Author : Palash Chandra Das, Neeti Zaman Bintee and A.K. Azad

Congestion, a problem that is not new for a third world country like Bangladesh. Like the other megacities of Bangladesh, Khulna is gradually experiencing this phenomenon in an acute manner. Khulna-Jessore highway that runs through the city and gives it a linear form works as the backbone of the traffic skeleton of Khulna city. This study focuses on developing a speed based model for predicting congestion scenario of Khulna city. Fulbarigate to Dakbangla midblock is chosen as the study corridor. The whole study corridor (10.73km) is divided into two equal halves and travelled 10 times with the stream and against the stream to generate 20 data set. Speedometer GPS survey along with moving observed method is used to collect necessary data. A latitude longitude based speed profile is found at each travel based on which travel speed index (TSI) is calculated and a multiple linear regression (MLR) model for predicting congestion is developed. Independent variables are selected based on the possibility of affecting congestion of the road section. The MLR model is found as TSI = 0.0167083 + 0.0001826*Volume (PCU/hr.) + 0.0019625*Speed Drop Factor, with 8.86 % error. Congestion is found severe at Doulatpur Bazar and Shibbari to Dakbangla circle within the study corridor. Due to illegal parking spot, road capacity is not utilizing properly. As a result in peak hour, speed drops more than average peak hour speed and congestion occurs.

Keywords: Traffic Congestion, Moving Observer Method, MLR Model, Speedometer GPS, Speed profile

Article published in International Journal of Current Engineering and Technology, Vol.8, No.6 (Nov/Dec 2018)