Dynamic real-time indoor environment mapping for Unmanned Autonomous Vehicle navigation

Unmanned aerial and ground vehicles today excel at autonomous navigation due to their reliance on GNSS sensors. GNSS sensors are an unviable option for navigation in indoor environments due to lack of direct line-of-sight with satellites and environmental awareness. This paper introduces a robust method for unmanned autonomous vehicles (UAVs and UGVs) that generates a knowledge base in the form of a pointset about an indoor environment using geometric primitives and range imaging sensors. The method uses depth data from range imaging sensors along with pose data from visual odometry sensors to generate a map of the environment while exploring in order to build a knowledge base. This knowledge base can be used to navigate in the mapped environment while avoiding obstacles and finding optimal route from start to goal. This paper discusses the efficacy of the proposed method, it also demonstrates results generated by implementing the method in a controlled indoor environment.