No abstract

No abstract

Detection and identification of the driving behavior is an issue that has attention broadly in the study of the intelligent automotive systems. This research study presents the detection of the risk of drowsiness and distraction while driving. If his or her face not in the right direction when driving for more than 2 seconds, then alert to the driver depend on the detected risk level. From two reasons mentioned above. The system can be divided into three parts: the first part consists of the normalization of the image size to optimize the system performance and improve image quality by adjusting illumination using Histogram Equalization, the second part is procedural to detect the eyes and nose, then create a risk feature name as "Feature of Driver Risk (FODR)" to know the possible direction of the faces with HaarLike Feature, the third part is procedural of data classification. In addition, calculation of risky for alert by used BPNN and Fuzzy Logic. This study uses a mobile phone camera by shooting in front of the driver during day time by 5 people with 6000 frames for each person. The study found that, the accuracy in calculating the risk was 78.43 and 87.12 percent, respectively.

No abstract

No abstract