The anticipated opening up of white space by the FCC for unlicensed use has created exciting new opportunities. It also presents unique challenges: (i) it requires accurate wide-band spectrum sensing technique to avoid interfering with primary users, and (ii) it needs dynamic spectrum sharing in heterogeneous setting both within networks and across networks.
This project develops a holistic approach for spectrum sensing to address spatio-temporal variability and spectrum sharing among multiple secondary users and networks in white space. In particular, it first develops efficient and accurate spectrum sensing techniques that simultaneously exploit time- and frequency-domain features, sparsity in active transmitters, and temporal and spatial locality in the transmissions. Second, it studies the cost and benefits of sharing network state in a heterogeneous network where different nodes have different information views. Third, it studies spectrum sharing across networks (e.g., CSMA/MaxWeight, TDMA/CSMA co-located networks). It develops algorithms and studies performance and fairness both with implicit sharing (i.e., no state-exchange) and with explicit sharing of information across networks. Fourth, it proposes a ground-up design for white space that leverages random CDMA-like codes for spectrum sharing with an OFDM physical layer that leads to a new efficient architecture for sharing across users and networks in white space.
The algorithms, techniques, and software resulting from this research will help enable effective communication in white space, create new wireless network technologies, and deepen our understanding of wireless networks. The research results will be incorporated into networking courses and widely disseminated through conference/journal publications and software distribution.
As wireless carriers are running out of spectrum to accommodate users' ever increasing hunger for network bandwidth, new technologies have to be developed to increase spectrum efficiency. This project developed a series of novel spectrum management approaches. Our system side contributions are as follow: - Dynamic spectrum access (DSA) is one of the most promising solutions to address spectrum crisis. An important challenge in DSA is how can a sender and receiver agree on the same channel in order to establish communication. One option to enable this is to coordinate the spectrum usage before communication (e.g., using RTS/CTS exchange or control channel). Such coordination incurs significant overhead. Moreover, RTS/CTS exchange may not be possible without a shared channel between the sender and receiver. We developed a Maximum Likelihood (ML) detection algorithm that detects the spectrum of the sender in-band using the IEEE 802.11 preamble. In addition, we developed a fine-grained spectrum access design that allows a sender and receiver to change their transmission and reception bands on a per-packet basis, and an effective spectrum adaptation algorithm that determines which spectrum should be used for each transmission by taking into account frequency diversity and interference. This is the first per-packet spectrum adaptation prototype for WiFi. - We proposed a spectrum sharing protocol -- Collision-Resistant Multiple Access (CRMA). It allows multiple transmitters to efficiently share the spectrum among multiple users without fine-grained coordination. In CRMA, each transmitter views the OFDM physical layer as multiple orthogonal but sharable channels, and independently selects a few channels for transmission. The transmissions that share the same channel naturally add up in the air. The receiver extracts the received signals from all the channels and efficiently decodes the transmissions by solving a simple linear system. This work was a finalist for best paper award at MobiCom'11. - Motivated by increasing popularity of wideband communication, we proposed a series of techniques to explicitly harness such frequency diversity. The Channel State Information (CSI) captures the SNR on each subcarrier. We use CSI to (i) map symbols to subcarriers according to SNR of each subcarrier and the importance of each symbol, (ii) effectively recover partially corrupted FEC groups and facilitate FEC decoding, and (iii) develop MAC-layer FEC to offer different degrees of protection to the symbols according to their importance and error rates at the PHY layer. We further developed a rate adaptation approach that works together with these optimization schemes. Our theoretical contributions are as follow: - We studied spectrum auction as a way to support efficient sharing of spare spectrum. Spectrum auction is fundamentally different from traditional auction problems in that (i) spectrum can be re-used and (ii) not everyone competes with each other; instead, the competition pattern is dictated by wireless interference and can be very complicated. We proposed the first double auction design for spectrum allocation that explicitly decouples the spectrum buyer and seller side auction design while achieving (i) truthfulness (i.e., bidders cannot benefit from bidding differently from their true valuation), (ii) individual rationality (i.e., bidders get non-negative utilities, that is, sellers are paid no less than their asks and buyers do not pay more than their bids), and (iii) budget balance (i.e., the total amount paid to the sellers is no more than the total amount received from the buyers so that an auctioneer does not lose money). The algorithms, techniques, and software resulting from this research help enable effective wireless communication, create new wireless network technologies, and deepen our understanding of wireless networks. The research results have been incorporated into networking courses and widely disseminated through conference/journal publications.
