Driver distraction is a leading cause of traffic accidents,with fatigue being a significant contributor.This paper introduces a novel method for detecting driver distraction by analyzing facial features using machine ...Driver distraction is a leading cause of traffic accidents,with fatigue being a significant contributor.This paper introduces a novel method for detecting driver distraction by analyzing facial features using machine deep learning and 68 face model.The proposed system assesses driver tiredness by measuring the distance between key facial landmarks,such as the distance between the eyes and the angle of the mouth,to evaluate signs of drowsiness or disengagement.Real-time video feed analysis allows for continuous monitoring of the driver’s face,enabling the system to detect behavioral cues associated with distraction,such as eye closures or changes in facial expressions.The effectiveness of this method is demonstrated through a series of experiments on a dataset of driver videos,which proves that the approach can accurately assess tiredness and distraction levels under various driving conditions.By focusing on facial landmarks,the system is computationally efficient and capable of operating in real-time,making it a practical solution for in-vehicle safety systems.This paper discusses the system’s performance,limitations,and potential for future enhancements,including integration with other in-vehicle technologies to provide comprehensive driver monitoring.展开更多
The aim of this study was to explore the effectiveness of behavioral evaluation measures for predicting drivers＇ subjective drowsiness. Behavioral measures included neck bending angle, back pressure, foot pressure, C...The aim of this study was to explore the effectiveness of behavioral evaluation measures for predicting drivers＇ subjective drowsiness. Behavioral measures included neck bending angle, back pressure, foot pressure, COP （center of pressure） movement on sitting surface and tracking error in driving simulator task. Drowsy states were predicted by means of the multinomial logistic regression model where behavioral measures and subjective evaluation of drowsiness corresponded to independent variables and a dependent variable, respectively. First, we compared the effectiveness of two methods （correlation coefficient-based method and odds ratio-based method） for determining the order of entering behavioral measures into the prediction model. It was found that the prediction accuracy did not differ between both methods. Second, the prediction accuracy was compared among the numbers of behavioral measures. The prediction accuracy did not differ among four, five and six behavioral measures and it was concluded that entering at least four behavioral measures into the prediction model is enough to achieve higher prediction accuracy. Third, the prediction accuracy was compared between the strongly drowsy and the weakly drowsy groups. The prediction accuracy differed between the two groups and the proposed method was effective under the condition where drowsiness was induced to a larger extent.展开更多
Today,fatalities,physical injuries,and significant economic losses occur due to car accidents.Among the leading causes of car accidents is drowsiness behind the wheel,which can affect any driver.Drowsiness and sleepin...Today,fatalities,physical injuries,and significant economic losses occur due to car accidents.Among the leading causes of car accidents is drowsiness behind the wheel,which can affect any driver.Drowsiness and sleepiness often have associated indicators that researchers can use to identify and promptly warn drowsy drivers to avoid potential accidents.This paper proposes a spatiotemporal model for monitoring drowsiness visual indicators from videos.This model depends on integrating a 3D convolutional neural network(3D-CNN)and long short-term memory(LSTM).The 3DCNN-LSTM can analyze long sequences by applying the 3D-CNN to extract spatiotemporal features within adjacent frames.The learned features are then used as the input of the LSTM component for modeling high-level temporal features.In addition,we investigate how the training of the proposed model can be affected by changing the position of the batch normalization(BN)layers in the 3D-CNN units.The BN layer is examined in two different placement settings:before the non-linear activation function and after the non-linear activation function.The study was conducted on two publicly available drowsy drivers datasets named 3MDAD and YawDD.3MDAD is mainly composed of two synchronized datasets recorded from the frontal and side views of the drivers.We show that the position of the BN layers increases the convergence speed and reduces overfitting on one dataset but not the other.As a result,the model achieves a test detection accuracy of 96%,93%,and 90%on YawDD,Side-3MDAD,and Front-3MDAD,respectively.展开更多
