Challenges in Implementing Crowdsourcing on automatic real-time transit tracking system

Implementing Crowdsourcing in arrival time prediction system for public transportation lets people share their pieces of information that they have, such as the transport location or situation, and combine the information together to get the complete picture of the interesting public transportation. However, the system that requires human to input the information has critical problems, which are the needed of the user effort and the fraud by false reporting. So the automatic system has been proposed to solve those problems by detecting necessary information automatically. But still there are some new challenges coming with it. In this paper, we aim to demonstrate those new challenges in automatic crowdsourcing system, which are Performance, Consistency, and Continuity of Tracking and also introduce some possible solutions for those challenges. We tested our solutions by simulating the problem situations with our boat arrival time prediction system, which is an automatic arrival time prediction system for Chao Phraya Express Boat in Bangkok, Thailand.