Modelling acceleration decisions in traffic streams with weak lane discipline: A latent leader approach

Abstract

Acceleration is an important driving maneuver that has been modelled for decades as a critical element of the microscopic traffic simulation tools. The state-of-the art acceleration models have however primarily focused on lane based traffic. In lane based traffic, every driver has a single distinct lead vehicle in the front and the acceleration of the driver is typically modelled as a function of the relative speed, position and/or type of the corresponding leader. On the contrary, in a traffic stream with weak lane discipline, the subject driver may have multiple vehicles in the front. The subject driver is therefore subjected to multiple sources of stimulus for acceleration and reacts to the stimulus from the governing leader. Hence, only the applied accelerations are observed in the trajectory data, and the governing leader is unobserved or latent. The state-of-the-art models therefore cannot be directly applied to traffic streams with weak lane discipline.This prompts the current research where we present a latent leader acceleration model. The model has two components: a random utility based dynamic class membership model (latent leader component) and a class-specific acceleration model (acceleration component). The parameters of the model have been calibrated using detailed trajectory data collected from Dhaka, Bangladesh. Results indicate that the probability of a given front vehicle of being the governing leader can depend on the type of the lead vehicle and the extent of lateral overlap with the subject driver. The estimation results are compared against a simpler acceleration model (where the leader is determined deterministically) and a significant improvement in the goodness-of-fit is observed. The proposed models, when implemented in microscopic traffic simulation tools, are expected to result more realistic representation of traffic streams with weak lane discipline.

The Roles of Car Following and Lane Changing Drivers’ Anticipations during Vehicle Inserting Process: A Structural Equation Model Approach

“…The longitudinal anticipation is modelled as a CF driver's estimation of one or more leaders' movements in one or more future moments [18][19][20]. Lateral anticipation is presented as that a staggered leader in the same lane or the neighbouring lane of the CFV brings the driver's lateral disturbance and triggers his or her deceleration [21,22]. In some LC cases, the FV driver, who originally follows the LV in target lane, could anticipate adjacent LCV driver's inserting intention and decelerate voluntarily to create a larger inserting gap for the LCV [21].…”

Simulation Strategies for Mixed Traffic Conditions: A Review of Car-Following Models and Simulation Frameworks

“…In heterogeneous traffic conditions, a single following vehicle affected by several leading vehicles (front, staggered, and right) receives stimuli from multiple sources while the leading vehicle affects only the vehicle next to it. e use of the GM model to simulate the acceleration of the leading vehicle in an approach for identifying the leading vehicle was reported by Choudhury et al [33]. Amini et al [34] proposed a case study on carfollowing-motorcycle behaviour using the GHR model, which compares the longitudinal behaviour of a car following an erratic motorcycle to that of the car-following-car scenario.…”

Flexible car–following models for mixed traffic and weak lane–discipline conditions

“…Metkari et al [9] have modified an existing car-following model in order to take into account lateral movements and include mixed traffic conditions. Choudhury and Islam [10] have developed a latent leader acceleration model.…”

“…The probability of a given front vehicle to be the governing leader depends on the type of the lead vehicle and the extent of lateral overlap with the following vehicle [10].…”