This paper presents an investigation of spectrum co-existence between IEEE 802.11b and 802.16a networks in the same shared frequency band using cognitive radio techniques with different levels of complexity. Simple reactive interference avoidance algorithms as well as proactive spectrum coordination policies based on etiquette protocols are proposed and compared in terms of achievable spectrum efficiency in a shared Wi-Fi/Wi-Max scenario. In reactive interference avoidance methods, radio nodes coordinate spectrum usage without exchange of explicit control information-this is done by adaptively adjusting transmit PHY parameters such as frequency, power and time occupancy based on local observations of the radio band. Because local observations provide information only about transmitters, they may not be sufficient for resolving spectrum contention in scenarios with "hidden receivers". Proactive coordination techniques solve the hidden-receiver problem by utilizing a common spectrum coordination channel (CSCC) for exchange of transmitter and receiver parameters. Radio nodes can cooperatively select key PHY-layer variables such as frequency and power by broadcasting messages in the CSCC channel and then following specified spectrum etiquette policies. An ns2 simulation model is developed to evaluate both reactive and proactive etiquette policies in scenarios with co-existing IEEE 802.11b and 802.16a networks. The density of radio nodes in the coverage region, and their degree of spatial clustering are key parameters in the system evaluation. Detailed simulation studies were carried out for a variety of scenarios including both single and multiple 802.11b hotspots per 802.16a cell with and without spatial clustering. Our results show that simple reactive algorithms can improve system throughput when sufficient "free space" (in frequency, power or time) is available for PHY adaptation. In more congested scenarios with spatially clustered nodes and hidden receivers, the proposed CSCC etiquette can significantly improve overall system performance over reactive schemes.
