Advancing Robotics with Graph Neural Networks

This chapter explores the challenges and innovations in utilizing floor plan datasets for machine learning in robotics, particularly focusing on room classification. It discusses the transition from traditional map representations to graph-based models, emphasizing the role of graph neural networks and their algorithms in enhancing spatial awareness and navigation. Through detailed comparisons and analyses, the chapter highlights the potential of advanced graph-based approaches in improving performance for room classification tasks.