Science Robotics丨No Map, No Localization: How a Group of "Deaf and Dumb" Robots "Hear" the Shape of the Space They Are In
Science Robotics丨不用地图、不用定位：一群"聋哑"机器人如何"听"出自己身处什么形状的空间？
A study published in Science Robotics proposes a new approach called Spectral Swarm Robotics, where robot swarms infer their spatial shape through local communication patterns. By analyzing the diffusion of information and the eigenvalues of the communication graph, the method enables robots to determine their environment's geometry without global positioning or mapping. Experiments with Kilobots and Pogobots show the method's effectiveness in distinguishing between circular and annular shapes.
This research introduces a novel method for shape inference in robot swarms using spectral analysis of communication patterns, without requiring global positioning or mapping. The approach leverages diffusion processes and eigenvalues of the Laplacian matrix to estimate global geometry from local interactions, demonstrating robustness in low-bandwidth environments.
Source: Chinese robotics — Research news discovery · Read original article ↗
English translation unavailable. Choose the original article to read it.
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
Reported accuracy
94%
Reported accuracy
73%
number of robots
550 robots
number of iterations
30
coverage ratio
≈17
View original evidence
α 太小（<7.5） ：集群分裂成多个不连通子图，信息无法全局传播，λ₂ 估计不准 α 太大（>32） ：每个机器人的邻居过多，邻接矩阵局部近似于全连接，通信图拓扑趋同，且数值不稳定 α ≈ 17 ：覆盖率和连通性的最佳平衡点 这个定律的工程意义很直接： 给定场地面积和通信半径，就能算出需要多少机器人；给定机器人数量和场地，就能算出通信半径应该设多大。
Open source S21communication radius
<7.5
View original evidence
α 太小（<7.5） ：集群分裂成多个不连通子图，信息无法全局传播，λ₂ 估计不准 α 太大（>32） ：每个机器人的邻居过多，邻接矩阵局部近似于全连接，通信图拓扑趋同，且数值不稳定 α ≈ 17 ：覆盖率和连通性的最佳平衡点 这个定律的工程意义很直接： 给定场地面积和通信半径，就能算出需要多少机器人；给定机器人数量和场地，就能算出通信半径应该设多大。
Open source S21
- paper: not_reported
- code: not_reported
- weights: not_reported
- dataset: not_reported
What remains unknown
Not established in the collected evidence: Environment, Control, Data origin.
Reported performance applies to the described task. It does not establish general autonomy or deployment readiness.
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
经过30次迭代，分类准确率从单次迭代的66%提升到94%（圆盘）。
Open source S23
分类准确率73%。
Open source S24
在25台Kilobots和25台Pogobots的真实实验中，集群成功区分了圆盘和圆环两种形状，Pogobots达到100%分类准确率。
Open source S6
五、实验验证：Kilobots和Pogobots 5.1 仿真结果 在7种几何形状上仿真，最多550台机器人。
Open source S22
α 太小（<7.5） ：集群分裂成多个不连通子图，信息无法全局传播，λ₂ 估计不准 α 太大（>32） ：每个机器人的邻居过多，邻接矩阵局部近似于全连接，通信图拓扑趋同，且数值不稳定 α ≈ 17 ：覆盖率和连通性的最佳平衡点 这个定律的工程意义很直接： 给定场地面积和通信半径，就能算出需要多少机器人；给定机器人数量和场地，就能算出通信半径应该设多大。
Open source S21
Source：Chinese robotics — Research news discovery · sohu.com