Abstract:In the case of peak spreading, the traditional static analysis on the duration of congestion using peak 15 minutes of hour has lost its meaning. To deal with congestion, we need dynamic analysis on the onset and dissipation of congestion. This paper aims to reflect the severity of congestion, even the spatial interaction of congestion, by the relative size of peak period. Two thresholds, the minimum value in the peak hour and 95 percent of peak hour volume respectively, were chosen to define the peak using archived stop-bar detector data. The thresholds and the time-granularity of the data were cross-compared to choose appropriate threshold and data time interval, and the result is 95 percent of peak hour volume under 10min interval data. Then the measures of duration of peak period, including the length of peak period, the beginning and ending time of peak period, were calculated for different signalized intersections inlet approaches. Further, the measures of peak period of different intersections in the same direction of the same radial road were presented to find out the commute traffic patterns. Lastly, the spatial-temporal association pattern of the measures of peak period of different intersections in downtown Shanghai was performed by ArcGIS.

Spatial-Temporal Association of Duration of Peak Period at Signalized Intersections