IV. Conclusion

To provide real-time traffic information at low cost, the present Floating Car Data system with virtual trip line (VTL) that exploits the proliferation of smart-phones faces a few challenges such as the privacy protection of the probe vehicle, the optimal length of the VTL line, and minimizing huge error in the geo-location and velocity data. This work has addressed some of those issues by deploying a few simple prediction methods. The work leads us to conclusions:  a VTL length of 26 m can provide 95% chance that the probe vehicle will cross the line, the delay time can be maintained below 1~min for 80% of the cases, and the prediction of the velocity and position can be improved significantly by applying the weighted moving average method.