-
Demonstration of training and operation of artificial intelligence signal processing models inside a 6G physical layer
-
AI models are trained with simulated and real-world data
-
Operation of models are validated in an over-the-air system
Keysight Technologies, Inc. has collaborated with mobile, video and AI technology research and development company InterDigital Inc. to show how Artificial Intelligence (AI) channel estimators can be trained, integrated into a physical layer, and evaluated with an over-the-air demonstration at Mobile World Congress Barcelona 2024 (MWC24). Presented at Keysight’s booth, Hall 5 Stand 5E12, the demonstration will feature a 144 GHz communications link that uses AI models for channel estimation developed by InterDigital and Keysight.
As AI technology matures, it is being integrated into wireless communications systems and will be a foundational technology for 6G. One of the applications of AI in wireless is channel estimation. When the channel estimation block is replaced with a neural network (NN), site-specific training data used for the NN can optimize its performance and robustness to new use cases.
“Through partnership with Keysight, we are able to extend our ability to validate AI solutions for wireless networks. These types of collaborations and their outcomes provide valuable insight and build industry alignment as we work to shape how AI will be used in both 5G and 6G.”
Recommended : Key Insights Into The Utility Of AI In The College Application Process
The demonstration uses Keysight’s communications system-specific Electronic Design Automation (EDA) software to model a 3GPP physical layer and wireless channels. Instead of using the traditional channel estimation signal processing, this block has been replaced with various neural networks for channel estimation. The first NN was fully developed and trained by InterDigital and delivered to Keysight. The second NN was developed by Keysight and trained in two different ways. First using a combination of data generated by Keysight’s Electronic Design Automation System (SystemVue) software and then again using real-world measurement data. The demonstration compares performance and complexity of InterDigital and Keysight’s Machine Learning approaches to that of traditional channel estimation, and uses an EDA software environment and a real-world 144 GHz over-the-air environment.
Being able to benchmark the performance of NN and compare them to traditional methods as well as other NN is critical for the advancement of AI in 6G. Without these sorts of measurements and comparisons, the true performance gains (or losses) of AI cannot be known and the adoption of the technology will be hindered.
Milind Kulkarni, Vice President and Head of Wireless Labs at InterDigital, said: “Through partnership with Keysight, we are able to extend our ability to validate AI solutions for wireless networks. These types of collaborations and their outcomes provide valuable insight and build industry alignment as we work to shape how AI will be used in both 5G and 6G.”
Giampaolo Tardioli, Vice President, 6G and Next Generation Technology at Keysight, said: “The world is embarking on a transformative journey to revolutionize the future of connectivity with the use of AI in wireless communications. Our collaboration with strategic partners like InterDigital enables us to hone our solutions and capabilities to meet the needs of our customers as they continue pushing the boundaries of what is possible when AI and wireless are combined.”
This demonstration includes technology that was developed as part of CENTRIC and 6G-SANDBOX, projects that have received funding from the Smart Networks and Services Joint Undertaking (SNS JU) under the European Union’s Horizon Europe research and innovation program under Grant Agreement No 101096379 & 101096328 respectively.
Latest HRtech Interview Insights : HRTech Interview With Lavonne Monroe, VP Of Global Talent Acquisition And Onboarding At HPE
[To share your insights with us, please write to pghosh@itechseries.com ]