INTEGRATING THE SEMANTIC INFORMATION INTO PLANNING FOR AUTONOMOUS ROBOT: A SURVEY

Abstract

Planning is an essential capability and plays a pivotal role in robotic systems. Utilizing semantic
information to enhance efficiency and capability in both low-level and high-level planning is a new field
that builds upon advancements in semantic segmentation and object detection. It holds the potential to
equip robots with more intelligent behaviors. Therefore, despite being a nascent field, it has garnered
significant attention from the community and has already achieved some practical results. However,
there is yet to be a comprehensive statistical analysis, classification, and synthesis of published works in
this area. The goal of this paper is to fill that gap by surveying related publications and deriving general
principles on how semantic information can enhance the planning process. These general principles can
serve as a foundation to guide future research in this domain.