Traffic Time-Intervals Split Based on the Traffic Flow Variation of Days of Week

A majority of urban pretimed traffic signal control systems deploy multiple signal time-intervals to account for traffic demand changes during a day. There have been some mature studies on the algorithms of time-intervals split of TOD. However, these algorithms neglected the variation of traffic flow among days of week. In this research, a Nonparametric Test is applied to reflect the variation of traffic flow among days of week and classify the days of week into different categories, and then each day is divided into several timeintervals by adopting an ordered sample cluster algorithm according to the former categories. This method could eliminate the impact of traffic flow uncertainty on the timeintervals split and will provide more accurate time-intervals split for the urban pretimed traffic signal control systems.