NSF-BSF: Culturally Adaptive Social Robot Navigation

George Mason University, Fairfax VA

Investigators

Abstract

This National Science Foundation (NSF) and US-Israel Binational Science Foundation (BSF) Collaborative Research Opportunities (NSF-BSF) grant will support research that will contribute new knowledge related to social robot navigation, promoting the acceptance and adoption of mobile service robots in crowded public spaces and advancing national prosperity and welfare. Social robot navigation techniques autonomously move mobile robots from one point to another, serving people by, for example, delivering food or packages. When these autonomous robots move in crowded public spaces, they not only need to share the same space and avoid surrounding humans, but also observe unwritten social norms, e.g., leaving enough social space, following the flow, and yielding to the right. However, such unwritten social norms vary from culture to culture, making a mobile robot developed for one culture difficult to adapt to another and therefore hindering the acceptance of mobile service robots in the new culture. This award supports fundamental research to develop social robot navigation techniques that can easily adapt to different cultures. While providing mobile robot services, such techniques will enable robots to efficiently navigate through dense human crowds without any collision and simultaneously follow culturally dependent social norms. A smooth integration of mobile service robots into daily lives will free people’s burden from repetitive and laborious tasks and increase the prosperity and welfare of people from different cultures. Therefore, results from this research will benefit the US economy and society. This research involves several disciplines including artificial intelligence, machine learning, human-robot interaction, and human factors. The multi-disciplinary approach will help broaden participation of underrepresented groups in research and positively impact engineering education. The cultural adaptiveness developed from this grant can overcome several limitations that existing mobile robot navigation techniques suffer from, including the "frozen robot" problem that impedes efficiency, moving too close to walking humans which compromises safety, and behaving in a way that does not conform with the underlying social norms therefore causing public resistance. This research will fill such a knowledge gap using data-driven approaches. The research team will collect a large-scale cross-cultural social robot navigation demonstration dataset, devise new algorithms to first comply with one underlying culture while maintaining efficiency and safety, design evaluation benchmarks and metrics to assess robot social compliance, and finally develop techniques that leverage commonality within different available cultures and quickly adapt to a new culture. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.