Abstract
Cognitive sensor network (CSN) is a promising paradigm to address the spectrum scarcity problem in traditional wireless sensor networks. Reliable spectrum sensing is essential to enable the normal operation of a CSN. Existing researches showed that by exploiting spatial diversity, cooperative sensing can greatly improve the detection performance over noncooperative sensing in opportunity-homogeneous environment. At a given time, cognitive sensors at different locations, however, may experience heterogeneous spectrum opportunities making the cooperation among cognitive sensors intractable. In this paper, we show the limitations and drawbacks of merely using temporal-domain detection performance metrics and introduce novel spatio-temporal detection performance metrics to guide the design of joint spatio-temporal spectrum sensing. An efficient one-bit hard decision based three-phase (i.e., a global cooperation phase, a local cooperation phase, and a joint decision phase) spatio-temporal sensing algorithm is proposed and numerical results demonstrates the effectiveness of the proposed algorithm.