safety reaction time is equal to the lower limit time, and the risk value is set to the maximum value (for example, 100, but not limited thereto).

[0071] In addition to the exponential rate affecting the curve change of the exponential function, the upper limit time corresponding to the safety reaction time is also one of the influencing factors. For example, Figure 4A and Figure 4B are schematic diagrams of exponential functions with different upper limit times according to an embodiment of the present invention. Please refer to Figure 4A and Figure 4B . The upper limit time of the exponential function CF7 is 6, and the upper limit time of the exponential function CF8 is 8. As the upper limit time increases, the effective range of the safety reaction time can be increased, and it also affects the curve change of the exponential function. For example, after the input value of the exponential function CF7 is lower than the reference time T REF , its corresponding risk value will rise from 40 to 100. Compared with the exponential function CF8, the increase amplitude of the risk value corresponding to the exponential function CF7 after the input value is lower than the reference time T REF is larger. However, the risk value of the exponential function CF8 can increase significantly when the input value is less than approximately 2.75.

[0072] After applying the above Figures 3A to 3F numerical settings, the risk value can be simplified as follows:

[0073]

[0074] First, regarding formula (1): This part is actually a linear transformation that normalizes the numerical range of the safety reaction time to between 0 and 1.

[0075] The time range of [T MIN , T MAX is mapped from the original [1.6, 3.6] to [0, 1]. For example, when T Cur = 1.8, the mapped value is 0; when T Cur = 3.6, the mapped value is 1.

[0076] The [T MIN , T MAX is arranged in reverse / transposed order to [T MAX , T MIN , so the normalized maximum value 1 is subtracted from to reverse its time range, and the actual time range changes from [0, 1] to [1, 0].

[0077] Next, take the ER power of so that the numerical range of [T MAX , T MIN remains within [1, 0], but the values within this range are exponentially mapped to a non-linear range. For example, when the exponential rate ER is greater than 1, the original values (e.g., safety reaction time) close to 0 will become even closer to 0 after mapping, and the original values close to 1 will become even closer to 1 after mapping.

[0078] Finally, since the range interval of the risk value is defined as [1, 100], multiply by 99, and the numerical range interval becomes [0, 99], and then add 1 to the obtained value.

[0079] Therefore, this formula (2) maps the range of [T MAX , T MIN from [3.6, 1.6] to [1, 100]. In addition, during the process where the safety reaction time T Cur decreases from 3.6 to 1.6, the numerical value of the risk value I Safe will increase exponentially from 1 to 100; when the safety reaction time T Cur decreases below 1.8, the numerical value of the risk value I Safe will increase rapidly.

[0080] In another embodiment, the cut-off function is not limited to an exponential function and may be a function with the same or similar characteristics.

[0081] In one embodiment, the processor 15 may issue an alarm according to the risk value through the alarm processing module 143. Specifically, the magnitude of the risk value determines whether to trigger an alarm. The greater the risk value, the higher the chance of triggering an alarm; the smaller the risk value, the lower the chance of triggering an alarm. In one embodiment, the processor 15 accumulates the risk values at multiple consecutive time points and triggers an alarm based on the statistical value of these risk values (such as the sum or arithmetic mean). For example, the greater the statistical value of these risk values, the higher the chance of triggering an alarm; the smaller the statistical value of these risk values, the lower the chance of triggering an alarm.

[0082] In one embodiment, the processor 15 may issue an audible-related alarm through a speaker (such as emitting a warning sound or a reminder voice message); issue a visual-related alarm through a display (such as a warning light flashing or presenting a reminder icon); and issue an alarm related to text, symbols, or patterns to a device of a relevant unit through a communication transceiver. In one embodiment, the processor 15 may slow down the vehicle, switch to a slow lane, or stop the vehicle through the vehicle system.

