Floating mobile data pilot in the Helsinki Metropolitan Area: Validation of travel time data

The floating mobile data (FMD) pilot aimed to provide anonymous travel time information from data provided by mobile phones. The main purpose of the pilot was to validate the quality of FMD by comparing it with camera-based travel time data. Additionally, the usability of data was assessed from a traffic management point of view and some principles were tested. A travel time camera system was used as a reference. The main conclusion is that for traffic management operations, monitoring of cell handovers of active 2G phone calls does not produce a high enough number of observations. Consequently, it is recommended to select an FMD technology that is able to monitor a larger proportion of mobile phones per aggregation period per link. Independence of active phone calls would help to cover also night time traffic as well as smaller roads and main streets. Another main conclusion is that the estimation of median travel time should be developed further. Although traffic-wise homogeneous links were targeted, obviously traffic on many links is dynamic in both space and time. Consequently, it is recommended that the estimation of traffic flow status be based on part-observations, and that the weight of different zones of the travel time link be balanced. If full observations on fluent parts of a link are allowed to dominate, the estimate will be biased. The main implication related to ad hoc service is that the precision of ad hoc links was not high enough. The precision does not have to be as high as for static links, but nonetheless high enough to make validation of the data possible. Thus our recommendation is to improve it. Another recommendation is to increase the number of observations in ad hoc service to suffice for real time operations based on 5-minute medians.