The k-means method, a nonhierarchical clustering algorithm, is applied to traffic volume data to determine time-of-day (TOD) breakpoints for traffic signal timing plans. Other methods, including hierarchical clustering techniques, have been applied to traffic signal timings, but they are computationally intensive and require a substantial amount of data storage space. The procedure requires that the analyst specify the number of clusters prior to engaging the algorithm. The resultant allocations of volumes to clusters may be "noisy"; smoothing may be needed to avoid having an inoperable number of TOD breakpoints. The algorithm is applied to a small case study involving a two-intersection corridor and just under four hours of volume data. Three time intervals were identified, with a peak, two-hour period sandwiched by two off-peak segments. An expanded application of the algorithm on a longer corridor or network, over a longer time period, is recommended. Subsequent steps would be to develop the signal timing plans for the study intersections, evaluate the proposed plans, and assess the potential for their implementation. The k-means method can develop TOD breakpoints from traffic volumes, making it a potentially useful procedure where detectors supplying additional traffic information are either sparse or nonexistent.
Using k-means clustering to identify time-of-day break points for traffic signal timing plans
2005-01-01
203955 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Using K-Means Clustering to Identify Time-Of-Day Break Points for Traffic Signal Timing Plans
British Library Conference Proceedings | 2005
|Generating More Equitable Traffic Signal Timing Plans
Online Contents | 2010
|Generating More Equitable Traffic Signal Timing Plans
Transportation Research Record | 2010
|Using connected vehicle data to optimize traffic signal timing plans
Europäisches Patentamt | 2020
|