Comparative Evaluation of Density Estimation Methods on Different Uninterrupted Roadway Facilities: Few Case Studies in India

Traffic density is one of the fundamental macroscopic characteristics and it is of prime importance when assessment of a facility has to be done based on both users as well as planner’s perspective. According to the Highway Capacity Manual (HCM), USA, for expressways and multilane roads, the Level-of-Service (LoS) has been defined taking density as the governing factor. Density is treated as the fundamental macroscopic spatial parameter of traffic flow, as it directly indicates the quality of traffic and ease with which one can drive. The present research study focuses on applicability of density estimation methods on multi-class, heterogeneous, non-lane based Indian traffic condition. In this research study, mid-block road sections namely, on Delhi-Gurgaon Expressway, Ahmedabad-Vadodara Expressway, National Highway (NH)-8 and urban arterial road in Chennai are considered. The expressways are the highest class of roads in Indian road network with different roadway, traffic conditions and with access control facilities. Delhi–Gurgoan Expressway is eight-lane divided facility and Ahemdabad –Vadodara Expressway is four-lane divided with 2.6 m paved shoulder. On the other hand, selected section on National Highway (NH8) is four-lane divided road. Real traffic data was collected through video-graphic survey on all these roads with different traffic conditions. HCVs (Heavy Commercial Vehicles) and MCVs (Multi-axle Commercial Vehicles) were the major traffic composition in congested regime. Thus, the present study focuses on the comparison of density estimation methods on different roadway and traffic conditions. The three methods to be employed for the purpose of estimating traffic density on the study section are: (1) using traffic flow fundamental equation relating speed, flow and density, HCM definition (2) using Cumulative input–output Plots (Input–Output method) (3) Eddie’s (x, t) method. This paper aims to empirically quantify the difference in the density estimation based on the aforementioned methods. Assuming input–output provides theoretical density, the errors in the estimation of density using fundamental equation under different traffic flow conditions are also quantified. In spite of growing body of literature disputing about the effectiveness and applicability of various density estimation methods, the key finding from this research indicates that all three abovementioned methods works very well under uncongested traffic flow condition. However, for oversaturated traffic conditions the density estimation using fundamental relationship has errors, primarily due to errors in estimation of the space mean speed since the vehicles which persisted within trap length for period longer than time-interval under consideration are not incorporated in the calculation, which is not the case for other two methods. Moreover, the research study concludes that smaller trap length (<100 m) can have errors in estimation of density values as compared to actual density values. The findings are useful for condition assessment of traffic flow for design and operation purposes.