Abstract:
Smart factories powered by Multi-Robot Systems (MRSs) play a central role in Industry 4.0 and smart manufac-turing. MRS operating under dynamic task assignments of col-la...Show MoreMetadata
Abstract:
Smart factories powered by Multi-Robot Systems (MRSs) play a central role in Industry 4.0 and smart manufac-turing. MRS operating under dynamic task assignments of col-laborative robots in production flows suggests a new technology paradigm to achieve productivity, flexibility, and energy efficiency to revolutionize the industry. However, dynamic collaborative MRSs, given resource-limited wireless communication, robotic AI computing, and lacking globally accurate references lead to a technological challenge to facilitate resilient, reliable, and precise operation of smart factories. This paper innovatively resolves this cyber-physical challenge by aiming to align robotic actions in dynamic production flows based on the concept of grouping and propose a social learning-based method that utilizes Bayesian network and reinforcement learning (RL) to provide robust coordination that improves productivity and resilience against bursty cyber and physical inaccuracies in difficult dynamic sys-tems. Numerical experiments demonstrate successful fulfillment of technical requirements for the smart factory.
Published in: 2021 IEEE Global Communications Conference (GLOBECOM)
Date of Conference: 07-11 December 2021
Date Added to IEEE Xplore: 02 February 2022
ISBN Information: