Communication Assisted Dynamic Scheduling of Public Transportation Systems

Abstract

In the developing countries, traffic and congestion on the roads are likely to be seen. Mostly the congested road deteriorates itself rapidly without proper maintenance. Moreover, the capacity of vehicles are also exceeded by the load capacity of road, leading to potholes and bumps and roughness. On the contrary, this also leads to bad driving behavior, which affects the safety of commuter and arrival time of public transportation systems. The efficient way to detect these anomalies is to collect the data from inbuilt sensors of smartphone. The data collected from the smartphone were normalized and analyzed to detect the events where the “Smart-Patrolling” prototype able to find potholes and bumps with the accuracy of 88.66% and 88.89% respectively. Driving behavior of driver was detected by observing the braking patterns and aggressive lateral maneuver, where the proposed algorithm was able to detect with an accuracy of 100% (harsh braking) and 97% (normal left/right turns) & 86.67% (aggressive left/right turns). Lastly, the arrival time of public buses has been predicted where the regression model produces better results when compared with other prediction models.