August 24, 2011-
Flexible and reconfigurable wireless systems are the future of telecommunications, according to last week’s SURF-IT seminar speaker, who gave his CALIT2 audience a sneak peek at some innovative offerings under development.
To optimize often-shrinking resources, said Konstantin Nikitopoulos, a postdoctoral researcher in the Mobile Ad hoc Networks Lab, tomorrow’s telecommunications platforms will feature software-defined radio and MANETs – ad hoc mobile networks – that operate without requiring pre-existing infrastructure.
Mobile wireless users want multiple applications, which often require different parameters. For example, communications and mobile multimedia processing require real-time delivery, while a surveillance photo would require more intricate detail. “Different applications, different trade-offs. We need devices that think,” Nikitopoulos said.
Interference, multi-path and noise constraints are dynamic, meaning they constantly change, so smart devices should be able to analyze situations and find optimal communication protocols and frequency band transmission modes.
One way to meet these challenges is through software-defined radio (SDR), a system which utilizes algorithms and parameters that are typically built in as dedicated hardware. Instead, in SDR, they are implemented as choices in software running on applications or embedded on general computational devices – thereby allowing devices to quickly adapt to a variety of networks and situations.
Optimal conditions, however, are different for different situations, Nikitopoulos said, so device systems must solve multiple optimization problems by reconfiguring and constantly adapting to changing mobile environments.
These software-defined radios leverage three driving forces: Moore’s law, which defines rapid advances in technology; machine learning and advanced optimization theory (machines should be able to learn); and communication/information theory – the ability to make decisions. “This is what we expect from our future telecommunication systems,” he said.
In all cases, flexibility is a key attribute.
Radio/system flexibility does not refer to just one idea; it’s an umbrella term that includes the ability to adapt and reconfigure, make dynamic adjustments, reuse components and be scalable. Radio flexibility will support more applications and standards, reduce engineering costs, allow access to a dynamic spectrum, and optimize performance to meet user needs.
Nikitopoulos shows students where it all happens.
Nikitopoulos said many people think of MANETs as the “holy grail of future technology applications.” Mobile nodes permit dynamic topology; they operate independently of base-system infrastructure; and they support multiple applications, including medical, emergency and rescue, sensor networks and military scenarios. “So MANETs describe all the problems we talked about,” he concluded.
He also discussed the Universal-Software-Radio-Peripheral (USRP) platform, which powers existing MANET architecture, calling it “a way to enable rapid design and prototyping of flexible software-defined radio systems.”
Nikitopoulos concluded his presentation by recapping: future wireless telecommunications systems will be highly complex and flexible; MANETs are rapidly-deployable, self-configuring systems that allow stand-alone communication; software-defined radio is a promising way to implement flexible wireless communications systems; and USRPs enable rapid design and prototyping of these systems. In addition, there is a fundamental trade-off between flexibility and implementation efficiency.
Finally, design and implementation of software-defined radios is an interdisciplinary task with inherent challenges that apply to all high-performance flexible systems. “We should be able to find solutions to all the challenges,” he said. “This is something we all need to work together on.”
Following his presentation, Nikitopoulos led interested audience members on a tour of his lab, demonstrating the software-defined radios and answering additional questions. He works in the lab with electrical engineering and computer science professors Hamid Jafarkhani and Homayoun Yousefi’zadeh, and this year’s undergraduate SURF-IT Fellow David Ho.
— Anna Lynn Spitzer
