A macroscopic first-order lane-group-based signalized node model

Abstract

In macroscopic dynamic network loading algorithms, node models that find the possible transfer flow at signalized intersections often disregard the role of turning lanes. This may result in finding wrong solutions when different turning lanes have different green times. This paper extends the recently presented node model of (Yahyamozdarani & Tampère, 2022a) to a lane-group-based signalized node model. The presented node model can consider different green times for different turns coming from the same link while it does not impose any extra requirements on the link model. In fact, the link model remains ignorant of the lane compositions. Consequently, the signalized node model would be more accurate regarding the exact green time of each turn while the computational burden is kept limited and the compatibility with the connecting link model is unchanged.