Nowadays days,the chief grounds of automobile accidents are driver fatigue and distractions.With the development of computer vision technology,a cutting-edge system has the potential to spot driver distractions or sle...Nowadays days,the chief grounds of automobile accidents are driver fatigue and distractions.With the development of computer vision technology,a cutting-edge system has the potential to spot driver distractions or sleepiness and alert them,reducing accidents.This paper presents a novel approach to detecting driver tiredness based on eye and mouth movements and object identification that causes a distraction while operating a motor vehicle.Employing the facial landmarks that the camera picks up and sends to classify using a Convolutional Neural Network(CNN)any changes by focusing on the eyes and mouth zone,precision is achieved.One of the tasks that must be performed in the transit system is seat belt detection to lessen accidents caused by sudden stops or high-speed collisions with other cars.A method is put forth to use convolution neural networks to determine whether the motorist is wearing a seat belt when a driver is sleepy,preoccupied,or not wearing their seat belt,this system alerts them with an alarm,and if they don’t wake up by a predetermined time of 3 s threshold,an automatic message is sent to law enforcement agencies.The suggested CNN-based model exhibits greater accuracy with 97%.It can be utilized to develop a system that detects driver attention or sleeps in real-time.展开更多
One of the major causes of road accidents is sleepy drivers.Such accidents typically result in fatalities and financial losses and disadvantage other road users.Numerous studies have been conducted to identify the dri...One of the major causes of road accidents is sleepy drivers.Such accidents typically result in fatalities and financial losses and disadvantage other road users.Numerous studies have been conducted to identify the driver’s sleepiness and integrate it into a warning system.Most studies have examined how the mouth and eyelids move.However,this limits the system’s ability to identify drowsiness traits.Therefore,this study designed an Accident Detection Framework(RPK)that could be used to reduce road accidents due to sleepiness and detect the location of accidents.The drowsiness detectionmodel used three facial parameters:Yawning,closed eyes(blinking),and an upright head position.This model used a Convolutional Neural Network(CNN)consisting of two phases.The initial phase involves video processing and facial landmark coordinate detection.The second phase involves developing the extraction of frame-based features using normalization methods.All these phases used OpenCV and TensorFlow.The dataset contained 5017 images with 874 open eyes images,850 closed eyes images,723 open-mouth images,725 closed-mouth images,761 sleepy-head images,and 1084 non-sleepy head images.The dataset of 5017 images was divided into the training set with 4505 images and the testing set with 512 images,with a ratio of 90:10.The results showed that the RPK design could detect sleepiness by using deep learning techniques with high accuracy on all three parameters;namely 98%for eye blinking,96%for mouth yawning,and 97%for head movement.Overall,the test results have provided an overview of how the developed RPK prototype can accurately identify drowsy drivers.These findings will have a significant impact on the improvement of road users’safety and mobility.展开更多
In the modern world,the increasing prevalence of driving poses a risk to road safety and necessitates the development and implementation of effective monitoring systems.This study aims to enhance road safety by propos...In the modern world,the increasing prevalence of driving poses a risk to road safety and necessitates the development and implementation of effective monitoring systems.This study aims to enhance road safety by proposing a dual-modal solution for detecting driver drowsiness,which combines heart rate monitoring and face recognition technologies.The research objectives include developing a non-contact method for detecting driver drowsiness,training and assessing the proposed system using pre-trained machine learning models,and implementing a real-time alert feature to trigger warnings when drowsiness is detected.Deep learning models based on convolutional neural networks(CNNs),including ResNet and DenseNet,were trained and evaluated.The CNN model emerged as the top performer compared to ResNet50,ResNet152v2,and DenseNet.Laboratory tests,employing different camera angles using Logitech BRIO 4K Ultra HD Pro Stream webcam produces accurate face recognition and heart rate monitoring.Real-world vehicle tests involved six participants and showcased the system’s stability in calculating heart rates and its ability to correlate lower heart rates with drowsiness.The incorporation of heart rate and face recognition technologies underscores the effectiveness of the proposed system in enhancing road safety and mitigating the risks associated with drowsy driving.展开更多
