Reproduce “Legibot: Generating Legible Motions for Service Robots Using Cost-Based Local Planners” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Javad Amirian, Mouad Abrini, Mohamed Chetouani
Legibot, developed at Sorbonne University's ISIR, introduces a framework for service robots to generate legible motions by embedding observer-aware similarity and field-of-view costs directly into a real-time local planner. This approach improved human understanding of robot intentions in simulations and demonstrated practical deployability on a physical Pepper robot.
Key insights
- Integrating legibility as a cost directly into a local planner enables real-time adaptation of robot motion to convey intent in dynamic environments, moving beyond global or learning-intensive methods.
- Observer-aware factors like field of view and weighted importance of early cues significantly enhance legibility by tailoring the robot's communicative effort to the human's perception.
- The chosen cosine similarity metric for velocity vectors effectively quantifies trajectory similarity, making it suitable for discerning between intended and unintended goal interpretations by observers.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.