Open Access
Add to BibSonomy
Abstract
A driver safety assisting system is essential to reduce the probability of traffic accidents. But most of the existing driver safety assisting systems are simple reminders that cannot improve the driver's driving status. This paper proposes a driver safety assisting system to reduce the driver's fatigue degree by the light with different wavelengths that affect people's moods. The system consists of a camera, an image processing chip, an algorithm processing chip, and an adjustment module based on quantum dot LEDs (QLEDs). Through this intelligent atmosphere lamp system, the experimental results show that blue light reduced the driver’s fatigue degree when just turned on; but as time went on, the driver’s fatigue degree rebounded rapidly. Meanwhile, red light prolonged the driver's awake time. Different from blue light alone, this effect can remain stable for a long time. Based on these observations, an algorith was designed to quantify the degree of fatigue and detect its rising trend. In the early stage, the red light is used to prolong the awake time and the blue light to suppress when the fatigue value increases, so as to maximize the awake driving time. The result showed that our device prolonged the awake driving time of the drivers by 1.95 times and reduced fatigue during driving: the quantitative value of fatigue degree generally decreased by about 0.2 times. In most experiments, the subjects were able to complete four hours of safe driving, which reached the maximum length of continuous driving at night allowed by China laws. In conclusion, our system changes the assisting system from a reminder to a helper, thus effectively reducing the driving risk.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Driver safety assisting systems consist of sensing devices and processors installed in the vehicle which assist the drivers to drive the car safely. For example, an augmented reality head-up display (AR HUD) enables a driver to know the speed, gas quantity and other information at night more clearly [1], a vehicle mounted radar lets the driver know the distance from the obstacle when the line of sight is poor [2], and a fatigue detector tells the driver’s fatigue degree [3]. Basically, these systems collect and analyze external information then provide car or road condition reminders to the drivers [4–6], or remind the drivers’ own status so that they can have a rest in time [7–10]. But all these assisting systems are lack of feedback; in other words, they only provide information, not actual help.
Recently, researchers have found that lights with different wavelengths can adjust human’s psychological and physiological activities: blue light inhibits melatonin secretion which can keep human away from falling asleep, meanwhile the blue light with the wavelength near 460 nm is harmful to eyes, causing diseases like maculopathy [11,12]; red light provides a relaxing environment for human beings to relieve the tension of the brain and further research has found that a proper exposure to red light contributes to vision health [13,14]. Based on these findings, different health light lamps have been developed, such as a lamp that can simulate the change of sunlight in a day [15] and a special lamp capable of treating myopia [16]. In the field of optically assisted driving, the wavelength and intensity of light have an important influence on the drivers’ mental state [17]. For the drivers who are insensitive to blue light, the blue light can resist driver’s fatigue [18]. Further research found that when blue light was used to enhance alertness, its intensity should be limited [19]. However, the possible side effects of blue light have also been pointed out: alerting effects of light impairs accuracy in precision tasks, such as keeping a proper car position [20]. In order to obtain more credible conclusions, Fitri et al. carried out detailed research on the reaction process of blue light. The result showed that blue light exposure increased drivers’ alertness and attention when responding to visual distraction at night, especially in the first appearance [21]. Different from the previous conclusion, a new point of view has been put forward that the suppressive effect of light at night is not a yes-or-no response, but a dose–response relationship − light intensity should be more important than wavelength [22]. However, it can be found that light source used in the above experiment were made according to some characteristics of light at a specific wavelength, and because of the lack of interaction with the external environment they can only realize some pre-designed functions. However, the change of a driver's mental state is a dynamic and complex process, it is almost impossible to follow up a dynamic process with pre-set procedures.
In order to achieve the goal of using light to adjust mood state, a system consisted of sensing, image processing and light adjustment modules was provided. Many previous studies have found that blue light can wake up people because it can inhibit the production of melatonin to avoid drowsiness and stimulate visual rod cells to make people more alert. However, through experiments it can be found that although blue light did achieve the above effect in a short time, the degree of fatigue rose rapidly in the later stage, which led to an actually shorter awake time when there was blue light only. Therefore, it is obviously an unwise choice to irradiate with blue light all the time. The result is the whole fatigue change process could be divided into three stages, whilst using appropriate light source adjustment at different stages could maximize the length of safe driving. The result shows that our system can accurately identify the driver's mood state during driving, through light regulation to reduce driver’s fatigue by 20% during driving and extend awake driving for about one hour (1.95 times of the original time). From the perspective of the whole driving process with our system, the accidents caused by fatigue within four consecutive hours (the maximum continuous driving hours allowed by China laws) at night have been reduced to nearly zero. At last, our system is a versatile platform to extend with more functions, which will enhance the safe driving assistance and make night driving safer and more efficient.
