Publications / 2021 Proceedings of the 38th ISARC, Dubai, UAE

Context-appropriate Social Navigation in Various Density Construction Environment using Reinforcement Learning

Yeseul Kim, Bogyeong Lee, Robin Murphy and Changbum Ahn
Pages 505-512 (2021 Proceedings of the 38th ISARC, Dubai, UAE, ISBN 978-952-69524-1-3, ISSN 2413-5844)
Abstract:

Construction environments are often densely populated with multiple resources (e.g., workers, equipment, and materials). As an increasing number of mobile robots are expected to coexist and interact with humans at close proximity, it is necessary that these robots are capable of not only avoiding collisions during navigation but also navigating in a way that is appropriate and acceptable to coworkers based on the shared social convention of the construction workplace. That is to take into account appropriate normal proxemic consideration (e.g., personal space) as well as work-related proxemic considerations (e.g., work zone). Failing to do so can result in violation of personal and workspace, which can lead to fatal accidents and inefficiency. To accommodate this need, this study aims to develop a navigation model that enables robots to navigate in a socially and contextually compliant manner while following a globally planned trajectory. We created a simulation environment composed of five human workers and a robot agent where the agent learned socially aware policies using reinforcement learning algorithms. The result showed that although the overall success rate was slightly lower than the baseline algorithm, our agent successfully learned the corresponding personal space per different types of workers and secured the minimum distance respectively. This finding will contribute to building future construction mobile robots with social intelligence which are capable of understanding the context of the workplace and adapting to appropriate behaviors accordingly.

Keywords: Construction mobile robots; Socially/Contextually-aware Navigation; Human-Robot Interaction; Reinforcement Learning