A Trajectory Planning Model for Autonomous Vehicles: Lane Change Manoeuvres Using Fuzzy Logic
Abstract
The Intelligent Transportation Systems (ITS) are critical in the automated Lane Change (LC) manoeuvre. The paper presents a developed hybrid lane-change planning model that incorporates fuzzy logic and polynomial trajectory generation. In contrast to current quintic-based approaches, which optimise only smoothness, the proposed model also employs adaptive fuzzy weighting to reduce both lateral jerk and the duration of lane changes, thereby maximising passenger comfort and vehicle stability during travel. A dual-objective goal modulates the trajectory's curvature based on dynamically varying Vehicle-to-Vehicle (V2V) real-time communication data. The comparative simulation results indicate better performance than the traditional polynomial-only, Model Predictive Control (MPC), and Deep Q-Network (DQN) planners, with a minor trajectory deviation (I 0.2 m), lower lateral acceleration (I 0.18 g), and a stabilised yaw rate (I 5). The flexibility and accuracy of the proposed approach, as well as its usefulness in real-world autonomous driving in mixed traffic conditions, are emphasised.