Discussion on the dynamic travel time estimation based on multiple data sources - combining emerging data collection devices with conventional detectors

摘要

Accurate travel time estimation is considered as a key ingredient in the advanced traveler information systems (ATIS). Loop detector is conventionally the predominant on road data collector and now the probe vehicle system is also getting highlighted. Probe data can be either combined with loop data or used exclusively to estimate travel time. In both of the two cases, the estimation precision depends on the percentage of probe vehicles, and also the traffic condition. Through an urban expressway simulation study, the relationships between estimation accuracies and probe percentages in various traffic situations are discussed here. Different data fusion techniques are used to combine probe data with loop data to estimate travel time. It is found that combining even very limited probe data with loop data can also tremendously improve the overall travel time estimation accuracy; and the simplest data fusion technique happens to be also the most accurate one.