Abstract: In motion planning, efficiently navigating from a start state to a goal state in spaces with narrow passages remains a significant challenge. Recently, learning-based methods have attracted ...