Estimation of Queue Length at Signalized Intersections Using Artificial Neural Network

Abstract

Signalized intersections are points in the transportation network where vehicles from various directions meet. They are critical points for traffic jams, and this is an application of applied science in the technology field. The vehicle queue length is one of the performance parameters of a signalized intersection. Long queues of vehicles pose a high risk of accidents involving many vehicles. Feedback signal control (actuated signal control) can improve intersection performance. One variable that can be used as feedback input is the vehicle queue length. Traffic in Indonesia is mixed traffic where various vehicles use the same road lane and with low lane discipline. This causes the traffic system to become complex stochastic, and non-linear. Modeling queue length using a static linear algorithm cannot capture the phenomenon of this complex traffic system. Therefore, this research aims to build a machine learning-based queue length model using artificial neural networks (ANN). This model studies the traffic system with historical data so that it can model queue lengths with reasonable accuracy through the training process. The estimation model was built and applied to the Muara Rapak signalized intersection, Balikpapan. Data on queue length for 10 days, 2 hours/day, was obtained using CCTV and direct field surveys. The model testing results show that ANN has a good level of accuracy with MAE, RMSE, and MAPE of 3.8 m, 4.9 m, and 6%, respectively.