Nutaq has the pleasure of sharing with you how Georgios and Emrecan, along with their research team led by professor Dimitris Pados, engage in this project. They now have their PicoSDR development tools in their hands and have accelerated the development process thanks to Nutaq’s Model-Based Design Kit (MBDK) software integration tools, which provide a unique advantage for researchers who desire to exploit powerful FPGA resources without needing a PhD in VHDL coding.
Each section is preceded by the evaluation criteria used for evaluating each submission, along with its corresponding weighting out of 100 pts. We hope you enjoy reading this exciting project proposal!
Project title: Spectral Efficiency Maximization in Heterogeneous Wireless Systems: A New Approach
Proposed Project (15 pts)
Describe the proposed project and explain why it is innovative. Throughout the evaluation of the project description, the independent technical committee seeks a global understanding of the proposed project, its innovative aspects, and how it pushes the boundaries of modern technology.
The objective of this project is to study, implement, and demonstrate a new approach for spectral efficiency maximization in future heterogeneous wireless systems by exploring the agility offered by real-time reconfigurable wireless radios platforms in joint signal waveform allocation and routing. We consider a network of multiple-input-multiple-output (MIMO) antenna users and propose a new spread-spectrum management paradigm, where digital waveforms are designed to occupy the entire available spectrum, and adaptively track the interference profile at the receiver to maximize the link capacity. Preliminary extensive simulation results for the single-antenna case have already demonstrated significant performance gains in terms of network throughput, comparing to relevant baseline solutions. Moreover, we propose a flexible, cross-layer system architecture design for enabling the implementation of such reconfigurable wireless radios with self-optimization capabilities that can employ the aforementioned algorithmic developments in real-time and further benefit network spectral efficiency in practical setups. Successful testbed deployments promise a ten-fold (or more) improvement of spectral efficiency by taking advantage of both novel cross-layer theoretical developments in the three lowest layers of the protocol stack and system architecture abstractions offered by our proposal.
Features of the PicoSDR 2×2 (50 pts)
Identify features of the PicoSDR-2×2 that are essential to the project. This section will receive a lot of consideration; the more each feature is used to its full potential, the higher the score. On the software side, the independent technical committee will be focusing on the use of the model-based design kit (MBDK) and the use of GNU Radio, both of which are proven to substantially reduce development time. The SDR platform provided to the winning team as part of the grand prize is 2×2 MIMO, a large FPGA (Virtex-6 SX315T), and is equipped with a rear PCI Express connector. The independent technical committee will be looking for the use of Nutaq’s PCI Express drivers, the substantial use of FPGA resources, and the use of 2×2 MIMO in the proposed projects.
GNU Radio/C API support and PCI express drivers are two important features of PicoSDR, that we will take advantage during our preliminary implementation with GNU Radio on a host PC. Custom C++ software signal processing blocks will be developed and used with GNU Radio to drive the PicoSDR. We will exploit PCI express drivers to overcome the GigE interconnection bottleneck between the SDR platform and the host PC, currently existing in other commercial platforms, and upper bounds the system throughput, and therefore the system’s overall spectral efficiency. The MBDK will help us rapidly embed our cross-layer design into the Virtex-6 FPGA and thus enable higher system performance and throughput. Towards that end, we plan to come up with an abstract design structure that depends both on hardware (Virtex-6) and software (Quad Core i7 CPU) tools, where signal processing functionalities and waveform selection algorithms will be handled respectively. More specifically, we will take advantage of the capabilities offered by System Generator (SysGen) to rapidly target the implementation of computational intensive digital signal processing algorithms (e.g. Singular-Value-Decomposition) to the FPGA, as well as the HDL code reusability offered by SysGen’s primitive blocks. MBDK’s co-simulation features and Real-Time Data Exchange tool will enable full-duplex data transfer between hardware and software on-the-fly. As a result our system will be able to simultaneously drive waveform selection decisions from software to hardware, while collected environmental and received content data are passed from hardware to software in real-time. In this way, we shall offload the intensive processing carried out by the host PC and exploit the valuable FPGA resources. The MIMO-OFDM reference design will serve as a great starting basis and tutorial for our system implementation while the 2×2-MIMO will be used to prove the proposed concept for efficient space, time, and spectrum utilization.
