A Comparison of Macroscopic Traffic Stream Models Under Non-lane-Based Heterogenous Highway Traffic

Abstract

AbstractTraffic stream models characterize the relationship between the traffic parameters and represent the spatio-temporal variations in the traffic conditions. This research work is an attempt to compare different macroscopic traffic stream models of the heterogeneous traffic on Indian highways. This work intends to develop accurate traffic stream models and select the appropriate one that provides the precise traffic flow characteristics by analyzing the traffic speed, flow, and density parameters, including the dynamic passenger car units (DPCU) factors. Field data from the National Highways (NH) of the Southern Indian region was collected using the Infra-red sensor-based technique. The speed-flow-density model that reported high accuracy among the others based on the highest R2 value and least error values (RMSE and MAPE) was considered the best-fitted model for representing the macroscopic traffic behavior. Pipes’ family models was found to be the best fitted model as it gave realistic results. Further, the comparison of the three different methods of passenger car unit (PCU) estimation (speed-area ratio method, Indo-HCM: 2017, and IRC-64: 1990) was carried out. The results reveal that the traffic capacity estimated using the Indo-HCM: 2017 suggested DPCU values overestimated the traffic capacity, while the IRC-64: 1990 suggested PCU values underestimated the traffic capacity compared to the traffic capacity estimated using the DPCU values from the speed-area ratio method. The study can be employed in practical traffic flow modeling applications such as design, control, and management of highway facilities if transport practitioners, consulting authorities, stakeholders, and other agencies are to plan and operate the highways under the new paradigm.KeywordsMacroscopic traffic modelFundamental diagramNon-lane-based trafficHeterogenous trafficIndian highway