A Novel Decision-making Framework for Lane Change Considering Intentions of Traffic Participants

Abstract

Drivers' lane-changing behavior is limited by road-ways, adjacent traffic, etc. A lane-changing maneuver consists of whether, when, and how to change lanes. This paper proposes a lane-changing decision-making framework that considers these three processes. This framework incorporates lane-changing intention identification, generation, and planning. First, the Gaussian Hidden Markov model identified traffic vehicles' lane-changing intentions. This determines the ego vehicle's lane-changing intention and longitudinal safety corridor. The ego vehicle's lane-changing trajectory is designed using a fifth-degree polynomial to assure continuous acceleration and zero acceleration at the start and end positions. Finally, simulation findings reveal that the proposed lane change decision-making framework can predict traffic vehicles' lane-changing intentions 1.7 seconds (on average) in advance. Like a driver, the proposed method can make lane-changing decisions in advance.