2. Methods and design
Figure 1(a) is the flow chart of the whole system. This system includes the following modules: sensing, image processing and lighting adjustment modules.
Fig. 1. (a) Flow chart of the atmosphere lamp system, consisting of three modules: sensing, image processing and lighting adjustment modules. (b) and (c) are TEM images (insets show corresponding HRTEM images), (e) and (f) are the corresponding size distribution histograms, (d) and (g) are quantum dot spectrograms and photos, respectively.
The sensing module consists of an infrared enhanced camera and a data transfer interface. The infrared enhancement camera takes pictures of the driver's face in the cab, and the interface transmits them to the image processing module.
Image processing module consists of a FPGA&ARM development board (ZYNQ-7000). This module has three parts: acceleration, process and PERCLOS calculation and analysis. Acceleration is the programmable logic (PL) part of the development board, which can complete image preprocessing faster than the processor based on ARM chips.
The processing system (PS) part of the development board confirms the positions of eyes and mouth by comparing with the trained face database. The last part is used to calculate and analyze the change of fatigue value. An algorithm for calculating fatigue degree is called PERCLOS (percent eye closure). PERCLOS is an indicator to quantify driver fatigue value through the opening and closing degree of eyes. To carry out PERCLOS detection, it is necessary to collect a certain number of driver face images in a certain time. Then, the face images are transferred to a chip with image processing function to carry out face detection, eye detection and eye state classification. A region of interest (ROI) is selected on the detected face boundary to locate the eyes and then to classify the eye state as open or closed. Finally, the number of images with eyes closed divided by the total number of images collected during this period to get the PERCLOS value. [23]. Compared with other fatigue detection means such as heart rate and electroencephalogram, PERCLOS is non-invasive, less expensive and easy to realize on a running vehicle. Most importantly, this method has been verified by Federal Highway Administration (FHWA) and National Highway Traffic Safety Administration (NHTSA) of the US, which also has many successful practices [24,25]. When one PERCLOS is calculated, the driver's fatigue state in the sampling interval can be quantified; when multiple consecutive PERCLOS are calculated, the driver's current fatigue state can be accurately judged.
After got the fatigue state, the lighting adjustment module will be put into use. This module consists of quantum dot LEDs (QLEDs) with a control circuit. Appropriate light will be used to improve the driver's working condition. In order to get a better lighting, quantum dots were selected as the luminescent materials [26,27]. Compare with traditional luminescent materials, quantum dots have a narrower full-width at half maximum (FWHM), which means a higher color purity [28], this will make the atmosphere lamp richer in colors to meet the market demands. The photoluminescence quantum yield (PLQY) of quantum dots is very high (up to 100%) [29–31], and the luminous efficacy ranges from 80 to 100 lm/W. Higher luminous efficacy means higher energy efficiency, which is in line with the concept of energy conservation. Figure 1(b) and (c) shows TEM images of quantum dots (the insets show the corresponding HRTEM images), (d) and (e) show the corresponding size distribution histograms, (f) shows the spectra of quantum dots and (g) shows the photos of quantum dots and the LEDs made by them.
A driving simulation software was chosen to build the experimental platform in order to safely and repeatedly test the night driving fatigue states of the drivers. The software is European Truck Simulator 2 since it was reliable in precedent simulated driving tests [32]. The simulated steering wheel and pedal were connected to a PC to make the experimental platform closer to the real driving environment. Figure 2 reveals the completed experimental platform. In a real car, the QLEDs were installed on the center console, instrument panel, doors and control lever, which were similar to the installation position of traditional atmosphere lamps. So, in this simulated driving platform, the distance between QLEDs and the driver’s eyes is the same as the distance between the center console and the driver, which was 30-50 cm.
Fig. 2. Experimental platform. The driving simulator runs on a desktop computer. The driver fatigue detection system runs on ZYNQ.
