Mediating Human-Swarm Interactions Through a Robot Helper: A Shepherding-Inspired Approach
Organizations: Arihant Yadav & Eakta Jain are with Department of Computer & Information Science & Engineering , University of Florida, Gainesville, USA · Sophie Jörg is a Professor in the Information Systems and Applied Computer Sciences Department at University of Bamberg, Germany
Abstract
This work investigates human-swarm interaction (HSI) for steering large collectives through an intermediary robot helper. While prior robot-swarm control and HSI techniques commonly assume swarm members are addressable, these assumptions often fail for ``ungovernable'' swarms such as wildlife flocks and schools of fish. Motivated by pastoral shepherding, we compare low-level teleoperation to a higher-autonomy motion-primitives interface in a virtual reality (VR) shepherding task under varied swarm cohesion dynamics. Results indicate that motion primitives that operate on collective shape (e.g., boundary-based control) improve task performance and reduce perceived workload.
Figures & tables
| Command | Functional Meaning |
|---|---|
| Come-bye [ 6 , 14 ] | Flank clockwise around stock to adjust pressure and balance relative to the handler. |
| Away [ 6 , 14 ] | Flank counter-clockwise around stock to adjust pressure and balance relative to the handler. |
| Walk-up [ 14 ] | Approach directly toward the livestock to apply forward pressure and initiate movement. |
| Stand/Stop [ 6 , 14 ] | Halt or hold position to remove or modulate pressure on the stock. |
| Look back [ 14 ] | Disengage and gather left-behind livestock. |
| That’ll do / Call off [ 14 ] | Terminate work and return control focus to the handler. |
| Partici-pant ID | Snippet, When asked about what condition they prefer, and why: |
|---|---|
| p3 | ‘the teleoperation condition, just because I like to have control over the agent at all times’ |
| p10 | ‘the shepherding condition, because it’s easier on the shepherd and harder on the sheepdog’ |
| p12 | ‘the first (shepherding) condition; it seemed like the dog knew exactly what to do, and when I controlled the dog, I kept messing up’ |
| Partici-pant ID | Snippet |
|---|---|
| p6 | When asked to describe how the trial went : ‘(for the shepherding condition) it went great! It was almost as if the agent knew the best way to move the swarm’ |
| p15 | When asked about how demanding the two trials were compared to each other : ‘in the first condition (teleoperation), I felt like I was responsible for the dog’s movement at all times and I couldn’t take a break. the second condition was easier because i was more relaxed’ |
| Partici-pant ID | Snippet |
|---|---|
| p5 | When asked what their ideal control scheme for this task would be : ‘I like the sheepdog control scheme, but the arcs the dog made around the swarm were too tight sometimes. I would like for there to be an additional button that helps me fine-tune the arc radius’ |
| p13 | When asked what changes they would make to the agent to make the task easier : ‘I think having multiple helpers instead of a single helper would be good, because I can then send helpers to round up split groups’ |
Swarm Robotics
Swarm robotics studies how groups of robots coordinate through local sensing, communication, and interaction to accomplish shared tasks.
See all 104 papers