Interactive-Teb: Effective Trajectory Planning in Mixed Traffic Scenarios Considering Kinodynamic Constraints

Trajectory planning in mixed traffic scenarios is a critical and demanding task in autonomous driving, since it necessitates addressing intricate interactions among traffic agents while conforming to kinodynamic constraints of the vehicle. Most existing methods struggle to simultaneously meet the above-mentioned requirements, resulting in unsafe driving behavior. To take this issue, we propose Interactive-TEB algorithm, which utilizes a time-optimal optimization framework to generate a safe and comfortable motion trajectory, considering kinodynamic constraints of the vehicle and obstacle avoidance. Moreover, Interactive-TEB incorporates interactive predicted trajectories of other traffic agents into trajectory optimization to consider the inevitable interactivity in mixed traffic scenarios. Extensive experimental evaluations across diverse mixed traffic scenarios demonstrate that the proposed algorithm outperforms existing methods in terms of safety, comfort, and effectiveness.