3. Results and discussion
3.1 Effects of light wavelength
At first, the influence of different colors of light on drivers should be explored. There has been some consensus on the impact of different wavelengths of light on human beings. For example, short wavelength of blue light will damage vision [33], and blue light will inhibit melatonin secretion, so as to achieve the effect of inhibiting falling sleep [34]. In the study of the influence of light on human mood, exposure to bright white light and short wavelength monochromatic light significantly improved subjective and objective alertness indicators [35,36], meanwhile the light source containing more red light was conducive to fatigue recovery [37]. But there are also many disputes in this field [38–40]. In order to determine the specific effects of different wavelengths of light on driver’s fatigue, red, green and blue QLEDs with the same illumination (40 lux) were tested. These QLEDs were made by blue light GaN chips and cadmium-based quantum dots. The quantum dots were synthesized by the “one-pot” approach [41,42], and the emission wavelength was controlled by adjusting the proportion of reagents. After the quantum dots of the specified wavelength were synthesized, they were mixed with epoxy resin in the same proportion and spin coated on a 365 nm GaN light chip and waited the epoxy resin to solidify [43]. Figure 3 shows the light output characteristics of the system, and the luminous efficacy is very close to the commercial LEDs purchased from the electronic market.
Fig. 3. (a) Spatial distribution of the QLEDs. (b) Absolute spectra of QLEDs. (c) Luminous efficacy comparison between blue QLEDs and commercial LEDs. (d) Luminous efficacy comparison between red QLEDs and commercial LEDs.
The volunteers participating in the test were all young men aged between 22 and 25 years old, and all have vehicle driving licenses. All the experiments in the paper were conducted on weekdays to ensure that the experimenters have similar activity tracks. The diet of volunteers has been restricted and they have been told to avoid eating foods containing caffeine on the day of the experiment. The whole experiment lasted two months, random working days of this two months were selected for the experiment and each experiment lasted at least three hours. All the test processes were reported to and approved by the experimental management department, and the subjects had also been informed of the experimental contents and possible risks in detail before participating in the experiment.
The subjects were surrounded by monochromatic LEDs and their PERCLOS values were monitored throughout the whole simulated driving process. Figure 4 describes the PERCLOS value variation under different color illumination. On the whole, it can be found that the PERCLOS model has three stages: Waking stage where the PERCLOS is stable at a low level; Fatigue surge stage where the PERCLOS is rising rapidly; Severe fatigue stage where the PERCLOS is almost maintained at 1 (the max value of PERCLOS). From Fig. 4(a) it can be seen that compared with dark environment (i.e., in the night), in the beginning blue light can effectively reduce the fatigue value of the drivers at night, but soon the PERCLOS value is rebounded rapidly, this leads to the time that the PERCLOS remains at a low level even shorter than in the dark. On the contrary, Fig. 4(b) reveals that red light does not reduce the fatigue prominently like blue light, but it can effectively delay the arrival of severe fatigue, which can therefore extend the safe driving time. Figure 4(c) shows that green light does not make obvious difference from the dark environment, the PERCLOS was only slightly reduced. Figure 4(d) is the situation of half illumination (20 lux) blue light together with half illumination red light. The PERCLOS change tendency is similar to the scenario that only blue lights exist.
Fig. 4. PERCLOS values during one driving process in (a) blue, (b) red and (c) green LEDs, (d) is the same process in red and blue LEDs together.
To ensure the validity of the data, repeated experiments have been conducted. Figure 5(a) shows the awake time (the duration before severe fatigue stage) under different lights. The experiments were performed by 20 volunteers (each color of light was assigned to five volunteers, the final result was the average of them), each volunteer was randomly given a light to conduct the experiment. Before the experiments, volunteers should have the same work and rest time, work intensity and diet during the day. For the experimental data group, the analysis of variance is used to test the data before taking the average value to ensure that the p value of the hypothesis “there is significant difference between the data groups” is greater than 0.05.
Fig. 5. (a) Changes of the awake time in eight experiments. (b) Average values of the PERCLOS in eight experiments.
From the Fig. 5(a), it can be seen that the awake time is indeed the longest with red light, the shortest with blue light, and the green light is close to the dark environment. Figure 5(b) describes the average value of PERCLOS in each experiment. As the data represent, the PERCLOS can be significantly reduced by blue light, and a completely dark environment can result in the highest average.
