The objective of this project is to improve the efficiency of random medium access control protocols. The approach is to employ efficient scheduling to amortize the high control overhead of medium access over a longer sequence of data frames. Fully distributed protocols will be developed to implement this approach.
Intellectual Merit: This proposal outlines a research and education plan focusing on the theory and system design of efficient random access protocols. The theoretical and algorithmic thrust includes: (i) developing a general analytical framework, (ii) exploiting multi-user diversity under fading channels, and (iii) extension to multi-hop and multi-channel wireless networks. The experimental thrust includes (i) formalizing the protocols and prototyping open source device drivers, and (ii) testbed experiments. The proposed approach has the potential of solving the low throughput problem many wireless networks suffer, as demonstrated in our preliminary studies. The proposed theoretical study will provide underpinning for the proposed protocol as well as existing standard components. The proposed field experiments will yield useful experimental experience and insights under a realistic wireless network setting.
Broader Impact: Integration of education and research is an important goal of this project. Graduate students will gain useful theoretical and experimental research experience from this project. The PI will actively involve undergraduate students by advising senior projects and participating in the NSF Research Experiences for Undergraduate program. Research outcomes will be incorporated into the curriculum of Auburn University?s Bachelor of Wireless Engineering program, the first Accreditation Board for Engineering and Technology-accredited program of its kind in the nation.
Despite recent advances in wireless technology, today’s wireless networks still cannot offer comparable data rates as their wired counterparts. In wireless networks, medium access control (MAC) protocols are the "workhorse" of the protocol stack, making it possible for higher layer protocols to savor the benefits of emerging physical layer technologies. In order to accommodate current and emerging bandwidth-intensive applications, it is thus crucial to improve the efficiency of wireless MAC protocols, while new physical layer technologies are being adopted for higher channel data rates. The objective of this project is to exploit polling services for random access protocol design, which have the potential of significantly improving the throughput, delay, energy consumption, and fairness performance of wireless networks. The main idea is to employ efficient scheduling to amortize the high control overhead of medium access over a longer sequence of data frames. Unlike the centralized approach taken in Bluetooth piconet and IEEE 802.11e Hybrid Coordination Function (HCF) Controlled Channel Access (HCCA), this project aim to develop fully distributed random access schemes, which are especially useful for multi-hop wireless networks within which no centralized entity exists. This project consisted of the research and education thrusts on the theory and system design of polling service-based random access protocols (termed PSMAC). The research tasks included a number of theoretical, algorithmic, and experimental components. In particular, the problem of improving the efficiency of IEEE 802.11-like MAC protocols was investigated. Motivated by insights from polling system theory, three polling service-based MAC schemes, termed PSMACs, were developed. The main idea was to serve multiple data frames after a successful contention resolution, thus amortizing the high control overhead and making the protocols more efficient. This idea was then extended to the case of multi-channel wire-less local area networks. This project also explored location information to improve the efficiency of wireless MAC protocols, and investigated the problem of MAC protocol design in WLANs with downlink multi-user (MU) MIMO capability. In addition to the above studies that are focused on medium access control, this project also looked into a broader range of research problems, including cross-layer optimization for video over wireless networks (i.e., using variable-bit-rate (VBR) video as a band-width/spectrum hungry application to capitalize the improved spectrum efficiency). Complementary to the theoretical studies, the PSMAC protocol was formalized while the challenging issues for practical protocol design and deployment were addressed. A proof-of-concept implementation was developed for the distributed version of PSMAC, i.e., PSMAC 2, on the GNU Radio and Universal Software Radio Peripheral (USRP) platform. Various design considerations and challenges of prototyping PSMAC 2 were examined, and extensive experimental studies were carried out with the GNU Radio/USRP PSMAC testbed. This research is expected to have the potential of greatly improving the performance for a broad range of wireless networks. Outcomes from this project have been disseminated through conference and journal publications in various conferences and journals as well as the project website. Graduate students working on this project have gained substantial theoretical and experimental research experience, with focus on cross-layer design of wireless networks. The PI actively involved undergraduate students in this project by participating in the NSF Research Experiences for Undergraduate (REU) program. Outcomes from this project have been used to enhance the wireless curriculum at Auburn University. Furthermore, the PI was an active member of the NSF I/UCRC Wireless Internet Center for Advanced Technology (WICAT) site at Auburn University and fostered successful collaboration with the industry. The PI also collaborated with a faculty member with Tuskegee University, one of the Historically Black Colleges and Universities (HBCU), on research, curriculum development, and recruiting African American graduate students.