[0083] Figure 5 is a flowchart illustrating the alarm mechanism of the alarm processing module 143 according to an embodiment of the present invention. Please refer to Figure 5 , the alarm processing module 143 determines the alarm status (step S510). The alarm status includes "True / Yes" and "False / No". "True / Yes" represents that an alarm is being issued, and "False / No" represents aborting / stopping / interrupting / canceling the alarm.

[0084] If the alarm status is "False / No" (for example, f alarm = False, f alarm is the alarm status), then the alarm processing module 143 executes the alarm detection process (step S520). At this time, the vehicle driving risk assessment system 10 is not sending an alarm, and the alarm processing module 143 detects whether an alarm needs to be sent. On the other hand, if the alarm status is "True / Yes" (for example, f alarm = True), then the alarm processing module 143 executes the alarm cancellation process (step S530). At this time, the vehicle driving risk assessment system 10 may detect an abnormal physiological state and / or the vehicle is in an unsafe distance / speed. Therefore, the vehicle driving risk assessment system 10 is continuously sending an alarm. On the other hand, the alarm processing module 143 continuously determines whether the physiological condition and / or the vehicle state reaches the condition for stopping sending the alarm. If the condition for stopping sending the alarm is reached, the alarm processing module 143 may reset the alarm status f alarm to False; if the condition for stopping sending the alarm has not been reached, the alarm status f alarm remains True (i.e., continues to issue an alarm).

[0085] Figure 6 is a flowchart of the alarm detection process S520 according to the first embodiment of the present invention. Please refer to Figure 6 , the processor 15 can add the risk value of the current time to the queue (step S610). The risk value of the current time refers to the risk value obtained from the safety response time at the current time point. The queue may further include one or more risk values of previous times. The risk value of a previous time refers to the risk value obtained from the safety response time at a previous time point. The previous time is earlier than the current time. The processor 15 can set the maximum capacity of the queue. For example, 10, 20, or 50 risk values. The queue is based on the first-in, first-out principle, and outputs the risk value at the head of the queue when the number of risk values reflected in the queue is greater than the maximum capacity of the queue (assuming the most recent / newest risk value is input from the tail end of the queue).

[0086] Figure 7 is a schematic diagram illustrating the queue QE according to an embodiment of the present invention. Please refer to Figure 7 , the queue QE has a maximum capacity QL. For example, 30, 50, or 60. Figure 7 The queue QE above has not added a risk value of 15 yet. Figure 7 The queue QE below has added a risk value of 15 to the right end and outputs the risk value of 0.01 at the leftmost end. The processor 15 can delete, ignore, or retain the risk value output by the queue QE.

[0087] In one embodiment, the value of the maximum capacity QL is related to the frequency of calculating the risk value and the frequency of monitoring the vehicle state or physiological state. For example, the product of the two frequencies (in times per second).

[0088] Please refer to Figure 6 , the processor 15 issues an alarm based on the sum of the risk values in the queue (step S620). Specifically, the processor 15 can determine the sum of the risk values in the queue (step S621). That is, add up the values of all the risk values in the queue. The processor 15 can compare the sum of the risk values in the queue with a trigger threshold value and determine whether this sum is greater than the trigger threshold value (step S622).

[0089] In response to the sum of the risk values in the queue being greater than the trigger threshold value, the processor 15 can issue an alarm (step S623) and change the alarm state from "false" to "true". In response to the sum of the risk values in the queue not being greater than the trigger threshold value, the alarm state remains "false", and the risk value of the current time will be added to the queue at the next time point (step S610).

[0090] Figure 8 is a schematic diagram of an application scenario illustrating the sorting of risk values in the queue. Please refer toFigure 8 , Table (1) shows the corresponding relationship between speed, distance, and time in the application scenario:

[0091] Table (1)

[0092]

[0093] Within the first 10 seconds, the fatigued driver maintained the vehicle speed at 80 kilometers per hour and kept a distance of 60 meters from the vehicle in front. After 10 seconds, due to the deceleration of the vehicle in front, the distance to the vehicle in front decreased to 40 meters. At this time, the alarm sounded. Two seconds later, the driver decelerated to 60 kilometers per hour, and three seconds later, the distance to the vehicle in front returned to 60 meters again, and the vehicle speed was 80 kilometers per hour. In addition, assume that the trigger threshold is 60, the risk value is calculated once every 0.1 seconds, the physiological and / or vehicle state within the most recent 5 seconds is monitored, and the maximum capacity of the queue is 50.

