Analysis of travel time reliability on Indiana interstates.

685138187PDFTech ReportTravel timeMathematical modelsAnalysisIndianaNEXTRANS Center (U.S.)Martchouk, MariaMannering, FredPurdue University.United States. Dept. of Transportation. Research and Innovative Technology AdministrationNTL-REFERENCES AND DIRECTORIES-StatisticsNTL-PLANNING AND POLICY-PLANNING AND POLICYUS Transportation CollectionTravel-time reliability is a key performance measure in any transportation system. It is ameasure of quality of travel time experienced by transportation system users and reflects the efficiencyof the transportation system to serve citizens, businesses and visitors. Travel-time reliability (thevariability in travel times on the same route at the same time from one day to the next) is critical totravelers, shippers, receivers and carriers for trip decisions and on-time arrivals at destinations. Thus,understanding travel time reliability is important to travelers in order for them to plan their tripseffectively as well as shippers for them to plan and select routes appropriately.The first objective of this study is to formulate a methodology to obtain travel time data usingBluetooth technology on a freeway segment and collect travel time data. Bluetooth technology enablesto collect real travel time data with a high sampling rate of up to 10% of the traffic flow. It alsoeliminates that need to use complex and often inaccurate algorithms use to calculate travel time frompoint speed data. Another benefit of this technology is it\ue2\u20ac\u2122s the fact that it is relatively inexpensive toimplement; every station where travel time is desirable needs to be equipped with a processing unit,power source and Bluetooth dongle.The second objective is to observe daily and inter-daily variations as well as those due to poorweather conditions and estimate econometric models to predict travel time and variability. Within thecontext of this study, travel times were collected for two sections of freeway that experience heavycongestion during the peak hours. These travel times were then used to estimate three econometricmodels that predict travel time as a function of traffic flow parameters including speed and volume. Thefirst model, which is linear regression with lagged dependent variable terms, aims to predict individualtravel times during all times of day. The second model, a survival model, seeks to evaluate theprobability of the trip lasting any specified length of time. In addition, it can predict the probability ofexiting the freeway segment given that the vehicle has been traversing the segment up to that time. Thethird model seeks to describe travel time and the variability of travel time using the seemingly unrelatedregression equations