Frequency of Probe Reports and Variance of Travel Time Estimates

Hybrid Vehicle-Centric Route Guidance 1

“…Sen et al [53,54] notes that even in route guidance systems that rely on dynamic reports of arc travel times, the ability to produce good estimates when dynamic reporting is unreliable or unavailable significantly affects the performance of the system. Static estimates may be required at system startup when not enough probe vehicles exist or in situations of equipment malfunction.…”

“…Static estimates may be required at system startup when not enough probe vehicles exist or in situations of equipment malfunction. Thus, [53] proposes a method for generating such estimates and [54] provides results of a real-world test of these estimation methods.…”

Assessing Expected Accuracy of Probe Vehicle Travel Time Reports

“…Furthermore, the results from Sen et al (1997a) discussed earlier in this paper, are based on field data in which all probe vehicles traversed a set of links by using the same link entry and link exit movements. Thus, the sample used in their analysis has a level of bias that is similar to that associated with the examples provided in this paper.…”

“…Specifically, the work of Sen et al (1997a) can be seen as applicable to scenarios in which the probe reports constitute a biased sample of the population. Conversely, the work of Van Aerde et al (1993) is applicable when the probe reports are an unbiased sample of the population.…”

“…The same assumption was used by other researchers in determining the required level of market penetration or number of probe vehicles (Srinivasan and Jovanis, 1996;Turner and Holder, 1995;Boyce et al, 1991aBoyce et al, , 1991b) Sen et al (1997aSen et al ( , 1997b examined field data collected from probe vehicles as part of the ADVANCE project. On the basis of a statistical analysis of probe link travel times, they found that probe reports are not independent and therefore regardless of the sample size, the sample mean does not approach the population mean.…”