A main trend in current research on wireless communication technology is the realization of dynamic access to radio spectrum. This trend is not only a challenge to technology but also increases the pressure on regulation authorities to rethink their licensing and assignment of spectrum. Intelligent radio systems, so-called cognitive radios, are a promising way out of the regulatory dilemma of satisfying the consumer demand for freely available spectrum on the one hand and the underutilization of many licensed frequency bands on the other hand. Cognitive radios autonomously coordinate spectrum sharing in learning from their environment. They essentially increase the efficiency of spectrum utilization through interference avoidance and local re-use of frequency bands originally designated for exclusive operation. This thesis provides an overview about the options for radio spectrum regulation and dynamic spectrum access with a special focus on spectrum sharing. The rising demand in wireless communication for free available spectrum goes along with increasing restrictions to spectrum utilization, i.e., Quality-of-Service (QoS) requirements, as for instance in consumer electronics or other multimedia applications. Therefore, the widely spread standard for wireless local area networks IEEE 802.11 is extended through 802.11e with enhancements to support QoS. These enhancements are evaluated in this thesis in especially concentrating on distributed environments. The realization of spectrum sharing between different wireless networks is discussed in introducing a centralized approach for the coexistence and interworking of 802.11(e) and 802.16.
Spectrum sharing, either of unlicensed frequency bands or opportunistically used spectrum, requires a coordinated medium access of the competing cognitive radios to support nevertheless QoS. The distributed QoS support in modified 802.11e cognitive radio networks is evaluated in this thesis. Spectrum sharing among cognitive radios is discussed in applying and comparing solution concepts derived from information theory on the one hand and from game theory on the other hand. These approaches allow cognitive radios to support and guarantee QoS when sharing spectrum without requiring direct information exchange in observing past spectrum utilization.