Traffic accidents can be occurred because the government does not immediately repair damaged road. They can be imprisoned or fined according to Undang-undang Republik Indonesia Nomor 22 Tahun 2009 Tentang Lalu Lintas dan Angkutan Jalan. It is necessary to have an automated method to detect damaged road because manually checking the road condition is not a practical option. Studies have demonstrated that vibration technique has potential to detect damaged road. This research intends to explore the potential use of artificial neural network for detecting road anomalies on the basis of accelerometer data and measures its performance. Smart-phone which was enriched with a 3D accelerometer sensor and geo-location sensor will be put in a vehicle. The vehicle then will cross several road conditions: Normal, Pothole, Speedbump, and Expansion Joint. Five features that used in this study are: maximum acceleration in x and z axis, and dominant frequency of acceleration in x, y, and z axis.

Damaged Road, Vibration, Accelerometer, Smart-phone, Artificial Neural Network

• Gradiyanto Sanjaya, S.Kom, M.TI

• Yudy Purnama, S.Kom, M.TI