A Swarm-based AI Aided Wheel Bot System to Detect Cracks in Railway Tracks

Abstract

Railway is one of our country’s most important sources of transit, yet it is a source of great concern since our railway tracks are vulnerable to damage/cracks. The major causes of railway accidents are railway track crossings and unseen faults in railway tracks. Due to such primitive conditions a numerous number of accidents are seen every year with heavy tolls to human life and infrastructure. The aim of this project is to develop a swarm robotic system to provide a real-time solution to crack detection and danger mitigation. Crack detection has been manual which is time-consuming and tedious, until recently, when the crack detection system started to become automated. But the system still has major flaws such as inability in identifying the different types of cracks, distinguishing between intentional and actual cracks, and providing a real-time solution to the identified problem. The proposed system is aimed towards efficient detection of not just cracks, but obstacles and animal detection along with fault detection in case of agent failure. A swarm robotic approach is implemented where the agents are deployed on the tracks and they are able to communicate among themselves as well as the base station in real-time. The system is integrated with an android application which updates the status of the agents of the assigned cluster in real-time. The agent also has the mechanism of detaching itself off the tracks when a train approaches to avoid collisions.