Research On Spectrum Sensing Technology With Dynamical Uncertain Environment

Cognitive radio (CR) is an effective technology to improve the efficiency of spectrum utilization significantly, in a way that does not affect the quality of service (QoS) of primary users (PUs). Therefore, it has drawn general atten-tion and become an important part of the 5th generation (5G) communication.Supported by the Natural Science Foundation of China as well as the Ph.D. Pro-grams Foundation of the Ministry of Education, this thesis is mainly devoted to promoting CR techniques in the future wireless communications, which may have the great theoretic significance and the wide application prospect.This thesis mainly investigates the CR techniques in complicated wireless environment, with the emphases especially put on spectrum sensing and sens-ing allocation under time-varying fading channel and noise uncertainty. Ac-cordingly, the main contributions can be summarized as follows.1. We formulate a new dynamic state-space model (DSM) to character-ize the spectrum sensing process with time-varying fading channel. i) In this unified stochastic model, the PU's states and the time-varying channel are con-sidered as two hidden states to be estimated; ii) A two-state Markov chain is used to model the evolution of PU's states (active / inactive); iii) For the sin-gle path channel, the evolution of channel states is abstracted as a finite states Markov chain (FSMC), while the tapped delay line is utilized to model the mul-tipath channel; iv) The ED-based or MFD-based receiver can be chosen based on the physical receiving device and the demands for the detection precision.2. A new spectrum sensing frame for single-node is proposed to cope with time-varying fading channel.i) A joint estimation algorithm is designed based on the Bayesian sequential theory, which can jointly detect/ estimate the two hidden states in real time; ii) Given that the likelihood information of fading channel will completely disappear under inactive state of PU and the channel states are assumed to be slow-varying, coarse detection and accumulative mod-ification schemes are integrated; iii) An effective iterative estimation mecha-nism is proposed for a more complex environment with time-varying multipath channel, and hence significantly enhance the blind detection performance.3. A joint spectrum sensing scheme with noise uncertainty is investigated.i) A new method is presented with unknown but static noise parameters. Based on the conjugate prior theory and marginalized particle filtering, this method can estimate the noise parameter and the PU states jointly; ii) A joint spec-trum scheme with dynamical noise variance is proposed, this Bayesian solution can recover the dynamical noise variance by an adaptive mechanism of forget ting factor which can take advantage of the slow varying characteristic of noise variance and enhance the estimation accuracy; iii) The joint detection scheme is extended into a more complex environment with unknown and dynamical inference and noise.4. An adaptive sensing allocation schedule with time-varying fading chan-nel is designed to maximize the throughput of secondary user (SU) under the constraints on the detection probability. i) A DSM which considers sensing-throughput tradeoff problem with time-varying fading channel is formulated;ii) We prove that there indeed exists the optimum sensing time for different channel gain and propose the relation function between them; iii) The adaptive mechanism for SU to reconfigure the sensing duration is established based on the Markov property of channel.Finally, the investigations of the whole thesis are summarized. At the same time, several valuable research interests on CR techniques in the future com-munications are discussed.