Then the specific bands of each color light were tested respectively. This was to determine which wavelength range of light was more suitable for adjustment, and more detailed wavelength division also made the designed atmosphere lamp more colorful, which was why the QLEDs with narrow FWHM were chose as light-emitting materials. Each experiment of red and blue lights was performed by five volunteers and repeated after exchanging the volunteers, the results and the standard error of mean (SEM) are shown in Table 1. Blue light with a shorter wavelength than 470 nm is not selected because short wavelength blue light can damage vision [44]. The result shows that the wavelength of blue light should not be longer than 490 nm, because the PERCLOS variation curves of longer ones are closer to green light. While the best wavelength of red light is 620-660 nm, since when the wavelength shorter than 620 nm or longer than 660 nm, the fatigue suppressing effect of red light weakened rapidly.
Table 1. Effect of light wavelength
3.2 Effects of light intensity
In the previous experiment, the LED lamps with an illumination of 40 lux were chosen because the illumination of the road outside the city at night is about 40 lux [45]. If the illumination is too high so that the internal light intensity is greater than outside, then the windshield will produce mirror effect and the image of the objects inside the vehicle will affect the driver's line of sight.
In order to study the influence of illumination on the regulation effect, QLEDs with 0 to 80 lux were used for research. The results are shown in Fig. 6(a) and (b). The red light with higher illumination made one awake longer; stronger blue light had the opposite effect. On the other side the PERCLOS during the whole awake time was rapidly reduced with intensified blue light, while the red light did not cause such a drastic change. Both PERCLOS rebounded at 80 lux, this phenomenon of red light can be attributed to the long awake time, which led to a higher PERCLOS in the later stage and affected the calculation of the average value. As for the blue light, volunteers fed back that the blue light of 80 lux was very dazzling, which affected the eye closure frequency and led to higher PERCLOS. Based on the above experimental data, the upper limit of brightness is set as 40 lux.
Fig. 6. (a) Average values of the PERCLOS in different illuminance (from 0 to 80 lux). (b) Awake time in the same experimental conditions.
3.3 Coordination scheme design of lights
The above experiments determined the influence of light with different wavelengths and illuminations on human fatigue. According to the variation characteristics of driver's PERCLOS value under the influence of light, a set of algorithms were designed, which used different lights in different stages and cooperated with each other to prolong the sober driving time.
In general, a three-stage model of driver’s fatigue change was established. Combined with the characteristics of blue light and red light, it is expected that using red light at first to prolong the waking stage and then using blue light in the fatigue surge stage to reduce the degree of fatigue thus to obtain the longest awake driving time. So, the best switching chance between red light and blue light should be the transition time from the waking stage to the fatigue surge stage. This change reflected in the numerical value is that the PERCLOS curve changes from gentle to rapid rise. In order to deal with this rising trend, following algorithm were designed. Take five numbers in chronological order and use the least square method for first-order fitting. After obtaining the slope, select the next five numbers to fit again. In this way, a graph about the slope was drawn out (Fig. 7(a)). For the point above the zero-scale line, it means that PERCLOS has an upward trend in the first 5 min of this point. But at the same time, a problem also appeared. As a quantization of external physiological signals, PERCLOS has a large noise and there is no good way to filter it. Therefore, it is impossible to judge that the curve will have an upward trend only by one slope greater than zero. Obviously, there is also a conflict between the speed and accuracy of detection. More slopes used to judge the rising trend, more accurate the judgment is, but more time is needed. Therefore, the abnormal state of the driver cannot be found and the light source cannot be adjusted in time. In order to weaken the influence of noise, the sampling interval were broadened, took 10 and 20 min before the current time point as the interval length, and sampled again 5 min after each sampling (Fig. 7(b)). Obviously, the change of slope is much smoother, basically floating up and down around the zero-scale line. But multiple points are still needed to judge the upward trend. The relationship between judgment accuracy and the number of slopes above zero is shown in Table 2 (If the time difference between the deviation values by this method and by the true rising curve was less than 10 min under the interference of noise, it can be considered as a correct judgment, otherwise it was a wrong judgment).
