Search

Traffic congestion is one of the mobility problems characterised in terms of travel time, speed, and social, environmental, and economic costs. Various factors such as the traveller's perception of acceptability, variation in congestion measurement, the traveller's perception of acceptability, and the unreliability of the measurement. One of the crucial issues for policymakers in identifying and quantifying the congestion levels. The aim is to develop a fuzzy-based methodology to emulate human expertise in measuring congestion levels in arterial streets for heterogeneous traffic flow using two input variables: Speed and Density. The outcome was a single congestion index value between (0 and 1). The (FMM) Fuzzy Mamdani model was implemented using MATLAB R2020. The memberships were defined for inputs and outputs variables, and fuzzy rules were developed based on the available parameters. The model used real-time traffic data on the major road network of Latakia city, Syria. The results from the fuzzy system compared with the results from using one parameter, the speed performance index. The results show notable differences between the two approaches and conclude that using fuzzy logic for congestion measure by combining two or more variables reflects and represents the ambiguous nature of congestion better than one input or index. [ABSTRACT FROM AUTHOR]

The response variable on walking frequency requires special consideration when it has a large number of zeros arising from two underlying processes. The excess zeros exhibit variability that exceeds the hypothesized structure in the widely used Poisson regression model (PRM). The PRM model places a constraint on modelling for the zero-outcome probability and the extra proportion of observations in the zero class. The zero-inflated Poisson (ZIP) model considers two separate data generation processes for non-Poisson and Poisson zero outcomes, and therefore the model's predictive ability is increased. Although the process by which these zeros arise may not be a key focus of the analysis, the selection of a model for this process may have consequences for theoretical and empirical refinement. In this paper, we have developed ZIP regression to model the excess zeros in walk trip frequency analysis. The ZIP regression is examined with PRM. Our study findings suggest that ZIP should be used in modelling walk trip frequency given that the data-generating mechanism for zeros and positives reflects two different decision-making processes. We applied a special test for non-nested models proposed by Vuong to see the advantage of ZIP over PRM. Using the 2010 Greater Mumbai region activity-travel diary survey, the paper estimates Zero-inflated Poisson models at the individual level for the frequency of walking trips for home-based commute and non-commute purposes during a single travel day. The analysis indicates that pedestrian socio-demographic characteristics and built environment attributes affect work and non-work walking behaviour. The resulting pedestrian travel behaviour models are appropriate for the assessment of public health and sustainability policy analysis and planning. We conclude that the ZIP regression model is more appropriate for walking behaviour analysis and practice in the context of Indian metropolitan cities. [ABSTRACT FROM AUTHOR]