Research On Resource Management And Outage Probability For Cooperative Cognitive Radio Networks

Cognitive radio is envisioned as a promising technique to greatly improve the spectrum efficiency by adaptively detecting and exploiting under-utilized wireless resources. Meanwhile, cooperative communication technique can induce space diversity for transmitters by asking some potential relays for cooperation. Not only spectrum efficiency is enhanced, but space diversity order and system throughput gains are also obtained by combining cognitive radio with cooperative communication techniques. Cooperative cognitive relay networks(CCRN), which possesses advantages of both cognitive radio and cooperative communication, has been become one of the most popular research topics in wireless communication field. Especially, the network performance and resource management(RM) of CCRN interest many researchers and institutes around the world. In this dissertation, outage performance and RM that includes channel and power allocation with relay selection for CCRN are investigated, respectively.Cognitive relay channel and CCRN transmission models are firstly presented based on Gauss interference channel model and cooperative communication transmission model. The cognitive relay channels can be classed as three types of sub-channels, which termed as direct transmission channel, relay channel and heterogonous dual-hop channel. Throughput of the CCRN with decode-and-forward(DF) and amplified-and-forward(AF) are also presented. A general RM model including relay selection, joint channel and power allocation/control for CCRN with multiple orthogonal channels are deduced. Some mathematic principles and theorems for RM, such as convex optimization, geometric programming, game and inequality variation theories are introducedBased on the general RM model, joint channel and power allocation for a CCRN that consists of a cognitive source, relay and destination are investigated. In the CCRN, each cognitive user is with multiple cognitive relay channels. Since the RM only concerns one relay system, it is termed as intra-CCRN RM in this dissertation. The optimal channel allocation algorithm for intra-CCRN using DF transmission protocol is firstly presented and analyzed. This optimal algorithm is performed by choosing the optimal channel combination of cognitive relay channels that result in the maximum throughput. It shows that the computational complexity of the optimal algorithm increases exponentially with N usual relay channels, i.e. O(4N). For decreasing the complexity, a throughput increment based sub-optimal algorithm with linearly complexity to N usual relay channels is proposed and analyzed. By implementing channel allocation as establishing an assignment matrix, a similary sub-optimal channel allocation algorithm based on throughput increment is presented for CCRN with AF. Numerical results show that sub-optimal algorithms are greatly decreased the computational complexity with minus performance loss.Based on above channel allocation algorithms, joint channel and power allocation for intra-CCRN with DF and AF protocols are considered in sequence. For intra-CCRN RM with DF protocol, a logarithm approximation based series geometric programming algorithm(LASGPA) is proposed to turn the non-convex power allocation problem into a series of convex problems for easily obtaining the solution. It is provable that the solution of LASGPA is also the global optimal solution of the original RM problem. Furthermore, timeslot allocation is also considered, and joint channel, power and timeslot allocation algorithm is proposed. For intra-CCRN RM with AF protocol, an expectation maximization theory based power allocation algorithm is proposed, which also transforms such non-convex optimization problem into convex optimization problems for using existing convex optimized tools to achieve the optimal solution. And a joint channel and power allocation algorithm was also proposed. Following conclusions are obtained by numerical simulations; 1) channel allocation is much more efficient than power allocation for increasing the throughput; 2) joint channel and power allocation can further enhance the throughput performance. Especially in CCRN with DF protocol, it is concluded that 3) joint channel, power and timeslot allocation can achieve the best performance in terms of throughput under medial and low SINR; 4) joint channel and power allocation can acquire larger throughput increment than joint timeslot and power allocation.In cognitive radio networks(CRN), multiple cognitive users compete in spectrum resources for transmission. Thus, RM for not only one cognitive relay system but among them should be taken into account. The resource management among cognitive radio networks is termed as inter-CRN RM in this work. However, each user’s transmission power and relay selection factors are respectively coupled with other users in the inter-CRN RM. Thus, the relay selection, joint spectrum allocation and power control/allocation are formulated as a GNEP, and distributed algorithms are proposed. By exploiting variational inequality theory, such GNEP is decomposed into Nash equilibrium problems for easily acquiring the generalized Nash equilibrium. The existence and uniqueness of the generalized Nash equilibrium are discussed and analyzed; a rigorously sufficient condition to ensure the uniqueness is also derived.Outage probability is one of the key performance merits of the CCRN. The opportunistic spectrum access(OSA) model and the merit of spectrum access capability are proposed by joint considering primary users’ activities and cognitive users’ detection probability. A hybrid spectrum access(HSA) manner composed of OSA and underlay spectrum access is proposed. Outage performance for three-nodes CCRN under OSA, underlay spectrum access and HSA etiquettes are investigated. Closed outage probability expressions and diversity orders for CCRN experienced flat Rayleigh fading channels under such spectrum access etiquettes are presented. Moreover, time division repetition(TDR) transmission and relay selection transmission are also considered for formulating outage probability, respectively. All outage probability is simulated, and following conclusions are obtained. 1) With OSA etiquette, the outage probability does not decrease as the number of potential relays increasing in the low and medial SNR region under TDR transmission manner. However, under relay selection transmission manner, the outage probability decreased as the number of potential relays increasing. 2) Under hybrid spectrum access etiquette, the outage probability with proactive relay selection(PRS) criterion outperforms that with reactive relay selection(RRS) criterion. 3) The outage performance under HSA always outperforms OSA and underlay spectrum access.