Fig. 7. (a) Slope change polylines of 5 interval sampling and 10 interval sampling. (b) Slope change polylines of 10 interval sampling and double verification. (c) Performance of three different methods in the same PERCLOS change curve. (d) PERCLOS curve of double verification.
Table 2. The relationship between accuracy and the number of slopes
In order to judge the change of fatigue state quickly and accurately, a verification algorithm has been added to the calculated points. The added judgment method is a common trend detection method for time series data called Cox-Stuart. This test method judges whether the trend is caused by a random distribution while analyzing whether the data have an upward trend, which effectively eliminates the previous fluctuation data. When the least square method and Cox-Stuart were applied to the Interval 20 sampling at the same time, the change of fatigue state was accurately calculated by only one point that met the two criteria at the same time. The accuracy of this method (double verification) was 100% in the samples of driver fatigue changes, and from Fig. 7(c) it can be seen that it is also the fastest.
In order to test the effect of the algorithm, a simulated driving experiment was carried out. In the early stage, red light was used to prolong the awake time. After the algorithm detected the change of fatigue state, it was switched to blue light to alleviate the fatigue degree. One of the PERCLOS change curves is shown in Fig. 7(d). The waking stage was significantly prolonged, and fatigue surge stage was unlike the previous straight-line rise, the maximum value reached after many repetitions, and the time to rise to the maximum value was also significantly longer. From the whole simulated driving process, the safe driving time (PERCLOS less than 1) has been significantly extended.
Table 3 records the data from multiple experiments and the SEM of them which confirm the above conclusion. At the same time, they also show the superiority of the double verification method. At last, the number of accidents has been significantly reduced. Due to the difference between simulated driving and the reality, and everyone's driving skill is also different, the drivers were asked for each accident. According to the comprehensive judgment of the subjects’ self-reports and the fatigue values at the time of the accident, excluding the accidents caused by the subjects’ operational errors, the accidents caused by fatigue driving were reduced to zero in the four hours driving.
Table 3. Comparison of various indicators under different algorithms
4. Conclusions
More and more attention has been paid to the safe driving assistance systems, but the current equipment only provides information and cannot really improve the driver's status. In order to solve the deficiencies, the light-emitting devices are combined with the sensing system supplemented by algorithms, to not only monitored the driver's mental state, but also adjusted it in real time, so as to improve the safety of driving at night. The experimental results showed that our intelligent QLED atmosphere lamps prolonged the drivers’ awake time by 1.95 times, and the average value of PERCLOS also decreased by 0.014. In other words, the driver fatigue has been reduced and the duration of sober driving has also been extended at the same time. Finally, a ZYNQ-based safe driving assistant platform was built, which can also expand the laser radar module and automatic driving module in the follow-up work, and has excellent versatility in the field of automobile safety.
Funding
National Natural Science Foundation of China (11974142, 12104178, 12174151, 12204193, 61935009, U21A2068); National Key Research and Development Program of China (2021YFB3500400); Science and Technology Development Program of Jilin Province (20200201084JC, 20200401059GX, 20220101008JC).
