Tidd, Matthew, Dubay, Rickey et Cao, Hung.
2025.
« A fuzzy inference-based architecture for allocating multi-robot tasks ».
In Proceedings of the CSME-CFDSC-CSR 2025 International Congress (Montreal, QC, Canada, May 25-28, 2025)
Coll. « Progress in Canadian Mechanical Engineering », vol. 8.
Prévisualisation |
PDF
382 - A fuzzy inference-based architectu.pdf - Version publiée Licence d'utilisation : Tous les droits réservés aux détenteurs du droit d'auteur. Télécharger (396kB) | Prévisualisation |
Résumé
Multi-robot systems (MRS) are becoming more popular due to advancements in artificial intelligence (AI), as well as their ability to resolve task complexity. One of the most challenging problems with the use of MRS is the allocation of tasks to robots. This paper proposes a methodology for structuring bid formulation within distributed, auction-based Multi-Robot Task Allocation (MRTA). A fuzzy inference system (FIS) is proposed as a means of evaluating the suitability of robots for a given task by considering objective factors such as their load history, distance to the task, and the total distance travelled. The suitability of a robot is combined with its capability and put forth as a bid in the auctioning process. An adaptive neuro-fuzzy inference system (ANFIS) was developed as a custom implementation within Keras to extract a higher dimensional relationship between the input variables and to leverage parallelized computation. The previously designed FIS was used as a data generator, and the ANFIS was trained to perform regression between the objective parameters of a robot and its suitability. The results from this were a 9.8x decrease in inference time, down from 0.315 seconds to 0.032 seconds. Future work involves exploring using an artificial neural network (ANN) to better approximate the FIS while maintaining the parallelized computation, as well as deploying and evaluating the methodology in simulation and on real robotic agents.
| Type de document: | Compte rendu de conférence |
|---|---|
| Éditeurs: | Éditeurs ORCID Hof, Lucas A. NON SPÉCIFIÉ Di Labbio, Giuseppe NON SPÉCIFIÉ Tahan, Antoine NON SPÉCIFIÉ Sanjosé, Marlène NON SPÉCIFIÉ Lalonde, Sébastien NON SPÉCIFIÉ Demarquette, Nicole R. NON SPÉCIFIÉ |
| Date de dépôt: | 18 déc. 2025 15:33 |
| Dernière modification: | 18 déc. 2025 15:33 |
| URI: | https://espace2.etsmtl.ca/id/eprint/32511 |
Actions (Authentification requise)
 |
Dernière vérification avant le dépôt |