The harmful effects of technological devices, including smart phones have been increasingly suspected among university students;bedtimes have become increasingly later at night, and leisure activities often extend thr...The harmful effects of technological devices, including smart phones have been increasingly suspected among university students;bedtimes have become increasingly later at night, and leisure activities often extend through the night. Likewise, availability and need of increasing part-time job hours have been considered. The purpose of this research was to determine the relationship among lifestyles, quality of sleep, and daytime drowsiness of nursing students of University A. The research was conducted in June 2015, when student life rhythms were considered stable after two months of lectures. Responses with missing values or with inappropriate answers were excluded. Of the data collected from 96 respondents, only 71 were acceptable. The survey focused on lifestyle, daytime sleepiness (using ESS: Epworth Sleepiness Scale) and quality of subjective sleep (using the PSQI: Pittsburgh Sleep Quality Index). Approval was obtained from the Research Ethics Committee of Shikoku University. While in this study, more than half (63.4%) of the students had poor quality of sleep, however, there was no relationship between their quality of sleep and daytime drowsiness, or between their lifestyles and the quality of sleep. These findings suggest that while university students’ use of technological devices is suspected to influence on sleep deprivation and consequent daytime drowsiness, the findings did not provide the evidence.展开更多
In this paper, we proposed a new concept: depth of drowsiness, which can more precisely describe the drowsiness than existing binary description. A set of effective markers for drowsiness: normalized band norm was suc...In this paper, we proposed a new concept: depth of drowsiness, which can more precisely describe the drowsiness than existing binary description. A set of effective markers for drowsiness: normalized band norm was successfully developed. These markers are invariant from voltage amplitude of brain waves, eliminating the need for calibrating the voltage output of the brain-computer interface devices. A new polling algorithm was designed and implemented for computing the depth of drowsiness. The time cost of data acquisition and processing for each estimate is about one second, which is well suited for real-time applications. Test results with a portable brain-computer interface device show that the depth of drowsiness computed by the method in this paper is generally invariant from ages of test subjects and sensor channels (P3 and C4). The comparison between experiment and computing results indicate that the new method is noticeably better than one of the recent methods in terms of accuracy for predicting the drowsiness.展开更多
Introduction: Over-drowsiness is a condition with serious consequences, including road accidents. The condition, however, is often ignored both by carers as well as victims themselves. The aim of the present study was...Introduction: Over-drowsiness is a condition with serious consequences, including road accidents. The condition, however, is often ignored both by carers as well as victims themselves. The aim of the present study was to investigate the factors associated with excessive drowsiness in Cotonou, Benin 2023, along with its influence on the occurrence of crashes among truck drivers. Methods: This was a descriptive and analytical cross-sectional study, held from March 13 to April 10, 2023, focusing on large truck drivers over 18 years of age, selected by convenience from parking lots in and around the city of Cotonou. Data collected using questionnaires on socio-demographic and behavioral factors, sleeping habits and working conditions were processed using Stata 15.0 software. Excessive drowsiness was defined by a score above 10 on the Epworth scale. Associated factors were found by multiple logistic regression, at a threshold of 0.05. Results: Altogether 304 drivers, all male and aged 35.98 ± 8.42 years, were surveyed. The prevalence of excessive drowsiness was 29.2%. The associated factors identified were not practicing sport OR = 2.27, CI95% = [1.33;3.86], p = 0,002;sleep duration per working day OR = 1.82;CI95% = [1.06;3.11], p = 0,027 and average distance travelled per day OR = 3.40;CI95% = [1.53;7.56], p = 0,003. Excessive drowsiness was associated with a 1.73-fold increased risk of road accidents (CI95% [1.04;2.87];p = 0.03). Conclusion: Communicating excessive drowsiness and its associated factors to all the stakeholders in the haulage chain is essential to help prevent road accidents.展开更多
The aim of this study was to predict drivers＇ drowsy states with high risk of encountering a crash and prevent drivers from continuing to drive under such drowsy states with high risk of crash. While the participants...The aim of this study was to predict drivers＇ drowsy states with high risk of encountering a crash and prevent drivers from continuing to drive under such drowsy states with high risk of crash. While the participants were required to carry out a simulated driving task, EEG （Electro encephalography） （EEG-MPF and EEG-α/β）, ECG （Electrocradiogram） （RRV3）, t racking error, an d subjective rating on drowsiness were measured. On the basis of such measurements, an attempt was made to predict the point in time with high crash risk using Bayesian estimation of posterior probability of drowsiness, tracking error, and subjective drowsiness. As a result of applying the proposed method to the data of each participant, it was verified that the proposed method could predict the point in time with high crash risk before the point in time of crash.展开更多
