Decentralised construction of a global coordinate system in a large swarm of minimalistic robots

2023-02-28 · Michal Pluhacek, Simon Garnier, Andreagiovanni Reina

Collective intelligence and autonomy of robot swarms can be improved by enabling the individual robots to become aware they are the constituent units of a larger whole and what is their role. In this study, we present an algorithm to enable positional self-awareness in a swarm of minimalistic error-prone robots which can only locally broadcast messages and estimate the distance from their neighbours. Despite being unable to measure the bearing of incoming messages, the robots running our algorithm can calculate their position within a swarm deployed in a regular formation. We show through experiments with up to 200 Kilobot robots that such positional self-awareness can be employed by the robots to create a shared coordinate system and dynamically self-assign location-dependent tasks. Our solution has fewer requirements than state-of-the-art algorithms and contains collective noise-filtering mechanisms. Therefore, it has an extended range of robotic platforms on which it can run. All robots are interchangeable, run the same code, and do not need any prior knowledge. Through our algorithm, robots reach collective synchronisation, and can autonomously become self-aware of the swarm's spatial configuration and their position within it.

📄 PDF Abstract BibTeX arXiv:2302.14587

Code (1)

tbu-ailab/kilobot-self-localization 공식 구현

Tasks

Position

Methods 이 논문이 사용한 방법론

AWARE We propose to theoretically and empirically examine the effect of incorporating weighting schemes into walk-aggregating GNNs. To this end, we propose a simple, interpretable, and…

Multi-Agent Common Knowledge Reinforcement Learning

2018-10-27 · NeurIPS 2019 12 · Christian A. Schroeder de Witt, Jakob N. Foerster, Gregory Farquhar, Philip H. S. Torr 외

Cooperative multi-agent reinforcement learning often requires decentralised policies, which severely limit the agents' ability to coordinate their behaviour. In this paper, we show that common knowledge between agents al…

Multi-agent Reinforcement Learning reinforcement-learning Reinforcement Learning Reinforcement Learning (RL)+2

QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning

2018-03-30 · ICML 2018 7 · Tabish Rashid, Mikayel Samvelyan, Christian Schroeder de Witt, Gregory Farquhar 외

In many real-world settings, a team of agents must coordinate their behaviour while acting in a decentralised way. At the same time, it is often possible to train the agents in a centralised fashion in a simulated or lab…

Multi-agent Reinforcement Learning reinforcement-learning Reinforcement Learning Reinforcement Learning (RL)+3

Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning

2020-03-19 · Tabish Rashid, Mikayel Samvelyan, Christian Schroeder de Witt, Gregory Farquhar 외

In many real-world settings, a team of agents must coordinate its behaviour while acting in a decentralised fashion. At the same time, it is often possible to train the agents in a centralised fashion where global state …

Multi-agent Reinforcement Learning reinforcement-learning Reinforcement Learning Reinforcement Learning (RL)+3

An Asynchronous Decentralised Optimisation Algorithm for Nonconvex Problems

2025-07-30 · Behnam Mafakheri, Jonathan H. Manton, Iman Shames arxiv

In this paper, we consider nonconvex decentralised optimisation and learning over a network of distributed agents. We develop an ADMM algorithm based on the Randomised Block Coordinate Douglas-Rachford splitting method w…

Initialisation and Network Effects in Decentralised Federated Learning

2024-03-23 · Arash Badie-Modiri, Chiara Boldrini, Lorenzo Valerio, János Kertész 외

Fully decentralised federated learning enables collaborative training of individual machine learning models on a distributed network of communicating devices while keeping the training data localised on each node. This a…

Federated Learning