In the future, when embodied artificial intelligence is ubiquitous, multiple robots, vehicles, and other smart devices will need to communicate and coordinate their actions.
Robots. Image credit: Farshadarvin via Wikimedia, CC-BY-SA-4.0
A recent paper published on arXiv.org looks into the problem of distributed localization: a set of moving devices that move and observe each other within a space have to estimate their locations.
A breakthrough Robot Web solution is proposed to general, fully distributed, and asynchronous many-robot localization. Each robot stores and maintains its own part of the full factor graph and updates and publishes a Robot Web Page of outgoing messages for other robots to download and read.
The ad-hoc, asynchronous messages contain only small vectors and matrices. Robots do not need any privileged information about each other; therefore, the whole system is fully dynamic, with robots joining or leaving at will.
We show that a distributed network of robots or other devices which make measurements of each other can collaborate to globally localise via efficient ad-hoc peer to peer communication. Our Robot Web solution is based on Gaussian Belief Propagation on the fundamental non-linear factor graph describing the probabilistic structure of all of the observations robots make internally or of each other, and is flexible for any type of robot, motion or sensor. We define a simple and efficient communication protocol which can be implemented by the publishing and reading of web pages or other asynchronous communication technologies. We show in simulations with up to 1000 robots interacting in arbitrary patterns that our solution convergently achieves global accuracy as accurate as a centralised non-linear factor graph solver while operating with high distributed efficiency of computation and communication. Via the use of robust factors in GBP, our method is tolerant to a high percentage of faults in sensor measurements or dropped communication packets.
Research paper: Murai, R., Ortiz, J., Saeedi, S., Kelly, P. H. J., and Davison, A. J., “A Robot Web for Distributed Many-Device Localisation”, 2022. Link: https://arxiv.org/abs/2202.03314