为了提高疲劳检测的精度,通过驾驶模拟试验采集了15位中青年有经验驾驶员的车辆横向位置、方向盘操控、眼动等多源数据并计算疲劳特征指标,同时采集驾驶员主观疲劳程度并通过视频回放进行校核,在此基础上建立疲劳等级与特征指标的决策...为了提高疲劳检测的精度,通过驾驶模拟试验采集了15位中青年有经验驾驶员的车辆横向位置、方向盘操控、眼动等多源数据并计算疲劳特征指标,同时采集驾驶员主观疲劳程度并通过视频回放进行校核,在此基础上建立疲劳等级与特征指标的决策树模型,结果表明,对于区别疲劳等级最显著的变量有闭眼时间比例(percentage of eye closure,PERCLOS)、车道偏移标准差、越线时空面积、方向盘反转率,且上述变量与疲劳等级呈正相关;PERCLOS为最优的疲劳等级划分变量,并获取了2个重要阈值:当PERCLOS小于2.8%时,驾驶员处于严重疲劳状态的比例为零;当PERCLOS大于21.9%时,驾驶员处于未疲劳状态的比例为零;该模型预测的总正确率为64.31%.为了校验模型,从15位驾驶员中随机选取了4位进行模型校验试验.校核结果表明该模型的正确率达63.22%.模型在2次试验中都未发现将严重疲劳识别为未疲劳的情况.展开更多
文摘Driver distraction is a leading cause of traffic accidents,with fatigue being a significant contributor.This paper introduces a novel method for detecting driver distraction by analyzing facial features using machine deep learning and 68 face model.The proposed system assesses driver tiredness by measuring the distance between key facial landmarks,such as the distance between the eyes and the angle of the mouth,to evaluate signs of drowsiness or disengagement.Real-time video feed analysis allows for continuous monitoring of the driver’s face,enabling the system to detect behavioral cues associated with distraction,such as eye closures or changes in facial expressions.The effectiveness of this method is demonstrated through a series of experiments on a dataset of driver videos,which proves that the approach can accurately assess tiredness and distraction levels under various driving conditions.By focusing on facial landmarks,the system is computationally efficient and capable of operating in real-time,making it a practical solution for in-vehicle safety systems.This paper discusses the system’s performance,limitations,and potential for future enhancements,including integration with other in-vehicle technologies to provide comprehensive driver monitoring.
文摘The aim of this study was to explore the effectiveness of behavioral evaluation measures for predicting drivers＇ subjective drowsiness. Behavioral measures included neck bending angle, back pressure, foot pressure, COP （center of pressure） movement on sitting surface and tracking error in driving simulator task. Drowsy states were predicted by means of the multinomial logistic regression model where behavioral measures and subjective evaluation of drowsiness corresponded to independent variables and a dependent variable, respectively. First, we compared the effectiveness of two methods （correlation coefficient-based method and odds ratio-based method） for determining the order of entering behavioral measures into the prediction model. It was found that the prediction accuracy did not differ between both methods. Second, the prediction accuracy was compared among the numbers of behavioral measures. The prediction accuracy did not differ among four, five and six behavioral measures and it was concluded that entering at least four behavioral measures into the prediction model is enough to achieve higher prediction accuracy. Third, the prediction accuracy was compared between the strongly drowsy and the weakly drowsy groups. The prediction accuracy differed between the two groups and the proposed method was effective under the condition where drowsiness was induced to a larger extent.
文摘Today,fatalities,physical injuries,and significant economic losses occur due to car accidents.Among the leading causes of car accidents is drowsiness behind the wheel,which can affect any driver.Drowsiness and sleepiness often have associated indicators that researchers can use to identify and promptly warn drowsy drivers to avoid potential accidents.This paper proposes a spatiotemporal model for monitoring drowsiness visual indicators from videos.This model depends on integrating a 3D convolutional neural network(3D-CNN)and long short-term memory(LSTM).The 3DCNN-LSTM can analyze long sequences by applying the 3D-CNN to extract spatiotemporal features within adjacent frames.The learned features are then used as the input of the LSTM component for modeling high-level temporal features.In addition,we investigate how the training of the proposed model can be affected by changing the position of the batch normalization(BN)layers in the 3D-CNN units.The BN layer is examined in two different placement settings:before the non-linear activation function and after the non-linear activation function.The study was conducted on two publicly available drowsy drivers datasets named 3MDAD and YawDD.3MDAD is mainly composed of two synchronized datasets recorded from the frontal and side views of the drivers.We show that the position of the BN layers increases the convergence speed and reduces overfitting on one dataset but not the other.As a result,the model achieves a test detection accuracy of 96%,93%,and 90%on YawDD,Side-3MDAD,and Front-3MDAD,respectively.
基金Deputyship for Research&Innovation,Ministry of Education in Saudi Arabia for funding this research work through Project Number MoE-IF-UJ-22-4100409-1.
