In EDS ansehen

A distributed multi-robot task allocation method for time-constrained dynamic collective transport.

Gespeichert in:
Bibliographische Detailangaben
Titel: A distributed multi-robot task allocation method for time-constrained dynamic collective transport.
Autoren: Shan, Xiaotao1 (AUTHOR) Xiaotao.Shan@toshiba-bril.com, Jin, Yichao1 (AUTHOR), Jurt, Marius1 (AUTHOR), Li, Peizheng1 (AUTHOR)
Quelle: Robotics & Autonomous Systems. Aug2024, Vol. 178, pN.PAG-N.PAG. 1p.
Schlagwörter: *GREEDY algorithms, *DECISION making, *ROBOTS, *SCALABILITY
Abstract: Recent studies in warehouse logistics have highlighted the importance of multi-robot collaboration in collective transport scenarios, where multiple robots work together to lift and transport bulky and heavy items. However, limited attention has been given to task allocation in such scenarios, particularly when dealing with continuously arriving tasks and time constraints. In this paper, we propose a decentralized auction-based method to address this challenge. Our approach involves robots predicting the task choices of their peers, estimating the values and partnerships associated with multi-robot tasks, and ultimately determining their task choices and collaboration partners through an auction process. A unique "suggestion" mechanism is introduced to the auction process to mitigate the decision bias caused by the leader–follower mode inherent in typical auction-based methods. Additionally, an available time update mechanism is designed to prevent the accumulation of schedule deviations during the robots' operation process. Through extensive simulations, we demonstrate the superior performance and computational efficiency of the proposed algorithm compared to both the Agent-Based Sequential Greedy Algorithm and the Consensus-Based Time Table Algorithm, in both dynamic and static scenarios. • Proposes an auction-based method for task allocation in collective transport. • It manages continuously arriving tasks with time window constraints. • Introduces a suggestion mechanism to mitigate decision bias in leader–follower mode. • A state update mechanism accommodates unexpected schedule deviations. • Its efficacy and scalability in addressing large-scale issues is validated. [ABSTRACT FROM AUTHOR]
Datenbank: Academic Search Index
Beschreibung
Abstract:Recent studies in warehouse logistics have highlighted the importance of multi-robot collaboration in collective transport scenarios, where multiple robots work together to lift and transport bulky and heavy items. However, limited attention has been given to task allocation in such scenarios, particularly when dealing with continuously arriving tasks and time constraints. In this paper, we propose a decentralized auction-based method to address this challenge. Our approach involves robots predicting the task choices of their peers, estimating the values and partnerships associated with multi-robot tasks, and ultimately determining their task choices and collaboration partners through an auction process. A unique "suggestion" mechanism is introduced to the auction process to mitigate the decision bias caused by the leader–follower mode inherent in typical auction-based methods. Additionally, an available time update mechanism is designed to prevent the accumulation of schedule deviations during the robots' operation process. Through extensive simulations, we demonstrate the superior performance and computational efficiency of the proposed algorithm compared to both the Agent-Based Sequential Greedy Algorithm and the Consensus-Based Time Table Algorithm, in both dynamic and static scenarios. • Proposes an auction-based method for task allocation in collective transport. • It manages continuously arriving tasks with time window constraints. • Introduces a suggestion mechanism to mitigate decision bias in leader–follower mode. • A state update mechanism accommodates unexpected schedule deviations. • Its efficacy and scalability in addressing large-scale issues is validated. [ABSTRACT FROM AUTHOR]
ISSN:09218890
DOI:10.1016/j.robot.2024.104722