Beyond 5G AI, Terahertz, And Open RAN Transform RF Engineering

By John Oncea, Editor

RF engineers are advancing 6G through terahertz communications, AI-driven network optimization, and Open RAN architectures, with major university research centers and NSF funding driving innovation toward 2030 commercial deployment.

The RF engineering landscape is undergoing unprecedented transformation as researchers, industry leaders, and government agencies collaborate to define the next generation of wireless communications. From groundbreaking terahertz research to artificial intelligence integration and open network architectures, the field is advancing on multiple fronts simultaneously. For RF engineers, these developments represent both technical challenges and exciting opportunities to shape the wireless infrastructure of tomorrow.

6G Research Gains Momentum At Leading Universities

The race to develop sixth-generation wireless technology is well underway at major research institutions across the United States. The University of Texas at Austin launched the 6G@UT research center, bringing together faculty from the Cockrell School’s Wireless Networking & Communications Group with industry partners including Samsung, AT&T, NVIDIA, Qualcomm, and InterDigital. According to 6G@UT Director Jeffrey Andrews, the center focuses on developing wireless-specific machine learning algorithms, advanced sensing technologies, and core networking innovations that will form the backbone of 6G systems.

NYU WIRELESS continues to push boundaries in 5G and beyond wireless research, with particular emphasis on millimeter-wave and terahertz communications. The research center brings together over 25 faculty members and industrial partners to advance the fundamental technologies needed for next-generation networks.

Meanwhile, Northeastern University’s Institute for the Wireless Internet of Things received NSF funding to build programmable cellular platforms for testing 5G and 6G technologies. This infrastructure enables researchers to experiment with software-defined networks and higher frequency radio bands, departing from traditional hardware-dependent platforms.

These academic efforts align with a broader NSF initiative that brought together government, industry, and academia leaders in April 2023 to discuss the future of 6G technology. The workshop, titled “6G: Open and Resilient by Design,” featured speakers including NSF Director Sethuraman Panchanathan and Deputy National Security Advisor Anne Neuberger. The NSF has established public-private partnerships, including the $40 million Resilient and Intelligent Next-Generation Systems (RINGS) program, which accelerates research across multiple areas critical to next-generation networking and computing systems.

Terahertz Communications Break New Ground

One of the most exciting frontiers in 6G research involves terahertz frequencies, which promise unprecedented bandwidth for ultra-high-speed wireless communications. Researchers at Northeastern University, working with NASA and the U.S. Air Force, achieved a breakthrough by establishing a 2-kilometer terahertz link, the longest ever demonstrated on Earth. Professor Josep Jornet’s team overcame significant technical challenges by developing a novel approach that feeds information directly into the signal source, rather than using traditional mixer components that would fail at such high-power levels.

Research at Brown University and Rice University introduced another innovation: terahertz beams that can curve around obstacles rather than being blocked by them. This development addresses one of the fundamental challenges of high-frequency communications, where solid objects like walls and furniture typically disrupt signal transmission. By creating self-accelerating beams with carefully engineered electromagnetic wave patterns, the researchers demonstrated the world’s first curved data link, a critical milestone for making 6G systems practical in real-world environments.

The technical challenges of terahertz communications remain substantial. Research at the University of Missouri-Kansas City focuses on developing leaky-wave antennas optimized for the terahertz range, which offer engineering flexibility and can create highly directional beams with radiation patterns that vary by frequency. This characteristic enables multiple applications, including frequency-division multiplexing, sensing, localization, and imaging capabilities that will be essential for 6G networks.

Academic research funded by NSF has identified mobile terahertz communications as a pressing research target for 6G and even 7G systems. The challenge lies in the fact that mobile terahertz wireless systems will often operate in the near-field due to the large electrical size of high-gain antenna systems required at these frequencies, complicating the transition from laboratory demonstrations to practical mobile networks.

Artificial Intelligence Transforms RF Engineering Workflows

Machine learning and deep learning are revolutionizing how RF engineers approach design, optimization, and troubleshooting. Research at MIT demonstrated that machine learning techniques can improve RF integrated circuit simulation accuracy by 98%, dramatically reducing the need for multiple design-fabrication iterations. This improvement translates to significant time and cost savings in RFIC product development.

The MIT RF Challenge exemplifies the convergence of machine learning and RF engineering, providing datasets and benchmarks specifically designed for radiofrequency signal detection, identification, and geolocation tasks. While groundbreaking results have been achieved in the past decade for natural signals like images and audio, RF signal processing with machine learning is still emerging as a research area.

Work at Northeastern University analyzed a massive 400 GB dataset of in-phase and quadrature signal data transmitted by 10,000 radios to demonstrate how deep convolutional neural networks can identify devices through RF fingerprinting under various practical conditions. This research addresses scalability issues with very large device populations and investigates techniques like partial equalization to mitigate wireless channel effects.

