Swarm Learning in Manufacturing
Our MODUL4R colleague, Netcompany - Intrasoft, introduces a novel approach in manufacturing by utilizing the advantages of swarm learning at the edge, combined with supervised and incremental learning to enhance modularity, scalability and data preservation.
Within the scope of MODUL4R EU Project they exploit swarm learning at the edge across various PCB assembly operations. A supervised and incremental learning model is deployed at each robot. The model undergoes continuous training, dynamically adapting as it processes new data generated by the robot’s ongoing activities. Subsequently, this model provides timely and precise recommendations for optimizing the grabbing and cutting processes, directly assisting the user.
These optimized recommendations remain robust, even if the final product changes. As new data are incorporated into the training cycle, the model swiftly adjusts, refining its recommendations until it can deliver the optimal suggestions. Additionally, the flexibility of the swarm network, which can support an adaptable number of nodes, significantly enhances the scalability of the platform.
Georgios Bardas, Netcompany-Intrasoft. (2024, June 19). Swarm Learning in Manufacturing. LinkedIn. Retrieved from https://www.linkedin.com/pulse/swarm-learning-manufacturing-georgios-bardas-7jbnf/?trackingId=yOOChk8eSQSZdY%2FL5UprHQ%3D%3D .