文摘Nowadays days,the chief grounds of automobile accidents are driver fatigue and distractions.With the development of computer vision technology,a cutting-edge system has the potential to spot driver distractions or sleepiness and alert them,reducing accidents.This paper presents a novel approach to detecting driver tiredness based on eye and mouth movements and object identification that causes a distraction while operating a motor vehicle.Employing the facial landmarks that the camera picks up and sends to classify using a Convolutional Neural Network(CNN)any changes by focusing on the eyes and mouth zone,precision is achieved.One of the tasks that must be performed in the transit system is seat belt detection to lessen accidents caused by sudden stops or high-speed collisions with other cars.A method is put forth to use convolution neural networks to determine whether the motorist is wearing a seat belt when a driver is sleepy,preoccupied,or not wearing their seat belt,this system alerts them with an alarm,and if they don’t wake up by a predetermined time of 3 s threshold,an automatic message is sent to law enforcement agencies.The suggested CNN-based model exhibits greater accuracy with 97%.It can be utilized to develop a system that detects driver attention or sleeps in real-time.
基金The Faculty of Information Science and Technology,Universiti Kebangsaan Malaysia,provided funding for this research through the Research Grant“An Intelligent 4IR Mobile Technology for Express Bus Safety System Scheme DCP-2017-020/2”.
文摘One of the major causes of road accidents is sleepy drivers.Such accidents typically result in fatalities and financial losses and disadvantage other road users.Numerous studies have been conducted to identify the driver’s sleepiness and integrate it into a warning system.Most studies have examined how the mouth and eyelids move.However,this limits the system’s ability to identify drowsiness traits.Therefore,this study designed an Accident Detection Framework(RPK)that could be used to reduce road accidents due to sleepiness and detect the location of accidents.The drowsiness detectionmodel used three facial parameters:Yawning,closed eyes(blinking),and an upright head position.This model used a Convolutional Neural Network(CNN)consisting of two phases.The initial phase involves video processing and facial landmark coordinate detection.The second phase involves developing the extraction of frame-based features using normalization methods.All these phases used OpenCV and TensorFlow.The dataset contained 5017 images with 874 open eyes images,850 closed eyes images,723 open-mouth images,725 closed-mouth images,761 sleepy-head images,and 1084 non-sleepy head images.The dataset of 5017 images was divided into the training set with 4505 images and the testing set with 512 images,with a ratio of 90:10.The results showed that the RPK design could detect sleepiness by using deep learning techniques with high accuracy on all three parameters;namely 98%for eye blinking,96%for mouth yawning,and 97%for head movement.Overall,the test results have provided an overview of how the developed RPK prototype can accurately identify drowsy drivers.These findings will have a significant impact on the improvement of road users’safety and mobility.
文摘In the modern world,the increasing prevalence of driving poses a risk to road safety and necessitates the development and implementation of effective monitoring systems.This study aims to enhance road safety by proposing a dual-modal solution for detecting driver drowsiness,which combines heart rate monitoring and face recognition technologies.The research objectives include developing a non-contact method for detecting driver drowsiness,training and assessing the proposed system using pre-trained machine learning models,and implementing a real-time alert feature to trigger warnings when drowsiness is detected.Deep learning models based on convolutional neural networks(CNNs),including ResNet and DenseNet,were trained and evaluated.The CNN model emerged as the top performer compared to ResNet50,ResNet152v2,and DenseNet.Laboratory tests,employing different camera angles using Logitech BRIO 4K Ultra HD Pro Stream webcam produces accurate face recognition and heart rate monitoring.Real-world vehicle tests involved six participants and showcased the system’s stability in calculating heart rates and its ability to correlate lower heart rates with drowsiness.The incorporation of heart rate and face recognition technologies underscores the effectiveness of the proposed system in enhancing road safety and mitigating the risks associated with drowsy driving.
文摘The harmful effects of technological devices, including smart phones have been increasingly suspected among university students;bedtimes have become increasingly later at night, and leisure activities often extend through the night. Likewise, availability and need of increasing part-time job hours have been considered. The purpose of this research was to determine the relationship among lifestyles, quality of sleep, and daytime drowsiness of nursing students of University A. The research was conducted in June 2015, when student life rhythms were considered stable after two months of lectures. Responses with missing values or with inappropriate answers were excluded. Of the data collected from 96 respondents, only 71 were acceptable. The survey focused on lifestyle, daytime sleepiness (using ESS: Epworth Sleepiness Scale) and quality of subjective sleep (using the PSQI: Pittsburgh Sleep Quality Index). Approval was obtained from the Research Ethics Committee of Shikoku University. While in this study, more than half (63.4%) of the students had poor quality of sleep, however, there was no relationship between their quality of sleep and daytime drowsiness, or between their lifestyles and the quality of sleep. These findings suggest that while university students’ use of technological devices is suspected to influence on sleep deprivation and consequent daytime drowsiness, the findings did not provide the evidence.