Research at UC Santa Barbara focuses on learning RF signatures that distinguish between devices sending identical messages, exploiting subtle hardware imperfections unique to each transmitter. By employing complex-valued convolutional neural networks, researchers can achieve robust wireless fingerprinting even when using only signal preambles, significantly enhancing security against spoofing attacks.

Studies at Virginia Tech have established foundational knowledge for transfer learning in RF contexts, addressing the inflexibility of traditional supervised learning solutions that remain fixed after deployment. This research provides methods for predicting and quantifying RF transfer learning performance, improving the robustness and deployability of RF machine learning systems in dynamic communication environments.

Open RAN Architecture Drives Network Innovation

Open Radio Access Network technology represents a change in thinking from proprietary, vendor-specific implementations to standardized, interoperable systems. According to research presented at ACM MobiCom 2024, Open RAN replaces traditional RAN architecture with disaggregated functions running on general-purpose hardware, supporting equipment interoperability from multiple vendors. This openness enables rapid innovation in service automation, network visibility, intelligent control, and security.

The technology has gained significant traction with major carriers. Information from academic workshops indicates AT&T has set a goal of 70% of its RAN traffic operating on open platforms by 2026, while the U.K. government aims for 35% of network traffic over Open RAN by 2030. Vodafone similarly targets 30% of its European networks transitioning to Open RAN by the same period.

The NSF has invested $7 million in additional funding for the Platforms for Advanced Wireless Research (PAWR) program to augment testbed capacities for Open RAN testing and validation. Multiple PAWR platforms have gained O-RAN ALLIANCE approval as Open Testing and Integration Centers (OTICs), enabling comprehensive testing of next-generation wireless networks.

Northeastern University’s Open6G OTIC announced general availability of testing and integration solutions for Open RAN in May 2024, including conformance, interoperability, and end-to-end testing based on O-RAN ALLIANCE specifications. The facility supports end-to-end full-stack testing using both emulated and real-world environments, leveraging Digital Twinning technology and the Colosseum wireless network emulator.

Research at Northeastern’s WINES Lab has demonstrated large-scale integration of O-RAN-compliant software components with open-source full-stack softwarized cellular networks. Experiments on Colosseum showed closed-loop integration of real-time analytics and control through deep reinforcement learning agents, optimizing scheduling policies for co-existing network slices.

Virginia Tech researchers received $4 million from the Innovation Fund to develop O-RAN security and performance testing frameworks. One project focuses on holistic cybersecurity testing for 5G Open RAN, addressing vulnerabilities through multiple testing methods that uncover security issues highlighted by radio interface attacks and other threat vectors.

Government Investment And Industry Collaboration

Federal support for advanced wireless research has accelerated significantly in recent years. The NSF’s Directorate for Engineering issued guidance encouraging research proposals related to advanced wireless as an emerging industry, highlighting areas including novel devices and circuits for terahertz communications, integrated wireless transceivers, and wireless technologies for manufacturing.

NSF’s broader communications and wireless initiative coordinates with other federal agencies, industry, and nonprofits to share data, tools, and expertise while strengthening workforce development. The Spectrum Innovation Initiative supports research enabling fast, accurate, and dynamic coordination of limited spectrum resources, recognizing that the electromagnetic radio spectrum must be shared across wireless systems and applications, including mobile connectivity, air traffic control, weather prediction, and research.

The Future of Semiconductors (FuSe) program announced 24 research and education projects with a total investment of $45.6 million, including funding from the CHIPS and Science Act of 2022. This public-private partnership with Ericsson, IBM, Intel, and Samsung addresses the national need for a reliable, secure supply of innovative semiconductor technologies essential for next-generation communications systems.

The Path To 2030 And Beyond

While commercial 6G deployment is not expected until approximately 2030, the foundational research happening today will determine the capabilities and characteristics of these future networks. Academic research highlights that 6G will incorporate technologies that matured during the past decade: ubiquitous sensing, machine learning, and higher frequency spectrum at millimeter-wave and terahertz bands.

The convergence of artificial intelligence with wireless communications represents a particularly significant shift. Research published in IEEE Communications Magazine discusses how reconfigurable intelligent surfaces – programmable metasurfaces that can dynamically control electromagnetic wave propagation – will enable new applications and improve network performance in 6G systems.

For RF engineers, these developments demand continuous learning and adaptation. The field is moving from traditional hardware-centric approaches toward software-defined, AI-enabled systems with unprecedented flexibility and capability. Understanding terahertz propagation, machine learning algorithms, open interfaces, and security frameworks will become essential competencies alongside classical RF engineering knowledge.

The collaborative nature of current 6G research – bringing together universities, government agencies, and industry partners – reflects the complexity and importance of building the wireless infrastructure of the future. As these technologies mature and transition from laboratory demonstrations to commercial products, RF engineers will play a central role in turning ambitious research visions into practical, deployable systems that connect billions of devices and enable applications we are only beginning to imagine.