Project Outcome (15 pts)
Describe how the outcome of the research project will be shared with the community. Previous projects can be used as examples to describe how the group typically shares the outcome of their work (e.g. though publications or industry/government partnerships). The independent technical committee will be looking for the team’s previous participation in journal papers, conference papers, blogs, and industry whitepapers to determine the potential impact the proposed project could have on modern technology and ensure the outcome will be properly conveyed.
The research data collected for this project shall be archived in a digital format in multiple GIT repositories in a local fileserver. The purpose of GIT is to manage the set of files in each project as they change over the time, therefore avoiding possible overwrite or delete of important data. The data will be backed up nightly and the fileserver will be operating under the supervision of UB EE Department. The proposed activity findings shall be presented in top-rated international journals and conferences such as IEEE Transactions on Wireless Communications, IEEE/ACM Transactions on Networking, IEEE GLOBECOM, IEEE International Conference on Communications (ICC), ACM International Conference on Mobile Computing and Networking (MobiCom), IEEE International Conference on Computer Communications (INFOCOM) and other relevant periodicals. Moreover, graduate theses shall be published from the results of the proposed project, which shall be accessible to the public through the University’s e-library archive. Our team is supported by a variety of federal funding sources such as National Science Foundation (NSF), The Office of Naval Research, and Air Force Research Laboratory. The team members are also active members of online SDR blogs and mailing lists, and follow the latest updates regarding technical meet-ups and conferences. An upcoming presentation in the GNU Radio Conference 2014 and a demonstration in the 2014 NATO IST-123 Symposium on “Cognitive Radio and Future Networks” will discuss progress results regarding the proposed project. Associated selected publications discussing preliminary results of the proposed activity are “All-spectrum cognitive networking through joint distributed channelization and routing”, to appear in IEEE Transactions on Wireless Communications, “Cognitive code-division channelization”, in IEEE Transactions on Wireless Communications, April 2011, and “Cross-layer routing and dynamic spectrum allocation in Cognitive Radio Ad Hoc networks”, in IEEE Transaction on Vehicular Technology, May 2010. The team’s research efforts towards advancing cognitive wireless communications have been also recently blogged by WIRED.com (http://www.wired.com/2014/05/air-force-cognitive/).
Team Introduction (20 pts)
Introduce the team working on the proposed research project. In this part, the independent technical committee seeks information on the team structure. The lead professor, research assistants, Ph.D. students, and any resources to be involved in the project’s execution can be listed to help highlight the team’s capabilities to execute the proposed project. The name of the legal entity receiving the prize (hardware and software licences) must be stated.
Professor Dimitris A. Pados, the Director of the Signals, Communications, and Networking Research Group, will lead the proposed research project and will be the legal entity receiving the prize of this competition. Professor Stella N. Batalama, Electrical Engineering Department Chair and Associate Professor Tommaso Melodia, the Director of the Wireless Networks and Embedded Systems Laboratory, will also be part of the team working towards the proposed project. The research assistants leading the project will be Emrecan Demirors, senior Ph.D. student working under Professor Melodia, and Georgios Sklivanitis, senior Ph.D. student working under Professor Pados and Professor Batalama. Several master students will also be involved in both theoretical and practical assignments of the project. The research team is hosted by state-of-the art facilities at the University at Buffalo that offer simulation, testing, and early prototyping capabilities. Particularly, the team has access to the following: 20 software-defined radio devices (USRP2, and USRPN-210), 1 ZYNQ-7000 EPP ZC02 Xilinx Evaluation Kit, 1 Nutaq ZeptoSDR, 1 Agilent CXA Signal Analyzer N9000A from 9kHz-7.5GHz, 1 Agilent DSOX-2004A Oscilloscope up to 70MHz (4 channels), 2 Digilent ATLYS comprehensive Spartan-6 Design platforms, and licensed software tools such as Matlab/Simulink and Xilinx design tools (System Edition).
We sincerely hope you enjoyed reading this project proposal as much as we did. It is certainly motivating to see what the upcoming developments will be in wireless technologies for the next few years. We want to thank every member of the Signals, Communications, and Networking Research Group for letting us share this proposal with you, and Georgios Slkivanitis and his Professor Dimitris Pados, specially.