Modeling lateral placement and movement of vehicles on urban undivided roads in mixed traffic: A case study of India

In India, the majority of urban roads are undivided where the behavior of flows in a particular direction is predominantly influenced by the opposing traffic. Due to lack of lane segregation, the vehicles in ongoing direction occupy the opposing lane, which increases the lateral interactions between vehicles. These lateral interactions are influenced by various parameters such as vehicle types, driver behavior and vehicular speeds. Study of such complex interactions plays an important role in evaluating various management measures using microscopic simulation models. The lateral characteristics of vehicles, such as placement, separation and movement, act as necessary input for simulation models. The present study aims to analyze and model the lateral characteristics of vehicles on two-lane urban undivided roads. To achieve this, traffic flow data were collected from an urban undivided mid-block section in Bangalore City, India, using video graphic technique. Multiple linear regression model was developed for predicting the lateral placement of subject vehicle and it was found that lateral placement of subject vehicle is influenced by types and speeds of subject and opposing vehicles. Lateral separation for different types of ongoing (subject) and opposing pairs was also analyzed. The results show that both the ongoing and opposing vehicles have less freedom to move laterally when their sizes increase and hence, lateral separation decreases. The choice of path of vehicles' lateral shifts (left, current and right) on urban undivided roads was modeled using multinomial logistic regression. Lateral shift of a vehicle is influenced by speeds of subject vehicle and leader vehicle in current path, speed of leader vehicle in target path, and lateral gap between leader vehicles in current path and target path.