I. Introduction
Traversal of unknown environments remains a challenge for autonomous mobile robots (AMR), especially in the presence of static and dynamic obstacles. As humans, we perform such tasks with ease by learning to slow down when approaching busy intersections, or taking wider paths to obtain a better view around a corner. Robots are not at this level of autonomy yet. In this work, we insist that agile and safe motion in cluttered environments is achieved only when known obstacles and their occluded areas with associated uncertainties are considered equally during motion planning.