Helping robot teams work together
ECITON-R studies how robots share information and divide tasks when communication is limited. We’re building Swarm Lab to make those choices visible: what each robot knows, what it hears from others, and why it picks its next job.
What we’re working on
Imagine a team inspecting a building. One robot finds a job behind a doorway, while another hasn’t seen that part of the floor. Sharing the discovery could save a search, but sending every observation would use bandwidth, battery and computing time. ECITON-R studies which information is useful enough to pass on.
The aim is a team that can divide work from local knowledge, cope with delayed or missing messages, and replace a coordinator when it goes offline. A coordinator is a temporary organizing role, with rules for its election and replacement.
In Swarm Lab today
You can follow a robot’s known tasks, received reports and last reported peer positions. Scenes include task claims, message sharing and temporary coordinators. In the new Map exploration scene, one rover discovers walls, remembers its own map and plans routes through observed free space.
That rover experiment assumes an accurate position. Sharing obstacle maps and estimating position through SLAM are still future work.
The direction
We’re connecting more of each robot’s decision code to its local evidence and working toward a complete inspection mission: explore, inspect, return for service, and resume. Recorded comparisons will help us test different sharing rules under the same conditions.
Training decision rules in simulation and building a physical ECITON-R Node come later. The Node would connect to a supported robot above its existing navigation and safety systems. Today’s movement, radio and energy are simulation models.
Latest updates
ECITON-R: From a Local Decision to a Delivered Message
Direct-message decisions now run in the interactive simulator, the Node PCB design has progressed, and cooperative sensing is the next physical demonstration.
ECITON-R and the Local AI Direction
Small local models for useful robot communication, the ECITON-R Node hardware vision, and the next demonstration I want to build.
Follow the project
Occasional emails about new experiments, development notes, and Swarm Lab updates.