چكيده به لاتين
Abstract:
Population growth, increasing the number of vehicles, traffic network complexity and necessity of having a safe, comfortable and inexpensive transportation system shows the importance of transportation problems. Parameters like flow, speed, density and headway specify the state of traffic and thus forecasting such parameters during rush hours or off-peak traffic times can be effective in traffic management system. Due to the nature of basic flow parameters it seems that spatial analysis can be effective in forecasting problems because of surveying probable relation of events and specifying a trend for their occurrence. By using such forecasts and managements based on these forecasts it is possible to reduce the congestions and accordingly reduce the times spent in the roads, distances traveled by the vehicles, air pollution and noise and also the expense of fuel.
In this study for the first time the ability of spatio-temporal kriging models in short-term traffic forecast of flow rate in urban districts is studied. The model is used for forecasting flow rates of single highway and also a collection of highways. Results show that the model’s efficiency is high as 85 percent when using 13 hour data and in case of reducing them to two hours it will diminish for about 4 percent. In case of considering flow rates of adjacent stations in other highways the results show salient improve in efficiency in comparison with single highway analysis. Using flow rates of the same times of the same day of week as input show higher than 82 percent match despite the small sample size of input.
Keywords: Traffic congestion, Short-term traffic forecast, Spatio-Temporal kriging, Advanced Traveler Information Systems (ATIS)
