SocNavBench: A Grounded Simulation Testing Framework for Evaluating Social Navigation

• DOI: 10.1145/3476413
• Publication Date: 2022-02-14T18:54:53Z
• Authors: Abhijat Biswas; Allan Wang; Gustavo Silvera; Aaron Steinfeld; Henny Admoni
• Keywords: Bridge (graph theory); Baseline (sea); Test suite; Code (set theory)

The human-robot interaction community has developed many methods for robots to navigate safely and socially alongside humans. However, experimental procedures evaluate these works are usually constructed on a per-method basis. Such disparate evaluations make it difficult compare the performance of such across literature. To bridge this gap, we introduce SocNavBench , simulation framework evaluating social navigation algorithms. comprises simulator with photo-realistic capabilities curated scenarios grounded in real-world pedestrian data. We also provide an implementation suite metrics quantify algorithms scenarios. Altogether, provides test consistent interpretable manner. illustrate its use, demonstrate testing three existing baseline method showing how helps infer their trade-offs. Our code is open-source, allowing addition new by help evolve reflect advancements our understanding navigation.