Service time variability at manual operated tollbooths under mixed traffic environment: Towards level-of-service thresholds

Comprehending the variation in service time is essential for evaluating the capacity and level of service of the toll-plaza and developing policies for effective management of tollbooth operations. The present study analyzes the variation in service time using statistical distribution at manually operated toll plazas in India. For this purpose, traffic video data were collected for five toll-plazas in India. The vehicles approaching the toll plaza are divided broadly into seven different categories. The service time of different vehicle classes was extracted manually from the recorded video. The variation in service time for a total of 13 leader-follower pairs (small car as leader and follower) were modeled using statistical distributions. Based on the Chi-squared, Anderson Darling (AD), and Kolmogorov-Smirnoff (KS) test statistics, log-normal distribution was found as the best-fitted distribution among other potential statistical distributions. Investigations revealed that traffic composition, vehicular characteristics, variation in toll rates, drivers? and tollbooth operators? behavior significantly influence the variability in service time. The important outcome of this study is the development of service time based level of service (LOS) criteria, which were developed using k-mean clustering technique. The developed LOS thresholds can be very useful to field engineers and practitioners for evaluating the existing performance of tollbooth and accordingly facilitating the development of appropriate policies for effective management of toll plaza operations.