Traffic Pattern Prediction Based Spectrum Sharing for Cognitive Radios

In this chapter, we introduce different traffic prediction techniques and discuss the process of evaluating channel availability through predicting traffic pattern of primary users for cognitive radios. When cognitive and non-cognitive users share the licensed spectrum, compared with secondary users (cognitive users), primary users have higher priority in using licensed channels. Therefore, whenever a primary user is detected, secondary users must vacate the relevant channels or decrease their transmitted power to reduce the interference on primary users. However, in some situations, due to the activities of primary users, secondary users may need to switch to other available channels, terminate communication, or reduce the transmitted power frequently. This leads to temporal connection loss of secondary users. In addition, if secondary users can not vacate a channel in a timely manner, it would interfere with primary users. To reduce the temporal connection loss and interference on primary users, secondary users should avoid using the channels which would be claimed by primary users with a high probability within a given time period. A solution to this problem is to enable secondary users to evaluate the channel availability through predicting the traffic pattern of primary users. This chapter discusses the methods of evaluating the probability of channel availability for secondary users, and examines the probability that there is a successful secondary user’s transmission at one time instance. In addition, the cooperative prediction is briefly introduced. Finally, the applications of traffic pattern prediction technique to spectrum sharing are discussed.