Anti-bump: A bump/pothole monitoring and broadcasting system for driver awareness

Abstract

This paper presents a system for bump detection and alarming system for drivers. We have presented an architecture that adopts context awareness and Bump location broadcasting to detect and save bumps locations. This system uses motion sensor to get the readings of the bump then we classify it using Dynamic Time Wrapping, Hidden Markov Model and Neural Network. We keep records for the bump location through tracking its geographic position. We developed a system that alarms the driver within appropriate profiled distance for bump occurrence. We conducted two experiments for testing the system in a street modeled architect with different kinds of bumps and potholes. The other experiment was on real street bumps. The results show that the system can detect bumps and potholes with reasonably accepted accuracy. © 2013 Springer-Verlag.