Generating Legible and Glanceable Swarm Robot Motion through Trajectory, Collective Behavior, and Pre-attentive Processing Features
| Autor: | Sean Follmer, Lawrence H. Kim |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
0209 industrial biotechnology
Collective behavior Computer science business.industry Swarm robotics Swarm behaviour 020207 software engineering Robotics 02 engineering and technology Legibility Swarm intelligence Motion (physics) Human-Computer Interaction 020901 industrial engineering & automation Artificial Intelligence Human–computer interaction 0202 electrical engineering electronic engineering information engineering Robot Artificial intelligence business |
| Zdroj: | ACM Transactions on Human-Robot Interaction. 10:1-25 |
| ISSN: | 2573-9522 |
| DOI: | 10.1145/3442681 |
| Popis: | As swarm robots begin to share the same space with people, it is critical to design legible swarm robot motion that clearly and rapidly communicates the intent of the robots to nearby users. To address this, we apply concepts from intent-expressive robotics, swarm intelligence, and vision science. Specifically, we leverage the trajectory, collective behavior, and density of swarm robots to generate motion that implicitly guides people’s attention toward the goal of the robots. Through online evaluations, we compared different types of intent-expressive motions both in terms of legibility as well as glanceability, a measure we introduce to gauge an observer’s ability to predict robots’ intent pre-attentively. The results show that the collective behavior-based motion has the best legibility performance overall, whereas, for glanceability, trajectory-based legible motion is most effective. These results suggest that the optimal solution may involve a combination of these legibility cues based on the scenario and the desired properties of the motion. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|