Dynamic Trajectory Planning of Autonomous Lane Change at Medium and Low Speeds Based on Elastic Soft Constraint of the Safety Domain

Most current research on the trajectory planning of the autonomous lane change focuses on high-speed scenarios and assumes that the states of the surrounding vehicles keep stable during the lane change. The methods based on geometric-curve are mostly used for trajectory planning. In this paper, considering the inevitable development of the autonomous driving, the surrounding vehicles are assumed to be driven by human drivers, while the ego vehicles are able to autonomously change lanes. Representative local lane-change scenarios are then designed and analyzed in detail aiming at medium- and low-speed lane-change conditions. Additionally, in contrast with most research, dynamic trajectory planning which considers the possible state variations of the surrounding vehicles and the driver characteristics is studied and described by a fifth-order polynomial function. The safety and comfort of the dynamic trajectory planning are validated through simulation. Moreover, the elastic soft constraint of the safety domain is designed, whereby the sensitivity of the studied dynamic trajectory planning system is reduced under the premise of ensuring safety. The effectiveness of the elastic soft constraint in terms of improving comfort during the lane change is verified through simulation. The availability of the dynamic trajectory planning system with the elastic soft constraint is demonstrated with the addition of trajectory tracking based on model predictive control, showing its potential in practical applications.