Analysis of optimal temporal aggregation interval of traffic flow data for urban road traffic monitoring

作者：Lu Zhenbo; Xia Jingxin^*; Jiao Tiantian; Shi Xingfei; Huang Wei

来源：Journal of Southeast University(Natural Science Edition), 2012, 42(5): 1000-1005.

DOI：10.3969/j.issn.1001-0505.2012.05.036

• Hypothesis tests
• Inflection points
• Operating state
• Real traffic
• T-test
• Temporal aggregation
• Traffic flow
• Traffic monitoring
• Traffic state
• Traffic volumes
• Urban road
• Urban road traffic
• Vehicle detector

To overcome the shortcomings of the traditional cross-validated mean-square error model in determining the optimal temporal aggregation interval for traffic flow monitoring data, three fundamental traffic flow parameters, traffic volume, time-mean speed, and occupancy, were used to represent traffic flow operating states for urban road. Based on the traditional model for estimating the cross-validated mean-square error of traffic state, an improved cross-validated mean-square error model for traffic state vector was proposed to estimate the fluctuations of traffic flow monitoring data at different temporal aggregation intervals. Then, a hypothesis test was established for the traffic state vector mean difference, and the t-test method was applied to find the inflection point of the changes in the cross-validated mean-square errors to determine the optimal temporal aggregation interval for traffic monitoring data. Taking the real traffic flow data collected by vehicle detectors on the urban roads of Kunshan City as an example, the optimal temporal aggregation intervals of traffic flow monitoring data for different types of urban roads were quantitatively analyzed. The results show that in practical applications, an optimal temporal aggregation interval of 5 min can be selected for the traffic flow monitoring data for urban roads.

• 出版日期2012
• 单位东南大学