ECE grad student Hanxiang Zhang, left, and Assistant Professor of Electrical and Computer Engineering, Bayaner Arigong, place a chip prototype inside a soundproof chamber in Arigong’s lab at the FAMU-FSU College of Engineering. (M Wallheiser/FAMU-FSU Engineering)
FAMU-FSU Engineering researchers have developed a novel technique for real-time RF signal processing.
The technology boosts the computation speed needed for operations that involve artificial intelligence, wideband wireless communication, quantum computing and brain-to-machine interaction.
The new circuit system will revolutionize the speed of performing complex mathematical operations directly at the analog and electromagnetic waveform levels. Bayaner Arigong, an electrical and computer engineering assistant professor at the FAMU-FSU College of Engineering, advanced the novel design.
“My graduate student, Hanxiang Zhang, and I are working together to perform Hilbert transformations at multiple electromagnetic frequency points,” Arigong said. “What that means is we are moving the most complex operations to the realm of the electromagnetic analog waveform where we can reduce the need for complex systems to translate the information.”
The researchers developed a novel RF Hilbert transformer from a hybrid coupler and tunable transmission line so that its operating frequency can be configured from 1.9 GHz to 2.4 GHz with a simple control method. The device will enable real-time RF signal processing for everything from cordless cellphones to satellite communications systems in the wide spectrum.
“Emerging 5G, AI, and scientific computing require a lot of data and a network that processes a high volume of data with minimal resources,” Arigong said. “We are reducing the system’s complexity and energy consumption with this approach, and it will greatly boost the computation speed of the system.”
The available bandwidth from the technology requires high-end analog to digital conversion and high-speed data transmission with low power consumption. Conventional data signal processing needs to be improved because of excessive memory storage, costs, and consumption and other critical problems they have. Current analog signal processors are bulky, complex, and do not support wideband and frequency configuration.
But semiconductor technologies can limit the performance. There is an increasing demand for high computation and moving the load from digital to analog significantly reduces the effort, power and cost.
“Compared to conventional digital implementation, our system transforms RF signals and doesn’t need high-speed data conversion,” Arigong said. “We can reduce power consumption and accelerate the computation process for wideband and high-speed signals.”
The researchers are working to verify the design concept and develop a prototype circuit. They hope to create a fully integrated RF/analog signal processing chip that performs multiple complex operations directly at analog higher frequencies.
Their research was published in the IEEE Xplore research journal.
The National Science Foundation partially supports this project, and the team expects more funding to develop the chip further.
RELATED ARTICLES
New Grants for Advanced Research Equipment Will Enhance Training for Future Minority STEM Faculty
Engineering Researchers Discover New Algorithm for Faster Digital Signal Processing