Structuring Motion-Based Communication as a Language for Robotics: Analysis with Dense Optical Flow

The construction industry is currently facing a labor shortage, prompting many researchers and practitioners to explore construction automation through advancements in robotics. For successful construction operations with multiple agents, effective and intuitive communication is essential. To this end, this research proposes a Motion-Based Communication (MBC) framework, which enables collaboration using movement. MBC provides a robust alternative that remains unaffected by various types of disruptions. This work contributes to the field of robotics by developing a motion-based language represented through an ontology. The ontological representation is encoded via directional shifts in motion rather than poses and is decoded by analyzing optical flow vectors. The experimental results in a lab environment using a UR5e robot indicate feasibility, with a shift inference accuracy of 98.8%. These findings can contribute to the development of a multi-agent MBC language to support communication in real time during construction operations.