[0094] Since the initial alarm state is "false", the alarm detection process S520 is executed. A total of 100 risk values are calculated during the first 10 seconds. Since the vehicle speed has been maintained at 80 kilometers per hour, the safe reaction time can be calculated as 60 / (80*1000 / 3600) = 2.7 seconds. Assume that the corresponding output risk value of the cut-off function is 0.01. Therefore, this 0.01 risk value is stored in the queue QE with a maximum capacity of 50 values. If the maximum capacity has been reached, the processor 15 can cyclically remove the first element in the queue QE (for example, the risk value at the leftmost end of the figure) and add a new risk value at the tail end (for example, at the rightmost end of the figure). Therefore, at the 10th second, the sum of all risk values in the queue QE is 0.01×50 = 0.5, and it has not yet exceeded the alarm trigger threshold (i.e., the alarm will not be triggered).

[0095] In the next detection stage (10 seconds + 0.1 seconds = 10.1 seconds), since the distance between the vehicle and another vehicle suddenly reduces to 40 meters, the safety response time is reduced to 40 / (80 * 1000 / 3600) = 1.8 seconds, and the risk value corresponding to the output of the cut-off function is 15. The value of the risk value newly added to the end of the queue QE is 15. At this time, the sum of all risk values in the queue QE is 0.01×49 + 15 = 15.49, but it is still less than the alarm trigger threshold value, so the alarm will not be triggered. After continuously filling the risk value of 15 into the end of the queue QE three times, the sum of all risk values in the queue QE becomes 0.01×49 + 15×4 = 60.49 and is greater than the alarm trigger threshold value, so the alarm is issued and the alarm status is changed from "false" to "true". It can be seen from this that if the value of the risk value is small (for example, less than 20 or less than 30, and corresponding to a larger safety response time), it may be necessary to accumulate the risk values for a period of time (for example, 0.4 or 0.5 seconds) before the alarm status can be triggered.

[0096] Figure 9 is a flowchart of the alarm cancellation process S530 according to the first embodiment of the present invention. Please refer to Figure 9 , the processor 15 can sort the risk values in the queue according to the magnitude of the risk value (step S910). If the queue outputs a risk value exceeding the maximum capacity at its head end, all the risk values in the queue are re-sorted from largest to smallest starting from the head end of the queue. For example, Figure 10 is a schematic diagram showing the sorting of risk values in a queue in an application scenario. Please refer to Figure 8 and Figure 10 , Figure 8 The alarm is triggered at 10.4 seconds in. Re-sorting the risk values in the queue corresponding to this 10.4 seconds can obtain the sorting of the queue QE shown in the top row as shown in Figure 10 . Since 15 is the maximum value, it is sorted at the head end of the queue QE (the leftmost end in the figure).

[0097] Please refer to Figure 9 , the processor 15 can add the risk values of subsequent times to the queue (step S920). These subsequent times are later than the aforementioned current time. That is to say, at subsequent time points, the risk values will still be determined continuously, and the determined risk values will be added to the queue in sequence. It should be noted that the subsequent time is used to distinguish the aforementioned current time. When the subsequent time arrives, it becomes the current time.

[0098] The processor 15 may eliminate the alarm according to the sum of the risk values ​​in the queue (step S930). Specifically, the processor 15 may determine the sum of the risk values ​​in the queue (step S931). That is, the values ​​of all the risk values ​​in the queue are summed up. The processor 15 may compare the sum of the risk values ​​in the queue with the elimination threshold value, and determine whether the sum is less than the elimination threshold value (step S932).

