Cognitive radio (CR) refers to radios that can learn the environment and adapt their transmissions to the environment. We focus on the investigation of CR with two special capabilities: (i) Radios that can detect and access vacant wireless spectrum, and (ii) detect and transmit cooperatively with each other. More specifically, we will investigate the use of CR to support dynamic spectrum access (DSA) for spectrum efficiency, and the use of CR to form cooperative transmission arrays for enhancing transmission security.

First, we investigate the performance of CR when used for DSA. We begin by analyzing the capacity of a CR-based secondary user in a broadcasting system consisting of one primary transmitter and multiple primary receivers. The average transmission power and capacity of this secondary user are derived under two different models of primary users. The results are then extended to include multiple secondary users. Closed-form upper and lower bounds are also derived. The results show that CRs can achieve a significant level of extra transmission capacity when allowed to access dynamically the same spectrum with the primary users.

Then, we address one of the major challenges for DSA, i.e., uncoordinated secondary transceivers can synchronize to the same spectrum that varies unpredictably in both time and space. By using a special frequency hopping spread spectrum transmission scheme, each CR independently detects usable spectrums, and CR selects one of them to start transmission or receiving according to a pre-defined frequency hopping pattern. The scheme works without any a prior assumption on handshaking, and is robust to even large spectrum sensing errors.

Next, we study the application of CR to assist other transmitters by forming a virtual transmission antenna array. We propose the use of this idea to secure wireless transmissions at the physical layer. This approach addresses a unique weakness of wireless networks whereby network traffic traverses a public wireless medium making traditional boundary controls ineffective. A randomized array transmission scheme is developed to guarantee wireless transmissions with inherent low-probability-of-interception (LPI).

Finally, we conclude this dissertation by summarizing the work and contribution, and by suggesting some future research directions.