Abstract

Multiple-input multiple-output (MIMO) networks are known to be able to achieve throughput performance superior to that available from single-input single-output (SISO) systems. However, when applying MIMO in multi-user networks, achieving this throughput advantage requires efficient precoding and optimal network scheduling. Furthermore, MIMO radios can help ensure security in a multi-user network. Previous work has proposed various precoding techniques for the MIMO broadcast channel, based either on channel state information (CSI) or channel distribution information (CDI), which achieve optimal or near- optimal MIMO channel throughput. The performance of these techniques largely depends on the availability of the channel information at the transmitter that must be fed back from the receiver. However, the past work has not examined the impact of latency caused by feedback of channel information and computation. This research proposes a performance metric to measure the throughput degradation caused by compression and feedback of channel information. We further propose an effective data compression technique based on the Karhunen-Lo`eve (KL) Transform and show that linear precoding (beamforming) based on CDI can achieve superior performance by providing stable channel throughput in both time- varying and frequency-selective channels. Very little prior work exists on optimal scheduling for multi-user MIMO networks, particularly in time-varying channels. One reason for this is that hybrid MIMO channels permit much more complex channel structures, such as broadcast channel (BC) and multiple access channel (MAC), whose capacity is limited not only by random channel noise but also by the multi-user interference. Furthermore, the achieved MIMO channel throughput depends on the spatial characteristics of the multi-user channels, a feature not captured by traditional network models based on signal-to-noise ratio and Doppler. Therefore, achieving near optimal performance requires development of scheduling techniques that depend on detailed channel characteristics. This dissertation proposes a novel parametric representation of the channel that simply describes the complex multi-user MIMO channels and allows for efficient scheduling. Because of the computational and feedback efficiency enabled by this parametric approach, it achieves low latency and therefore excellent performance.Finally, in any network setting, security is an important consideration. Specifically authentication ensures that unauthorized users do not gain network access. Unfortunately, user identity can be relatively easy to forge. This work therefore explores the user of radio- metric fingerprinting that uniquely identifies a device by unique imperfections in its transmitted waveform. This work shows that by applying this fingerprinting technique to MIMO devices, authentication reliability can be dramatically improved. The work also develops an information-theoretic approach to identify the optimal set of radiometric features to use for authentication and further considers the impact of drift in radiometric features on authentication performance.

Degree

PhD

College and Department

Ira A. Fulton College of Engineering and Technology; Electrical and Computer Engineering

Rights

http://lib.byu.edu/about/copyright/

Date Submitted

2011-07-11

Document Type

Dissertation

Handle

http://hdl.lib.byu.edu/1877/etd4638

Keywords

MIMO, CDI, feedback reduction, link scheduling, device identification

Share

COinS