Service levels of sidewalks for pedestrians under mixed traffic environment using Genetic Programming clustering

This study aims at developing a model for pedestrian to evaluate service measure at the roadside walking environment from the qualitative and quantitative analysis in developing countries. The potential primary factors influencing service measure of pedestrian were summarized taking consideration to heterogeneous traffic flow condition and perception of road users. Total 735 real-time sense of participant’s satisfaction when they are using the sidewalks, road geometric as well as operational characteristics of the road segments were collected from the 73 sites. The selected sites belong to midsized cities with varying geometric features. Different statistical investigations like factor analysis, Wilks’ lambda test and stepwise regression analysis were carried out to develop PLOS model for sidewalk facilities. From the results it has been observed that the variables taken such as width of sidewalk, vehicle volume, pedestrian volume etc. are significantly influencing the pedestrian service measure and the model was validated with a significance value of R 2 = 0.972. The PLOS scores got from the model are then classified using Genetic Programming clustering to find the six PLOS ranges (A-F) which show score for PLOS A ≤1.8 with best condition and PLOS F > 5.1 having worst condition of service.