Safety Gap Factor-Based Lane-Changing Trajectory Planning Model

Abstract

In discretionary lane changing (DLC) decision models, software-based vehicles should be controlled using safe and comfortable trajectories. The simulated lateral and longitudinal trajectories are approximate trajectories, whereas the calibrated models provide more appropriate trajectories. Very few studies have used the calibration method to find safe and comfortable lateral trajectories at the starting and ending places of lane changing (LC); however, no study has provided safe and comfortable lateral and longitudinal trajectories at these places. This study uses calibrated lateral and longitudinal trajectory models with a comfortable lateral trajectory to pinpoint the safety gap at the target lane during LC. The updated LC trajectory model, in which the adopted lateral and longitudinal trajectory parameters are calculated, is calibrated using a genetic algorithm. This study indicated that the average root mean square error (RMSE) value is 0.93 ( f ) of calibrated data decreasing more than 70%, whereas the average RMSE value of simulation data is 1.93 ( f ). Additionally, the longitudinal positions during LC have an average RMSE value of 0.93 ( f ), while the simulation model's average RMSE value is 1.94 ( f ). Depending on the dataset used, the proposed safety gap can be applied in traffic software while DLC decision models such as binary logistic and game theory models are used. Keywords: lane changing, transportation planning, gap acceptance, safety factors, game theory model. https://doi.org/10.55463/issn.1674-2974.50.8.14