Method for Identifying Homogeneous Periods to Evaluate Urban Travel Time Reliability

Travel time reliability influences drivers’ decisions in urban trips as they adapt to uncertainty levels of congestion. While measures of central tendency have traditionally characterized travel time reliability, recent research has expanded to consider alternative metrics for describing this feature of traffic flow. This paper proposes a method to determine time slots of homogeneous periods to assess travel time reliability with a case study on arterial streets of Belo Horizonte, Brazil. The method comprises an initial data collection stage through an application programming interface (API) on selected street segments and a second stage to group periods into hourly time slots with homogeneous travel time distributions to calculate reliability metrics identified in an extensive literature review. The results refer to travel time reliability on streets in Belo Horizonte during different times of the day and days of the week, highlighting low reliability between 5 and 10 a.m. on Monday and Tuesday from the neighborhood to the downtown direction and between 4 and 8 p.m. on Wednesday and Friday from the downtown to the neighborhood direction.