Continuous 3D sensing for navigation and SLAM in cluttered and dynamic environments

Navigation and mapping is well understood for two-dimensional or static environments since valuable approaches exist for reacting to dynamic obstacles or for extracting static 3D information. However, today's challenge lies in combining the strengths of these approaches to obtain a system capable of performing safe navigation and obstacle avoidance based on rich 3D information of the environment while still being capable of reacting to sudden dynamic changes. In this paper we will present a methodology for continuously sensing environments in 3D and the necessary representations for exploiting the so gathered data in a way efficient enough to perform real-time 3D data based obstacle avoidance and online mapping.