Abnormal Driving Detection Based on Accelerometer and Gyroscope Sensor on Smartphone using Artificial Neural Network (ANN) Algorithm

Transportation has become one of the needs of human life. One of the most used transportation in Indonesia is a motorcycle. This transportation is reliable, has a wide variety of uses, is capable of handling traffic jams, and is very economical. But motorbikes have a relatively high risk for the safety of the rider compared to cars. With driver behavior and road congestion conditions, accidents often occur. In this research, we observed the movement of the driver through a smartphone and built an application that was able to provide a warning to motorists. In this system we use 2 types of sensors, Accelerometer and Gyroscope sensors which are generally the basic features of a smartphone. We adopt a Machine Learning-based movement identification process with an Artificial Neural Network (ANN) algorithm. ANN will do what movements it does based on the data obtained from the Accelerometer and Gyroscope sensor values. The application is installed on a smartphone and put it on a vehicle with a fixed position. The device will calculate the activity based on several predetermined categories of the driver status. The status categories are normal, zig-zag, sleepy, turn right, turn left, U-turn, sudden braking, sudden acceleration, and speed bumps. The level of accuracy generated by Machine Learning using the Artificial Neural Network Algorithm in this research was 96,2%.