도시고속도로의 유입연결로 합류영향권내 최적의 속도-밀도특성모형 구축에 관한 연구

Abstract

Most of the urban freeways become a primary arterial in the urban transportation system, but they don't play their roles an more as a high-speed roadway in the big cities. They have suffered from severe traffic congestion due to the increased traffic exceeding their capacities regardless of the morning and afternoon rush-hours or the inbound and outbound directions since 1980's. Additionally the transportation systems management(TSM) and intelligent transportation systems(ITS) were emphasized on for solving the urban transportation problem with the construction and extension of the new transportation facilities limited in the late 1980's, but they didn't reduce the traffic congestion on the urban freeways and increased the inflow traffic to the urban freeways rather. Thus, urban freeway traffic management system could be systematically devised with the appropriate transportation demand management for promoting the efficiency of urban freeway and also maintaining the function of urban freeway if possible.

The purpose of this study was to investigate the real-time traffic characteristic data for identifying the speed-density characteristics within the on-ramp merge influence area of the urban freeway, compare the speed-density characteristic relationship for constructing the linear or non-linear speed-density models, and finally suggest the optimal speed-density characteristic models within the on-ramp merge influence area of the urban freeway. First, the real-time traffic characteristic data were collected on the 3 locations within the on-ramp merge influence areas of the urban freeway in Busan. Second, databases were constructed based on every 15 minute data for 24 hours a day during weekday. Third, model construction and verification procedures were conducted for optimally modeling and verifying the relationship between the speed-density. Finally, the optimal speed-density characteristic models were suggested within the on-ramp merge influence area of the urban freeway.

From the traffic characteristic analyses and modeling/verifying between the speed and density within the on-ramp merge influence area of the urban freeway, the following conclusions were drawn:

ⅰ) Flow characteristics did not show a distinct peak period on the on-ramp junctions located between the urban and suburban areas in the urban freeway under the study, but they did a distinct peak period on the on-ramp junctions study close to the urban or suburban areas in the daytime.

ⅱ) Speed characteristics showed a considerable difference in speed distribution depending on the locations within the merge influence area of the urban freeway under the study. Especially, upstream speed showed higher than downstream one by about 10% on the on-ramp junctions regardless of urban or suburban areas.

ⅲ) Density characteristics were shown to be a distinct difference in density distribution within the merge influence area of the urban freeway under the study. Especially, upstream density showed less than downstream one by about 30% on the on-ramp junctions regardless of urban or suburban areas.

ⅳ) Linear models showed a considerable explanatory power with the determination coefficients of about 0.788 to 0.961 at the upstream location, 0.906 to 969 at the downstream location, and 0.598 to 0.796 at the ramp location of the urban freeway under the study, respectively.

ⅴ）Linear models proved a considerably high correlation between speed and density with the correlation coefficients of about 0.888 to 0.980 at the upstream location, 0.952 to 984 at the downstream location, and 0.773 to 0.892 at the ramp location of the urban freeway under the study, respectively.

ⅵ) Non-linear models such as the existing Greenberg, Underwood, May, Drew, and Pipes & Munjal ones were compared with the cubic models suggested in this study at the locations of the urban freeway. On the whole, cubic models proved a higher explanatory power and correlation than the non-linear ones in the above.

ⅶ) Cubic models showed a high explanatory power with the determination coefficients of about 0.832 to 0.974 at the upstream location, 0.927 to 985 at the downstream location, and 0.614 to 0.800 at the ramp location of the urban freeway under the study, respectively.

ⅷ）Cubic models proved a high correlation between speed and density with the correlation coefficients of about 0.912 to 0.987 at the upstream location, 0.963 to 986 at the downstream location, and 0.783 to 0.894 at the ramp location of the urban freeway under the study, respectively.

ⅸ) Non-linear speed-density characteristic model such as Cubic models suggested in this study showed a higher explanatory power and correlation than linear model in the construction of the speed-density at the locations of the on-ramp junction in the urban freeway.

ⅹ) Cubic models showed the increase of speed with the increase of density under the low or high densities for maintaining the fitness between the observed and estimated data, which was based on the difference in speed-density relationship between the vehicles, driving habits, and etc.

Thus, it was concluded that the cubic models would be very valid in the analysis of the speed-density characteristic relationship on the urban freeway on-ramp junctions in the country.

Keywords: speed, density, linear model, non-linear model, urban freeway, merge influence area, on-ramp junction, correlation.