文摘In this paper, we proposed a new concept: depth of drowsiness, which can more precisely describe the drowsiness than existing binary description. A set of effective markers for drowsiness: normalized band norm was successfully developed. These markers are invariant from voltage amplitude of brain waves, eliminating the need for calibrating the voltage output of the brain-computer interface devices. A new polling algorithm was designed and implemented for computing the depth of drowsiness. The time cost of data acquisition and processing for each estimate is about one second, which is well suited for real-time applications. Test results with a portable brain-computer interface device show that the depth of drowsiness computed by the method in this paper is generally invariant from ages of test subjects and sensor channels (P3 and C4). The comparison between experiment and computing results indicate that the new method is noticeably better than one of the recent methods in terms of accuracy for predicting the drowsiness.
文摘Introduction: Over-drowsiness is a condition with serious consequences, including road accidents. The condition, however, is often ignored both by carers as well as victims themselves. The aim of the present study was to investigate the factors associated with excessive drowsiness in Cotonou, Benin 2023, along with its influence on the occurrence of crashes among truck drivers. Methods: This was a descriptive and analytical cross-sectional study, held from March 13 to April 10, 2023, focusing on large truck drivers over 18 years of age, selected by convenience from parking lots in and around the city of Cotonou. Data collected using questionnaires on socio-demographic and behavioral factors, sleeping habits and working conditions were processed using Stata 15.0 software. Excessive drowsiness was defined by a score above 10 on the Epworth scale. Associated factors were found by multiple logistic regression, at a threshold of 0.05. Results: Altogether 304 drivers, all male and aged 35.98 ± 8.42 years, were surveyed. The prevalence of excessive drowsiness was 29.2%. The associated factors identified were not practicing sport OR = 2.27, CI95% = [1.33;3.86], p = 0,002;sleep duration per working day OR = 1.82;CI95% = [1.06;3.11], p = 0,027 and average distance travelled per day OR = 3.40;CI95% = [1.53;7.56], p = 0,003. Excessive drowsiness was associated with a 1.73-fold increased risk of road accidents (CI95% [1.04;2.87];p = 0.03). Conclusion: Communicating excessive drowsiness and its associated factors to all the stakeholders in the haulage chain is essential to help prevent road accidents.
文摘The aim of this study was to predict drivers＇ drowsy states with high risk of encountering a crash and prevent drivers from continuing to drive under such drowsy states with high risk of crash. While the participants were required to carry out a simulated driving task, EEG （Electro encephalography） （EEG-MPF and EEG-α/β）, ECG （Electrocradiogram） （RRV3）, t racking error, an d subjective rating on drowsiness were measured. On the basis of such measurements, an attempt was made to predict the point in time with high crash risk using Bayesian estimation of posterior probability of drowsiness, tracking error, and subjective drowsiness. As a result of applying the proposed method to the data of each participant, it was verified that the proposed method could predict the point in time with high crash risk before the point in time of crash.
文摘为了提高疲劳检测的精度,通过驾驶模拟试验采集了15位中青年有经验驾驶员的车辆横向位置、方向盘操控、眼动等多源数据并计算疲劳特征指标,同时采集驾驶员主观疲劳程度并通过视频回放进行校核,在此基础上建立疲劳等级与特征指标的决策树模型,结果表明,对于区别疲劳等级最显著的变量有闭眼时间比例(percentage of eye closure,PERCLOS)、车道偏移标准差、越线时空面积、方向盘反转率,且上述变量与疲劳等级呈正相关;PERCLOS为最优的疲劳等级划分变量,并获取了2个重要阈值:当PERCLOS小于2.8%时,驾驶员处于严重疲劳状态的比例为零;当PERCLOS大于21.9%时,驾驶员处于未疲劳状态的比例为零;该模型预测的总正确率为64.31%.为了校验模型,从15位驾驶员中随机选取了4位进行模型校验试验.校核结果表明该模型的正确率达63.22%.模型在2次试验中都未发现将严重疲劳识别为未疲劳的情况.
