RSS Workshop on Social Robot Navigation

Focus Topics

Our goal is to organize a truly interactive workshop that will convene small, interdisciplinary working group sessions around each of the workshop’s focus topics. We will run a sequential set of three workshopping sessions, each focused on one of the key topics. Each session will start with a set of talks from our invited speakers, designed to provoke the audience with controversial ideas, stories, and/or arguments that lead to challenging questions. Then the organizers will divide up the workshop attendees into small, interdisciplinary working groups. Using the challenge questions as a prompt, we will task each small group to share their perspectives with each other and brainstorm answers. One team member in each small group will be tasked with capturing the ideas, which will then be shared with the full set of workshop participants. Together, the small groups will brainstorm, summarize, and present their key findings to the full workshop group at the end of each of the three working sessions.

Focus topic 1: What are some important tradeoffs in selecting representations for perception, planning, and control in social robot navigation?

Speakers: Reid Simmons, Jon How, Aleksandra Faust
Moderator: Chris Mavrogiannis, Marynel Vázquez

• What are the appropriate representations to capture important properties of collective crowd behavior?
• What is the right level of detail to include in context models? How does this change depend on the experimental setting?
• How do we encode behavior specifications into objective or reward functions?
• How do we verify that a robot navigation framework is correct and safe? How do we manage a safety critical system without safety guarantees?

Focus topic 2: How should we collect (or generate) and use data for social robot navigation?

Speakers: Anca Dragan, Todd Murphey
Moderator: Pete Trautman, Francesca Baldini

• What aspects of human behavior can be simulated? What aspects of human behavior are most important for social navigation?
• How does simulation limitation inform reinforcement learning? For example: if we train on social forces agents, can we quantify performance boundaries when deploying in real world settings?
• What is the correct reward function for reinforcement learning based social navigation?
• 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 account for “distribution shift” in social navigation settings?

Focus topic 3. What are the most important variables that influence social robot navigation performance? In other words, what levers should we be exploring?

Speakers: Takayuki Kanda, Wendy Ju
Moderator: Leila Takayama

• What robot behaviors will enable untrained bystanders/pedestrians to navigate effectively around robots? What behavior models and interaction metaphors are most effective for enabling effective robot navigation -- human-like, dog-like, what else?
• What social robot navigation performance metrics should we optimize for? How should we measure them? What are appropriate benchmarks? What are some good baselines? How do we ensure that these metrics are accessible to the community?
• How do we design social robot navigation studies to address and get beyond novelty effects in our data?