UWB radar object recognition for SLAM

In situations where the environment is filled with dust or smoke, like in many emergency scenarios, UWB Radar is a possible technology to still sense the surrounding and build maps, thus helping firemen and other rescue personal. In this paper, a combination of two algorithms is presented. A basic simultaneous localization and mapping (SLAM) algorithm relying on simple indoor features is used to explore the environment and build a map. This algorithm is augmented by an object-recognition (OR) algorithm that is capable of recognizing and localizing complex shapes. These shapes can be used to improve the map resolution accuracy and resolve ambiguities. First tests using an m-sequence UWB Radar indicate the feasibility of this concept.