DARPA Centibots Project
DARPA funded the Centibots project, a multi-institution effort to demonstrate coordinated autonomous mapping and search using up to 100 mobile robots simultaneously, with results reported from 2003. It addressed whether large teams of robots could self-organise without continuous human supervision.
Background
By the early 2000s, researchers had shown that a single autonomous robot could build a map of an unknown space while keeping track of its own position within it. This technique, called simultaneous localisation and mapping (SLAM), had become one of the central problems in mobile robotics, and several groups had workable solutions for a robot operating alone.
Scaling that up to many robots at once was a different problem. Each robot generates its own partial map, moves unpredictably relative to the others, and may lose contact with them. Merging those separate maps into one coherent picture, in real time, without a person in the loop to sort out the inconsistencies, had not been done at any meaningful scale. Most multi-robot work at the time involved small numbers of machines, often two or three, in controlled laboratory conditions. Whether the approach could hold together with tens or hundreds of robots in a real environment was an open question.
DARPA wanted an answer to that question at a scale that would matter for search operations: not three robots, but up to a hundred.
What happened
The Centibots project brought together several research institutions under DARPA funding: SRI International’s Artificial Intelligence Center, the University of Washington, and the University of Freiburg, among others. Kurt Konolige at SRI, Dieter Fox at the University of Washington, and Wolfram Burgard at Freiburg were key figures in the work. The goal was to field a team of up to 100 mobile robots that could enter an unknown building, spread out, map it, and search it without continuous human direction.
The robots used occupancy grid mapping, a method that represents space as a grid of cells each marked as free, occupied, or unknown. As each robot moved through the environment, it updated its own local grid using laser range sensors. The harder part was fusing those individual maps into a single shared map that all the robots could use to coordinate where to go next. The system had to handle the fact that robots might not be in radio contact continuously, that their position estimates would drift over time, and that the same corridor might appear in several different robots’ maps under slightly different coordinates.
Results from the project were published from 2003 onward, with a detailed account appearing at the IEEE International Conference on Robotics and Automation in 2004. The experiments demonstrated that the approach could scale considerably beyond the small teams that had been tried before, moving multi-robot coordination from a laboratory exercise toward something closer to a deployable capability. The work did not put a hundred robots into a live emergency response, and the gap between a structured evaluation and a real disaster site remained large. But the project established that the underlying methods, decentralised control combined with real-time map fusion, could hold together at that scale.
Why it mattered
Centibots was one of the first systematic demonstrations that a large team of autonomous ground robots could collaboratively explore and map an unknown environment without human direction, pushing the boundary of scalable multi-robot coordination from laboratory experiments toward field-scale deployments. The project's emphasis on decentralised control and real-time map fusion informed subsequent research in swarm robotics and autonomous search-and-rescue systems. Because the work was DARPA-funded and subject to structured evaluation, it produced benchmark results against which later multi-robot systems could be compared.
People
Kurt Konolige, Dieter Fox, Wolfram Burgard
Organisations
DARPA, SRI International, University of Washington, University of Freiburg
Sources
- Centibots Project, SRI International AI Center.SRI International.Primary source
- Large Scale Multi-Robot Mapping: The Centibots Project.IEEE International Conference on Robotics and Automation (ICRA 2004).Primary source
Cite this page
AI Achievements. (2003). DARPA Centibots Project. Retrieved 2026-08-22, from https://achievements.ai/milestone/darpas-centibots-project
@misc{achievements_darpas_centibots_project,
title = {DARPA Centibots Project},
author = {{AI Achievements}},
year = {2003},
url = {https://achievements.ai/milestone/darpas-centibots-project}
}