We are thrilled to share that our Smart Agriculture testbed has made substantial strides in recent months. Following the successful completion of the initial phase in February, where we achieved a significant milestone with our drone prototype and established a robust communication link with the tractor in a simulated environment, we have now entered the crucial second phase.

Real-World Deployment Begins

As part of this pivotal stage, we have begun deploying the two-node system in real-world scenarios. Extensive testing in the field is currently happening, with our primary emphasis on establishing real-time communication between the drone and the tractor. This phase is essential for gathering the necessary data to train our MLSysOps models effectively.

Data Collection and Analysis

Data collection is at the heart of this phase, enabling us to fine-tune our models for optimal performance. The real-time data gathered from the field is invaluable in enhancing the precision and efficiency of our agricultural operations. This step not only marks a substantial leap for our testbed but also sets the stage for the project’s concluding year, where we aim to fully engage MLSysOps for real-time application optimization.

Preparing for the Final Phase

With the successful deployment and ongoing data collection, we are laying a solid foundation for the upcoming third and final phase. The insights gained from this real-world testing will be instrumental in ensuring that our system operates seamlessly and effectively in live agricultural environments.

The advancements made thus far are a testament to the hard work and dedication of our team, and we are excited about the possibilities that lie ahead.

Stay tuned for more updates!

On October 14th, during the Open Innovation Week organized by NTT DATA Italia, the MLSysOps project was showcased in a presentation focused on the potential of 5G Standalone (SA) technology adoption in the enterprise market.

The event brought together representatives from leading companies, including Mediaset, Fastweb, Sky Italia, ACEA, IP Gruppo api, BNL BNP Paribas, and Crédit Agricole Bank Polska. Attendees also included a prominent mobile telecommunications operator in Italy, one of the country’s primary tower operators, a major international biomedical laser company, and a key player in the Aerospace, Defense, and Security sectors.

Additionally, members of the academic community, such as those from Università del Salento, were present, fostering dialogue between industry and academia on the transformative role of 5G technology.

From 20 to 22 November, NTT DATA Italia participated in the 6th International Conference on Industry 4.0 and Smart Manufacturing (ISM) in Prague, Czech Republic. At the conference, NTT DATA Italia presented a paper titled “Optimizing 5G Networks for Low-Latency Communication: A Reinforcement Learning-Based Approach“, detailing the initial results of their joint research with the University of Calabria related to MLSysOps project.

The study focused on applying reinforcement learning (RL) algorithms to minimize latency in 5G networks, with a particular emphasis on edge selection in urban areas. Recognizing the critical role of latency reduction in enhancing network efficiency, the research leveraged real-time data processing and optimization strategies. The paper evaluated several RL algorithms to determine the most effective approach for latency optimization. The results indicated that the Deep Q-Network algorithm achieved superior performance with stable convergence, underscoring the importance of algorithmic selection and hyperparameter tuning in complex network optimization tasks. Therefore, the study highlighted the potential for future research to explore alternative configurations aimed at further optimizing 5G network performance and promoting sustainability.

The conference provided a valuable platform for NTT DATA Italia to share its innovative work furthering the development of next-generation network technologies for MLSysOps.

Copyright © 2024, MLSysOps EU Project, All rights reserved.