Autonomous mobile robot development based on duckietown platform for recognizing and following the traffic sign

Abstract

This paper presents development of an autonomous mobile robot which can recognize the traffic sign and move according to that recognized sign. The mobile robot was developed based on Duckietown platform, which is a research which focuses on self-driving vehicles. The mobile robot was equipped with a fish eye camera as its vision device and raspberry pi.3 controller module as the main controller for doing image processing in order to recognize the signs and controlling the movement of the robot. The program runs on ROS platform. The system has been tested by moving mobile robot through all area and the results showed average recognition rate of 80%. The system also has been tested in variation of light intensity and the result showed average recognition rate of 80%. Another test also was done by adding disturbance object which has similar or same color as sign color and the result showed average recognition rate of 43.3%.