Acknowledgments
This work was supported by the National Natural Science Foundation of China (12104178, U21A2068, 11974142, 61935009, 12204193, 12174151), National Key Research and Development Program of China (2021YFB3500400) and Science and Technology Development Program of Jilin Province (20220101008JC, 20200401059GX, 20200201084JC).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. Y. W. Wang, Y. M. Wu, C. Chen, B. H. Wu, S. Ma, D. M. Wang, H. T. Li, and Z. Yang, “Inattentional Blindness in Augmented Reality Head-Up Display-Assisted Driving,” Int. J. Hum.-Comput. Interact. 38(9), 837–850 (2022). [CrossRef]
2. T. B. Liu, S. S. Du, C. C. Liang, B. Zhang, and R. Feng, “A Novel Multi-Sensor Fusion Based Object Detection and Recognition Algorithm for Intelligent Assisted Driving,” IEEE Access 9, 81564–81574 (2021). [CrossRef]
3. C. J. de Naurois, C. Bourdin, A. Stratulat, E. Diaz, and J. L. Vercher, “Detection and prediction of driver drowsiness using artificial neural network models,” Accid. Anal. Prev. 126, 95–104 (2019). [CrossRef]
4. Y. L. Liang, M. L. Reyes, and J. D. Lee, “Real-time detection of driver cognitive distraction using support vector machines,” IEEE Trans. Intell. Transport. Syst. 8(2), 340–350 (2007). [CrossRef]
5. P. M. Forsman, B. J. Vila, R. A. Short, C. G. Mott, and H. P. A. Van Dongen, “Efficient driver drowsiness detection at moderate levels of drowsiness,” Accid. Anal. Prev. 50, 341–350 (2013). [CrossRef]
6. J. J. Wang, S. H. Zhu, and Y. H. Gong, “Driving Safety Monitoring Using Semisupervised Learning on Time Series Data,” IEEE Trans. Intell. Transport. Syst. 11(3), 728–737 (2010). [CrossRef]
7. S. Kar, M. Bhagat, and A. Routray, “EEG signal analysis for the assessment and quantification of driver's fatigue,” Transp. Res. Pt. F-Traffic Psychol. Behav. 13(5), 297–306 (2010). [CrossRef]
8. C. Zhang, H. Wang, and R. R. Fu, “Automated Detection of Driver Fatigue Based on Entropy and Complexity Measures,” IEEE Trans. Intell. Transport. Syst. 15(1), 168–177 (2014). [CrossRef]
9. A. K. Kokonozi, E. M. Michail, I. C. Chouvarda, N. M. Maglaveras, and Ieee, “A Study of Heart Rate and Brain System Complexity and Their Interaction in Sleep-Deprived Subjects,” in 35th Annual Conference on Computers in Cardiology, (2008), 969–971.
10. S. Y. Hu and G. T. Zheng, “Driver drowsiness detection with eyelid related parameters by Support Vector Machine,” Expert Syst. Appl. 36(4), 7651–7658 (2009). [CrossRef]
11. R. J. Lucas, S. N. Peirson, D. M. Berson, T. M. Brown, H. M. Cooper, C. A. Czeisler, M. G. Figueiro, P. D. Gamlin, S. W. Lockley, J. B. O’Hagan, L. L. A. Price, I. Provencio, D. J. Skene, and G. C. Brainard, “Measuring and using light in the melanopsin age,” Trends Neurosci. 37(1), 1–9 (2014). [CrossRef]
12. I. E. Society, “Light and human health: An overview of the impact of optical radiation on visual, circadian, neuroendocrine, and neurobehavioral responses,” in Illuminating Engineering Society Technical memorandum, IES TM-18-08, (2018).
13. D. J. Dijk and M. von Schantz, “Timing and consolidation of human sleep, wakefulness, and performance by a symphony of oscillators,” J. Biol. Rhythms 20(4), 279–290 (2005). [CrossRef]
14. F. Xiong, T. Mao, H. F. Liao, X. Q. Hu, L. Shang, L. Yu, N. N. Lin, L. Huang, Y. M. Yi, R. Zhou, X. Y. Zhou, and J. L. Yi, “Orthokeratology and Low-Intensity Laser Therapy for Slowing the Progression of Myopia in Children,” Biomed Res. Int. 2021, 1–10 (2021). [CrossRef]
15. Y. Zhao, D. K. Xue, J. T. Wang, M. Lu, X. Y. Shen, X. P. Gao, W. W. Yu, and X. Bai, “Smart quantum dot LEDs with simulated solar spectrum for intelligent lighting,” Nanotechnology 31(50), 505207 (2020). [CrossRef]
16. Y. Jiang, Z. Zhu, X. Tan, X. Kong, H. Zhong, J. Zhang, R. Xiong, Y. Yuan, J. Zeng, I. G. Morgan, and M. He, “Effect of Repeated Low-Level Red-Light Therapy for Myopia Control in Children: A Multicenter Randomized Controlled Trial,” Ophthalmology 129(5), 509–519 (2022). [CrossRef]
17. Y. Komada and Y. Inoue, “Chapter 29 - Sleepiness at the Wheel and Countermeasures: Effects of Caffeine, Napping, and Blue Light,” in Modulation of Sleep by Obesity, Diabetes, Age, and Diet, R. R. Watson, ed. (Academic Press, San Diego, 2015), pp. 271–277.