[0099] In response to the fact that the sum is less than the elimination threshold, the processor 15 may determine whether the duration for which the sum of the subsequent time is less than the elimination threshold is greater than the time threshold (step S933). As long as the sum of the subsequent time is less than the elimination threshold, the duration is accumulated. In other words, within the time interval corresponding to the duration, the sum of all risk values ​​in the queue is less than the elimination threshold. Alternatively, as long as the sum of the subsequent time is less than the elimination threshold, the processor 15 may accumulate the elimination count.

[0100] In response to the duration being greater than the time threshold, the alarm is cleared (step S934), and the alarm state is changed from "true" to "false". Alternatively, in response to the elimination count being greater than the count threshold (e.g., 3, 5, or 10 times), the alarm is cleared (step S934), and the alarm state is changed from "true" to "false".

[0101] In response to the fact that the sum of the risk values ​​in the queue is not less than the elimination threshold value, the alarm state remains "true", the risk values ​​in the queue are reordered (step S910), and the risk value at the current time is added to the queue at the next time point (step S920). In addition, in response to the fact that the sum of the risk values ​​in the queue is not less than the elimination threshold value, the alarm state remains "true", the risk values ​​in the queue are reordered (step S910), and the risk value at the current time is added to the queue at the next time point (step S920).

[0102] In another embodiment, in response to the sum of the risk values ​​in the queue being less than the elimination threshold, the processor 15 may also directly eliminate the alarm.

[0103] Please refer to Figure 10 , starting from the 12.1 second, the speed of the vehicle changes to 60 kilometers per hour, the safety reaction time is 40 / (60*1000 / 3600)=2.4 seconds, and the corresponding output risk value of the cutoff function is 0.02. At this time, the sum of all risk values ​​of the queue QE is 0.01×46+0.02×1+15×3=45.48 and is less than the alarm elimination threshold value (e.g., 60). Next, the elimination counts whose sum is less than the alarm elimination threshold value begin to be accumulated. When the time reaches 15.1 seconds, the elimination count reaches the count threshold value (e.g., 3 times), and the processor 15 sets the alarm state to "no" and releases / eliminates the alarm.

[0104] After 15.2 seconds, since the alarm status is "No", the process returns to the alarm detection process S520.

[0105] Figure 11 It is a flowchart of the alarm detection process S520 according to the second embodiment of the present invention. Please refer to Figure 11 , and Figure 6 The difference from the first embodiment shown in is that before adding the risk value of the current time to the queue, the processor 15 also adjusts the risk value according to the abnormal state level (step S605). This abnormal state level is related to the physiological state of the driving vehicle. The physiological state has been described above and will not be elaborated here. For example, the adjusted risk value is the product of the original weight value and the weight value corresponding to the abnormal state level. The abnormal state level is divided into 5 levels from mild to severe, where the abnormal state levels 3 to 5 are defined as severe abnormal states and an alarm needs to be issued immediately. Therefore, the weight values corresponding to levels 3 to 5 can be set to be higher than the alarm trigger threshold. If it is desired to shorten the alarm trigger time of level 2 by half compared to level 1, the weight value of level 2 can be set to twice the weight value of level 1. Assume that the weight value of level 1 can be defined as 1. However, the weight values corresponding to the abnormal state levels can still be adjusted according to actual needs.

[0106] Figure 12 It is a flowchart of the alarm cancellation process according to the second embodiment of the present invention. Please refer to Figure 12 , and Figure 9 The difference from the first embodiment shown in is that before sorting the risk values in the queue according to the numerical size, the processor 15 also adjusts the risk value according to the abnormal state level (step S905). This abnormal state level is as described above Figure 11 and will not be elaborated here.

