A Novel Trajectory Planning Method for Automated Vehicles under Parameter Decision Framework

Abstract

Decision and control in all stack scenarios comprise a key issue in the design of automated vehicle control systems. Thus, in higher level, automated vehicles, the decision and the form of the decision should be able to adapt to diverse, changeable, and complex scenarios, which increase the complexity of trajectory planning. In this paper, a parameter decision framework in which the decision is described with key parameters, rather than specific behaviors, such as lane-changing or car-following, is considered. Under this framework, a novel trajectory planning method is proposed to implement behavior with integrated longitudinal and lateral control, in which a nonlinear motion control model is established. The nonlinear model predictive control (NMPC) method with terminal constraints without a predefined path form is applied, which presents more flexibility for changeable decisions. Both the trajectory planning controller and the overall framework are verified by simulation. The results show the validity of the controller and the framework.