Establishing Random Forest Model Based on Visual Variables to Detect Driving Fatigue

  • Chengxi MaEmail author
  • Yonggang Wang
  • Xu Chang
  • Yanhui Li
  • Hao Zhu
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 503)


Driving fatigue is a major cause of traffic accidents, and how to detect whether the driver is fatigue has become an urgent issue to be solved, therefore, propose a method of driving fatigue recognition based on visual behavior variables. In order to improve the accuracy of drowsiness detection, 64 professional drivers were recruited to participant in simulated driving experiment, and collect the data of driver’s visual variables and the degree of fatigue. Then the random forest algorithm trains 200 samples of 256 samples and tests the 56 samples data remaining, and compares the predictive state and the real state of the driver. The commonly used evaluation index and ROC curves were used to evaluate the forecasting accuracy and performance of random forest algorithm. These two indicators indicate the performance of the model is excellent. At the same time, the random forest algorithm determines the importance of the four groups of indexes in the evaluation model, the blink time was of the highest importance and the pupil diameter was the lowest.


Driving sleepiness Visual variables Random forest algorithm Driving sleepiness assessment Simulated driving test 



The authors wish to acknowledge the financial support of the National Natural Science Foundation of China (51208051), Natural Science Basic Research Plan in Shaanxi Province of China (2016JM5036), and the Special Fund for the Basic Scientific Research of Central Colleges, Chang’an University (310821172005).


  1. 1.
    Thiffault P, Bergeron J (2003) Monotony of road environment and driver fatigue: a simulator study. Accid Anal Prev 35(3):381CrossRefGoogle Scholar
  2. 2.
    Radun I, Ohisalo J, Radun J et al (2013) Driver fatigue and the law from the perspective of police officers and prosecutors. Transp Res Part F Traffic Psychol Behav 18(5):159–167CrossRefGoogle Scholar
  3. 3.
    Yang G, Lin Y, Bhattacharya P (2010) A driver fatigue recognition model based on information fusion and dynamic Bayesian network. Informationences 180(10):1942–1954Google Scholar
  4. 4.
    Mattsson K (2007) In-vehicle prediction of truck driver sleepiness: lane position variables. Luleå Tekniska UniversitetGoogle Scholar
  5. 5.
    Zhao C, Zhao M, Liu J et al (2012) Electroencephalogram and electrocardiograph assessment of mental fatigue in a driving simulator. Accid Anal Prev 45(1):83–90CrossRefGoogle Scholar
  6. 6.
    Patel M, Lal SKL, Kavanagh D et al (2011) Applying neural network analysis on heart rate variability data to assess driver fatigue. Expert Syst Appl 38(6):7235–7242CrossRefGoogle Scholar
  7. 7.
    Xu C, Wang X, Cheng X et al (2015) Driver drowsiness level analysis and prediction based on decision tree. J Tongji Univ (Nat Sci) 43(1):75–81Google Scholar
  8. 8.
    Horng WB, Chen CY, Chang Y et al (2004) Driver fatigue detection based on eye tracking and dynamk, template matching. In: Danesy D (ed) Proceedings of the Second Workshop on the Atmospheric Chemistry Validation of Envisat, vol 1, pp 7–12Google Scholar
  9. 9.
    Pang X, Li J, Xu X (2011) Threshold value of indices of eye states to monitor drive fatigue. J Tongji Univ (Nat Sci) 39(12):1811–1815Google Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2019

Authors and Affiliations

  • Chengxi Ma
    • 1
    Email author
  • Yonggang Wang
    • 1
  • Xu Chang
    • 1
  • Yanhui Li
    • 1
  • Hao Zhu
    • 1
  1. 1.School of HighwayChang’an UniversityXi’anChina

Personalised recommendations