Social Trajectory Planning for Urban Autonomous Surface Vessels

In this article, we propose a trajectory planning algorithm that enables autonomous surface vessels to perform socially compliant navigation in city's canal. The key idea behind the proposed is adopt an optimal control formulation which deviation of movements vessel from nominal human-operated penalized. Consequently, given pair origin and destination points, it finds trajectories resemble those vessels. To formulate this, kernel density estimation (KDE) build movement model prerecorded dataset, use Kullback-Leibler cost measure vessel's model. We establish analogy between our approach maximum entropy inverse reinforcement learning (MaxEntIRL) explain how can learn behavior On other hand, distinguish MaxEntIRL does not require well-defined bases, often referred as features, construct its function required many approaches context. Through experiments using dataset collected automatic identification system, demonstrate generated by them for canal beneficial reducing head-on encounters improving safety.