18. J. Taillard, A. Capelli, P. Sagaspe, A. Anund, T. Akerstedt, and P. Philip, “In-Car Nocturnal Blue Light Exposure Improves Motorway Driving: A Randomized Controlled Trial,” PLoS One 7(10), e46750 (2012). [CrossRef]
19. J. Phipps-Nelson, J. R. Redman, L. J. M. Schlangen, and S. M. W. Rajaratnam, “BLUE LIGHT Exposure Reduces Objective Measures of Sleepiness during Prolonged Nighttime Performance Testing,” Chronobiol. Int. 26(5), 891–912 (2009). [CrossRef]
20. B. Rodriguez-Morilla, J. A. Madrid, E. Molina, J. Perez-Navarro, and A. Correa, “Blue-Enriched Light Enhances Alertness but Impairs Accurate Performance in Evening Chronotypes Driving in the Morning,” Front. Psychol. 9, 1 (2018). [CrossRef]
21. L. L. Fitri, C. Chusnia, and Suprijanto, “Blue Light Exposure and Visual Distraction Effect of The Cognition Brain Waves of Driver in Car Simulator at Night,” in 2nd International Conference on Biosciences (ICoBio) - From Genes to Ecosystems, IOP Conference Series-Earth and Environmental Science (2017),
22. A. Lerchl, C. Schindler, K. Eichhorn, F. Kley, and T. C. Erren, “Indirect blue light does not suppress nocturnal salivary melatonin in humans in an automobile setting,” J. Pineal Res. 47(2), 143–146 (2009). [CrossRef]
23. A. Dasgupta, A. George, S. L. Happy, and A. Routray, “A Vision-Based System for Monitoring the Loss of Attention in Automotive Drivers,” IEEE Trans. Intell. Transport. Syst. 14(4), 1825–1838 (2013). [CrossRef]
24. L. M. Bergasa, J. Nuevo, M. A. Sotelo, R. Barea, and M. E. Lopez, “Real-time system for monitoring driver vigilance,” IEEE Trans. Intell. Transport. Syst. 7(1), 63–77 (2006). [CrossRef]
25. B. Mandal, L. Y. Li, G. S. Wang, and J. Lin, “Towards Detection of Bus Driver Fatigue Based on Robust Visual Analysis of Eye State,” IEEE Trans. Intell. Transport. Syst. 18(3), 545–557 (2017). [CrossRef]
26. C. J. Xu, M. G. Ou, H. Zhou, and C. L. Yang, “Preparation and properties of bifunctional Gd2O3/GQD composite nanoparticles,” J. Rare Earths 40(7), 1118–1125 (2022). [CrossRef]
27. Z. H. Wang, Z. C. Wen, F. Tang, K. Z. Tian, C. D. Zuo, and L. Z. Zhang, “Role of localized phonon vibration in luminescence performance of Pr doped Ba(Mg0.28Zr0.16Ta0.56)O3 transparent ceramics,” J. Rare Earths 40(11), 1728–1733 (2022). [CrossRef]
28. D. K. Xue, C. Ruan, Y. Zhang, H. B. Chen, X. B. Chen, C. F. Wu, C. T. Zheng, H. D. Chen, and W. W. Yu, “Enhanced bandwidth of white light communication using nanomaterial phosphors,” Nanotechnology 29(45), 455708 (2018). [CrossRef]
29. Y. H. Won, O. Cho, T. Kim, D. Y. Chung, T. Kim, H. Chung, H. Jang, J. Lee, D. Kim, and E. Jang, “Highly efficient and stable InP/ZnSe/ZnS quantum dot light-emitting diodes,” Nature 575(7784), 634–638 (2019). [CrossRef]
30. Y. Y. Wang, D. Yue, B. S. Tian, W. D. Li, B. Zhang, L. Yang, Y. Li, Z. L. Wang, and Y. L. Zhang, “Effectively enhanced photoluminescence of CePO4:Tb3 + nanorods combined with carbon dots*,” J. Rare Earths 40(7), 1007–1013 (2022). [CrossRef]
31. L. Shihua, B. Deng, J. Yang, J. H. Liu, J. Chen, F. H. Zeng, H. Y. Wang, R. J. Yu, and G. G. Zhang, “Multi-site occupancies and luminescence properties of cyan-emitting Ca9-xNaGd2/3(PO4)(7):Eu2 + phosphors for white light-emitting diodes,” J. Rare Earths 40(2), 243–252 (2022). [CrossRef]
32. D. S. Lee, T. W. Chong, and B. G. Lee, “Stress Events Detection of Driver by Wearable Glove System,” IEEE Sens. J. 17, 194–204 (2017). [CrossRef]
33. E. R. Gaillard, L. Zheng, J. C. Merriam, and J. Dillon, “Age-related changes in the absorption characteristics of the primate lens,” Invest. Ophthalmol. Vis. Sci.41, 1454–1459 (2000).
