Kein Folientitel -

Kein Folientitel -

Institute of Integrated Sensor Systems Department of Electrical and Computer Engineering Multi-Sensor Soft-Computing System for Driver Drowsiness Detection Li Li, Klaudius Werber, Carlos F. Calvillo, Khac Dong Dinh, Ander Guarde and Andreas Knig 10-Dec-2012

Introduction Driving Scene Modeling and Hardware Setup Software Components and Algorithms Experimental Results Conclusion and Future Work 1 Li Li [WSC17] Introduction Major factor in 20 percent of all accidents in the United States in 2006 The second most frequent cause of serious truck accidents on German highways

Major damage caused by drowsy truck or bus drivers Enhance active safety with advanced driver assistance 2 Li Li [WSC17] Hardware Setup DeCaDrive System Multi-sensing interfaces Depth camera

Steering angle sensor Pulse rate sensor PC-based softcomputing subsystem PC-based driving simulator 3 Li Li [WSC17] Hardware Setup SoA depth camera +

+ + + + Extention of 2D image with distance Wide field of view Relatively low computational cost Robust to lighting variations (active sensing) Non-intrusive and non-obstructive (eye-safe NIR light source) Microsoft Kinect PMD CamCube 4

SoftKinetic DepthSense Li Li [WSC17] Hardware Setup Pulse rate sensor Steering angle sensor Heart health and fitness Time domain analysis Frequency domain analysis Steering behavior of driver Correlation with driver

status and driver intention embedded 5 Li Li [WSC17] Software Components and Algorithms Overview of the data processing flow 6 Li Li [WSC17]

Feature Computation Features being computed from multiple sensor measurements 7 Li Li [WSC17] Experimental Results Test subjects Five male test subjects

22 to 25 years old (mean: 23.6, std:1.1) All have drivers license for at least 4 years No alcohol drinking before test Experiments One hour driving simulation for each test subject 588-minute driving sequence recorded Ground truth: not drowsy, a little drowsy, deep drowsy Through self-rated score and response time

8 Li Li [WSC17] Experimental Results Examples of different sensor features blink frequency low steering percentage mean pulse rate 9

Li Li [WSC17] Experimental Results Screenshot of online processing of various sensor data Depth image 10 Eye pupil and corners Li Li [WSC17] Experimental Results

Results of ANN based classifier with two training algorithms 11 Li Li [WSC17] Experimental Results Confusion matrix of SCG Confusion matrix of LM 40 hidden neurons 10-fold cross-validation 80 hidden neurons

10-fold cross-validation 12 Li Li [WSC17] Experimental Results Drowsiness level classification accuracy depending on selected features 13 Li Li [WSC17] Conclusion and Future Work

Contribution Emerging framework for driver status monitoring and intention detection with multi-sensor soft-computing system Classification of three different drowsiness levels with up to 98.9% accuracy based on data sets of ve test subjects. Future work Validation with more statistics and with data from real vehicles Variance compensation by adaptive learning

Optimization of feature selection with sophisticated heuristics Utilization of other advanced classification techniques, e.g., SVM Integration of more embedded sensors with wireless technology 14 Li Li [WSC17] Thank you! 15 Li Li [WSC17]

Recently Viewed Presentations

  • Visually Impaired Students with Vestibular Issues

    Visually Impaired Students with Vestibular Issues

    Vestibular organs Located in the bony chambers of the skull in the inner ear 3 semi-circular canals positioned in different planes for rotational movements Otoliths (Utricle and Saccule) for linear accelerations Semicircular Canals Bulge at base of canals - ampulla...
  • Transition from EU Centres to Jean Monnet EU

    Transition from EU Centres to Jean Monnet EU

    If you have already one EULogin account, you may use it, otherwise you must create an EULogin account. Applicant organisation (and for the Jean Monnet Networks also the partner institutions involved) must be registered in the Participant Portal and receive...


    BUILD OR ENHANCE YOUR RECURING REVENUE STREAM. OWS is designed with a dedicated reseller portal to allow you to control customer access to features and functionality based on your RMR business model. OWS even includes features to enable time based...
  • 2 - De Anza College

    2 - De Anza College

    Nine Pearly Peaks - Huang Gong Wang AD 1269 - 1354 #2 Chinese Horse, Lascaux caves France 17,000 ya , #3 Last Supper Leonardo da Vinci, 15th century mural #4 Maria Sybilla Merian Watercolor, 1705 #5. Greek ketos jar- 7000...
  • From Gatekeeping to Engagement

    From Gatekeeping to Engagement

    From Gatekeeping to Engagement: A Multicontextual, Mixed Method Study of Student Academic Engagement in Introductory STEM Courses. Association for Institutional Research . Annual Forum. Toronto. May 22, 2011. Jo Gasiewski. Kevin Eagan. Gina Garcia. Sylvia Hurtado. Mitchell Chang. Higher Education...
  • Welcome ISLN Meeting November 14, 2014 Agenda  Welcome

    Welcome ISLN Meeting November 14, 2014 Agenda Welcome

    SAMPLE VIDEO What defensible evidence is this teacher collecting about student learning? What evidence is the observing principal collecting about student learning? Table Talk Mid-Year Conferencing Student Voice Jeffrey Coles, KDE Program Reviews Phil Shepherd, KDE Than k you for...
  • What the tapestry programme is What an observation should ...

    What the tapestry programme is What an observation should ...

    Teachers and TAs have iPads in the classroom with the EExAT app on them. Each member of staff has a unique login to enable them to access the pupils. When an adult observes your child making a step in their...
  • Foam Cube Challenge Leader board after day 3

    Foam Cube Challenge Leader board after day 3 Foam Cube Challenge. Leader board after day 3 of the 2015 Big Bang Fair