[0107] In summary, in the driving monitoring method and vehicle driving risk assessment system according to the embodiments of the present invention, the safety response time based on speed and distance is converted into a risk value corresponding to the degree of urgency according to the characteristics of the cut-off function. The risk values at multiple time points are temporarily stored through a queue. Whether to trigger an alarm or cancel the alarm is determined according to the sum of the multiple risk values in the queue. Thereby, the driver is reminded timely at an appropriate vehicle distance and will not interfere with the driver too frequently.

[0108] Although the present invention has been disclosed as above with embodiments, it is not intended to limit the present invention. Any person skilled in the art within the technical field, without departing from the concept and scope of the present invention, can make some changes and modifications. Therefore, the protection scope of the present invention shall be subject to what is defined by the claims.

Claims

1. A driving monitoring method, comprising: Determine a safe reaction time relative to a vehicle driving state; as well as A risk value is evaluated by inputting the safety reaction time into a cutoff function, wherein the greater the safety reaction time, the smaller the risk value corresponding to the cutoff function, and the smaller the safety reaction time, the greater the risk value corresponding to the cutoff function.

2. The driving monitoring method as claimed in claim 1, wherein the cutoff function is an exponential function, and the driving monitoring method further comprises: Set an upper limit and a lower limit of the safety response time; Set a range of the risk value; as well as The exponential function is defined according to the upper limit time, the lower limit time and the range interval, wherein the upper limit time and the lower limit time are used in the exponential function to normalize the safety reaction time, and the range interval is related to the frequency of detecting the driving state of the vehicle.

3. The driving monitoring method as claimed in claim 2, wherein the exponential function is: T Cur is the safety response time, T MIN is the lower limit time, T MAX is the upper limit time, ER is an exponential rate, O MAX is the upper limit of the range, O MIN is the lower limit of the range, and I Safe is the risk value.

4. The driving monitoring method according to claim 3, further comprising: detecting a physiological state, wherein the physiological state is related to an ability to drive a vehicle; as well as The exponential rate is set according to the physiological state.

5. The driving monitoring method as claimed in claim 4, wherein the step of setting the exponential rate according to the physiological state comprises: In response to the physiological state being that the eyes are not focused, the index rate is set to a first rate; In response to the physiological state being that the yawning frequency is higher than the corresponding threshold value, setting the exponential rate to a second rate, wherein the second rate is greater than or equal to the first rate; In response to the physiological state being redness of the eyes or constriction of the pupil, setting the exponential rate to a third rate, wherein the third rate is greater than or equal to the second rate; In response to the physiological state being that the number of blinks is higher than a corresponding threshold value, setting the exponential rate to a fourth rate, wherein the fourth rate is greater than or equal to the third rate; and In response to the physiological state being a serious abnormal state, the safety response time is set to the lower limit time.

6. The driving monitoring method of claim 1, wherein after evaluating the risk value, the driving monitoring method further comprises: Adding the risk value at a current time to a queue, wherein the queue also includes the risk value at at least one previous time, the at least one previous time being earlier than the current time, and outputting the risk value in the queue in response to the number of risk values ​​in the queue being greater than a maximum capacity of the queue; as well as An alarm is issued based on the sum of the risk values ​​in the queue.

7. The driving monitoring method according to claim 1, further comprising: The risk value is adjusted according to an abnormal state level, wherein the abnormal state level is related to a physiological state of driving a vehicle.

8. The driving monitoring method as claimed in claim 6, wherein the step of issuing the alarm according to the sum of the risk values ​​in the queue comprises: comparing the sum of the risk values ​​in the queue with a trigger threshold; as well as In response to the sum of the risk values ​​in the queue being greater than the trigger threshold, the alarm is issued.

9. The driving monitoring method according to claim 6, further comprising: Sort the risk values ​​in the queue according to their numerical values; adding the risk value of a subsequent time to the queue, wherein the subsequent time is later than the current time; as well as The alarm is eliminated based on the sum of the risk values ​​in the queue.

