RSS Workshop on Social Robot Navigation

Call for Papers

We invite short papers (max 4 pages, excluding citations, in standard RSS paper format, but not anonymized) on topics related to social robot navigation or social trajectory prediction. Papers may be submitted in PDF format via email by June 20th (anywhere on earth). All accepted papers will be presented as posters and introduced via lightning talks. The papers are encouraged (but are not required to) address the workshop’s focus topics and the following themes of interest.

Themes of Interest

Design and experiments
-What robot behaviors will enable untrained pedestrians to navigate effectively around robots?
-How do we design user studies to address and get beyond novelty effects?
-How do we measure the performance of a social robot navigation system?
-How do we deal with uncontrollable variables in experimental settings?
-What are appropriate benchmarking experiments and baselines?
-How do we ensure these benchmarks are accessible to the community?
-What is the role of simulation in social robot navigation research?
-What aspects of human behavior can and should be simulated?

Representations for social robot navigation
-What are appropriate representations for capturing important properties of crowd behavior?
-What are appropriate context models?
-What is the right level of context detail and how should it change depending on experimental settings?
-How do we encode behavior specifications into objective or reward functions?
-Should robots try to “mimic” human behavior?
-What behavior models and interaction metaphors are most effective for enabling smooth robot navigation?
-How do we verify that a robot navigation framework is correct and safe?
-How do we manage and deploy a safety-critical system without safety guarantees?

Datasets and data-driven approaches
-What properties should a social navigation dataset capture?
-How do we deal with uncontrollable variables in experimental settings?
-What aspects of human behavior can be simulated? What aspects of human behavior are most important for social navigation?
-How do the limitations of simulation inform our understanding of reinforcement learning? For example: if we train on social forces agents, can we quantify performance boundaries when deploying in real world settings?
-What role does supervised (e.g. deep) learning play in social navigation? How do we verify that we have not learned dataset specific artifacts?
-How do we address transfer/out of distribution performance for both reinforcement or supervised learning approaches? How does the fragility of RL and supervised learning inform future approaches?
-What role does active learning play in social navigation? For example, specific people might have specific preferences for interactive robot behavior---can we handle such things during field deployment?

Important dates

Workshop paper submission deadline:	June 20th (AOE)
Notification of acceptance:	July 1st
Camera-ready paper:	July 10th
Workshop:	July 13 Virtual (Zoom link)