The effects of advanced traveller information system (ATIS) accuracy and the extent of ATIS roadway instrumentation on the on-time reliability benefits to routine users of ATISs were evaluated by using archived estimates of roadway travel times to re-create hypothetical, retrospective paired driving trials between travellers with and without ATISs. Previous research using this technique demonstrated that travellers who received notification of congestion before departure could realize time management benefits from improved on-time reliability and trip predictability. On the basis of millions of hypothetical trips over a 12-month period in Los Angeles, California, it was found that a net benefit to an average user of ATIS existed if the error in travel time estimation was in the range of 14% to 21% or better, a threshold that depended on regional day-to-day travel time variability. For less accurate ATIS services, only certain subsets of the driving population, such as those with relatively long or highly variable trips, may realize any benefit. Furthermore, it was observed that a nearly optimal geographical deployment of ATIS could garner as much as 56% of the benefit of full coverage from the first 30% of deployment. Yet, the best deployment strategy is not as simple as prioritising links on the basis of day-to-day variability. In making cost-effective trade-off decisions about how to invest in improved ATISs, be it expanding geographic coverage or improving accuracy, the findings underscore the importance of understanding the accuracy required to generate ATIS user benefits on the basis of regional day-to-day roadway variability.
Abstract