10. The driving monitoring method as claimed in claim 8, wherein the step of eliminating the alarm according to the sum of the risk values ​​in the queue comprises: comparing the sum of the risk values ​​in the queue with an elimination threshold; as well as In response to the sum of the risk values ​​in the queue being less than the elimination threshold, the alarm is eliminated. 11 . The driving monitoring method as claimed in claim 1 , wherein the vehicle driving state comprises a speed of a vehicle and a distance between the vehicle and another vehicle.

12. A vehicle driving risk assessment system, comprising: a sensor for detecting a driving state of a vehicle; a memory storing a program code; as well as A processor, coupled to the sensor and the memory, loads the program code and executes: Determine a safe reaction time relative to the driving state of the vehicle; and A risk value is evaluated by inputting the safety reaction time into a cutoff function, wherein the greater the safety reaction time, the smaller the risk value corresponding to the cutoff function, and the smaller the safety reaction time, the greater the risk value corresponding to the cutoff function.

13. The vehicle driving risk assessment system as claimed in claim 12, wherein the cutoff function is an exponential function, and the processor further executes: Set an upper limit and a lower limit of the safety response time; Setting a range of risk values; and The exponential function is defined according to the upper limit time, the lower limit time and the range interval, wherein the upper limit time and the lower limit time are used in the exponential function to normalize the safety reaction time, and the range interval is related to the frequency of detecting the driving state of the vehicle.

14. The vehicle driving risk assessment system as claimed in claim 13, wherein the exponential function is: T Cur is the safety response time, T MIN is the lower limit time, T MAX is the upper limit time, ER is an exponential rate, O MAX is the upper limit of the range, O MIN is the lower limit of the range, and I Safe is the risk value.

15. The vehicle driving risk assessment system as claimed in claim 14, wherein the processor further executes: detecting a physiological state, wherein the physiological state is related to a performance ability to drive a vehicle, and the vehicle driving state includes a speed of a vehicle and a distance between the vehicle and another vehicle; and The exponential rate is set according to the physiological state.

16. The vehicle driving risk assessment system as claimed in claim 15, wherein the processor further executes: In response to the physiological state being that the eyes are not focused, the index rate is set to a first rate; In response to the physiological state being that the yawning frequency is higher than the corresponding threshold value, setting the exponential rate to a second rate, wherein the second rate is greater than or equal to the first rate; In response to the physiological state being redness of the eyes or constriction of the pupil, setting the exponential rate to a third rate, wherein the third rate is greater than or equal to the second rate; In response to the physiological state being that the number of blinks is higher than a corresponding threshold value, setting the exponential rate to a fourth rate, wherein the fourth rate is greater than or equal to the third rate; and In response to the physiological state being a serious abnormal state, the safety response time is set to the lower limit time.

17. The vehicle driving risk assessment system as claimed in claim 12, wherein the processor further executes: Adding the risk value at a current time to a queue, wherein the queue also includes the risk value at at least one previous time, the at least one previous time being earlier than the current time, and outputting the risk value in the queue in response to the number of risk values ​​in the queue being greater than a maximum capacity of the queue; and An alarm is issued based on the sum of the risk values ​​in the queue.

18. The vehicle driving risk assessment system as claimed in claim 12, wherein the processor further executes: The risk value is adjusted according to an abnormal state level, wherein the abnormal state level is related to a physiological state of driving a vehicle.

19. The vehicle driving risk assessment system as claimed in claim 17, wherein the processor further executes: comparing the sum of the risk values ​​in the queue with a trigger threshold; In response to the sum of the risk values ​​in the queue being greater than the trigger threshold, the alarm is issued; comparing the sum of the risk values ​​in the queue with an elimination threshold; as well as In response to the sum of the risk values ​​in the queue being less than the elimination threshold, the alarm is eliminated.

20. The vehicle driving risk assessment system as claimed in claim 17, wherein the processor further executes: Sort the risk values ​​in the queue according to their numerical values; adding the risk value at a subsequent time to the queue, wherein the subsequent time is later than the current time; and The alarm is eliminated based on the sum of the risk values ​​in the queue.

Images