Requirement Analysis for Perception on Assistant Robots in Multi-Modal Environment Conditions

Abstract

This work presents the results of the preliminary analysis for realizing a robust multi-modal perception framework on mobile platforms at the German Aerospace Center (DLR). Within the evaluation, our focus is directed towards investigating robust approaches for localization, place recognition, and navigation in the visual domain. In addition to the analytical examination of promising state-of-the-art methods, an experimental study is carried out based on real-world datasets from mission-related environments. With this, the prevailing environmental properties are evaluated to identify the best-suited visual abstraction and characterization frameworks. In the end, we summarize our findings and realizations in recommendations for improving situation awareness within the considered projects.