34. G. C. Brainard, J. P. Hanifin, J. M. Greeson, B. Byrne, G. Glickman, E. Gerner, and M. D. Rollag, “Action spectrum for melatonin regulation in humans: Evidence for a novel circadian photoreceptor,” J. Neurosci. 21(16), 6405–6412 (2001). [CrossRef]
35. A. M. Chang, F. Scheer, C. A. Czeisler, and D. Aeschbach, “Direct Effects of Light on Alertness, Vigilance, and the Waking Electroencephalogram in Humans Depend on Prior Light History,” Sleep 36(8), 1239–1246 (2013). [CrossRef]
36. M. Ruger, M. C. M. Gordijn, D. G. M. Beersma, B. De Vries, and S. Daan, “Weak relationships between suppression of melatonin and suppression of sleepiness/fatigue in response to light exposure,” J. Sleep Res. 14(3), 221–227 (2005). [CrossRef]
37. L. Shi, T. Katsuura, Y. Shimomura, and K. Iwanaga, “Effects of Different Light Source Color Temperatures during Physical Exercise on Human EEG and Subjective Evaluation,” J. Hum. Envir. Syst. 12(1), 27–34 (2009). [CrossRef]
38. S. A. Rahman, M. A. St Hilaire, and S. W. Lockley, “The effects of spectral tuning of evening ambient light on melatonin suppression, alertness and sleep,” Physiol. Behav. 177, 221–229 (2017). [CrossRef]
39. A. Y. Segal, T. L. Sletten, E. E. Flynn-Evans, S. W. Lockley, and S. M. W. Rajaratnam, “Daytime Exposure to Short- and Medium-Wavelength Light Did Not Improve Alertness and Neurobehavioral Performance,” J. Biol. Rhythms 31(5), 470–482 (2016). [CrossRef]
40. C. Cajochen, S. Frey, D. Anders, J. Spati, M. Bues, A. Pross, R. Mager, A. Wirz-Justice, and O. Stefani, “Evening exposure to a light-emitting diodes (LED)-backlit computer screen affects circadian physiology and cognitive performance,” J. Appl. Physiol. 110(5), 1432–1438 (2011). [CrossRef]
41. J. Cho, Y. K. Jung, J. K. Lee, and H. S. Jung, “Highly efficient Blue-Emitting CdSe-derived Core/Shell Gradient Alloy Quantum Dots with Improved Photoluminescent Quantum Yield and Enhanced Photostability,” Langmuir 33(15), 3711–3719 (2017). [CrossRef]
42. O. Chen, J. Zhao, V. P. Chauhan, J. Cui, C. Wong, D. K. Harris, H. Wei, H. S. Han, D. Fukumura, R. K. Jain, and M. G. Bawendi, “Compact high-quality CdSe-CdS core-shell nanocrystals with narrow emission linewidths and suppressed blinking,” Nat. Mater. 12(5), 445–451 (2013). [CrossRef]
43. G. Wang, S. J. Yu, B. Liu, Z. M. Liu, and J. K. Li, “Nanocomposites of CsPbBr3 perovskite quantum dots embedded in Gd2O3:Eu3 + hollow spheres for LEDs application,” J. Rare Earths 40(10), 1509–1518 (2022). [CrossRef]
44. J. X. Tao, W. C. Zhou, and X. G. Zhu, “Mitochondria as Potential Targets and Initiators of the Blue Light Hazard to the Retina,” Oxid. Med. Cell. Longev. 2019, 6435364 (2019). [CrossRef]
45. M. A. A. bin Ramli, M. R. Uddin, S. J. Tan, E. M. M. Zahran, Y. H. Yap, and Ieee, “Development of a Road Illumination Measurement System,” in 4th International Conference on Intelligent Transportation Engineering (ICITE), (2019), 296–300.
