29 May 2023
Swarm robotics, inspired by swarms in nature, has the potential for resilient, robust, and redundant solutions to a wide range of problems such as mapping, logistics, search and rescue, disaster recovery, and environmental monitoring. Many relatively simple and cheap robots, each following simple rules, with local interactions between themselves and the environment are capable of producing a desired emergent swarm-level behaviour.
In this work, EMERGE partners from the University of Bristol investigate the Gaussian Belief Propagation (GBP) algorithm, which shows great potential as a general distributed knowledge inference algorithm for use within swarms of robots. Individual robots build local factor graphs describing the environment, continually running GBP within their fragment to infer state. Robots coming within communication range of each other exchange information that connects these graph fragments and widens perception of each agent.
The distributed paradigm of GBP is a natural fit with that of swarm robotics, and they intend to use it to realise their vision of Distributed Situational Awareness for swarms of robots. By giving swarms an enhanced awareness of the world, simple and robust swarm algorithms can approach the performance of centrally planned solutions in logistics applications.
Read the paper in the link below.
More Information
