Agile Spectrum-Sharing Wireless Systems using Software-Defined Radios and Reconfigurable Antennas

Since the radio spectrum is a scarce resource, to satisfy the future demand for wireless capacity, more efficient methods to utilize the spectrum are needed. The recent trend of migrating radio devices from ASIC hardware to a more software-based nature enables a host of new flexible applications. In addition, modern RF frontend and antenna technologies have matured in terms of processing bandwidth, frequency coverage, and reconfigurability to enable emerging capabilities, such as wideband sensing, beamforming, and adaptive transmission. These advances constitute significant system design elements that are still underutilized in practice for enhancing spectrum use and network capacity. In this research work, we propose to leverage these new system design elements to realize effective MAC/PHY cross-layer designs that can substantially increase wireless network capacity. Specifically, we address three system design issues: i) enhancing software radios for timing and frequency agility in spectrum sharing scenarios, ii) enabling synchronous and directional channel access in wireless LANs to handle congestion and interference, and iii) maximizing indoor spatial reuse through adaptive antenna beamsteering. Software-defined radios (SDRs) have become instrumental for experimental prototyping and evaluations of modern wireless technologies. Nevertheless, with their physical layer programmability, SDRs incorporates long and non-deterministic processing latencies that severely limits its utility in cross-layer MAC/PHY designs. Existing approaches to overcome this drawback tend to focus on either optimizing PHY software implementations with highly specialized techniques, or moving the entire PHY onto reconfigurable hardware (FPGAs and DSPs). These solutions inadvertently make reprogramming the radio layer becomes a complex and time-consuming task. We propose an extension of the recently emerged split-functionality architecture to enhance the timing and frequency agility of SDRs specifically for spectrum sharing applications. The key idea is to split MAC and PHY functionalities for hardware/software co-processing, with all time-critical functionalities implemented in hardware for low latency and deterministic timing, and all non-time-critical parts remaining in software for flexibility. We demonstrate the effectiveness of this design pattern through two specific spectrum-agile operations: a protocol-aware reactive jamming radio, and a bandwidth-adaptive Orthogonal Frequency-Division Multiplexing (OFDM) communications system. Both experimental systems are verified to work consistently with prior studies and can serve as valuable instruments to investigate and develop agile spectrum-sharing techniques to enhance the spectral efficiency and security of cognitive radio networks. An undesired aspect of widespread wireless adoption is the chaotic, unplanned, and unmanaged deployments of wireless devices, especially in indoor environments. Existing wireless LAN protocols such as IEEE 802.11, however, scales poorly to handle the interference problem in dense networks. To address this problem, we design a novel synchronous directional wireless architecture that dramatically increase indoor wireless capacity using reconfigurable antennas. The core of this solution is an antenna orientation algorithm that applies reinforcement learning to consistently identify and track close-to-optimal antenna directions. In addition, this wireless design also incorporates a new Time Division Multiple Access (TDMA)-based channel access protocol that can ensure sub-microsecond synchronization among all wireless nodes, making it much more efficient than existing asynchronous 802.11 MACs in maintaining quality of service across the network at scale.Ph.D., Electrical Engineering -- Drexel University, 201