Boat arrival time prediction system using mobile phone sensing

Abstract

Boat is one of the most popular vehicles for local people and tourists to travel in Bangkok, Thailand, since there is no traffic jam. Without accurate arrival time, however, boat becomes an inconvenient transportation since the waiting time can be varied from 10 minutes to an hour. In this paper, we propose a novel boat arrival time prediction system by using passengers' mobile phones. The system can work without installing any infrastructure on boat or pier. It does not require cooperation with any boat company. This would make the system almost costless and easy to implement. Moreover, the system does not require any manual input from the passenger, so this means it can automatically predict the arrival time. We take advantages of the boat environment and design the system to use only energy-efficient sensors, including Wi-Fi, accelerometer, and microphone. No energy hungry sensor such as GPS is used. The system has been tested with one of the most crowded boat transit, Chao Phraya Express Boat, in Bangkok, Thailand. The result shows that our system can predict boat arrival time accurately. Specifically, the prediction error is less than one minute per pier.