Research On Interference-driven Precoding Techniques For Downlink In Cognitive Radio MIMO Systems

With the rapid development of global broadband wireless mobile communicationtechnology contradiction between supply and demand of spectrum becomes abottleneck constraint that hampers the sustainable development of wirelesscommunication. Cognitive Radio Multiple Input Multiple Output (CR MIMO)technology which can solve the contradiction between supply and demand ofspectrum receive widespread attention. CR MIMO takes advantage of the capabilityof dynamic spectrum allocation in cognitive radio and the unique advantages ofMultiple Input Multiple Output (MIMO) technology on spatial processing. Thereforeit maximally improves spectrum utilization efficiency of communication system andachieves higher data throughput and quality of service.Interference management, as one of the core problems need to be resolvedurgently in the cognitive radio network, is investigated in this paper. Conventionally,linear precoding methods in CR MIMO network mainly aim to limit or completelycancel the interference between Secondary Users (SUs) subject to the interferencepower constraint for the Primary Users (PUs) while achieving superior performanceof the CR system. When the user information and Channel State Information (CSI) atthe transmitter are available, the interference received by the target signal can bepredicted. Classifying or transforming the predicted interference, we can extract theconstructive part from the interference. By doing so, the Signal to Interference andNoise Ratio (SINR) is improved without extra transmitted power consumption. In thispaper, the precoding design for CR MIMO downlink from the perspective of usinginterference is studied. The main research contents and innovation points are asfollows:1) Cognitive Radio Partial Linear Precoding (CR-PLP) based on interferencedriven is proposed in the downlink of CR MIMO systems. First, according to thespace orthogonal projection theory, the secondary users’ signals are projected onto thenull space of interference channel matrix of primary user, in order to avoid the interference to primary users. Second, considering the property of multiuserinterference among secondary users, the interference is classified as eitherconstructive part or destructive part. The constructive interference with the samephase of the target signal is reserved and the destructive interference with theopposite phase of the target signal is zeroed by taking advantage of interferencejudgment criteria. CR-PLP not only avoids the interference to the primary but alsotakes advantage of constructive interference to enhance the energy of the target signal,which reduces the average Symbol Error Rate (SER) and enhances informationtransmission rate.2) Cognitive radio partial linear precoding based on interference classificationonly retains the constructive interference, abandons the use of destructive interference.So this article presents a novel precoding method, namely Cognitive Radio PhaseAlignment Linear Precoding (CR-PALP), which enhances the utilization rate of theinterference while converts destructive interference into constructive part and furtherimproves the performance of the system. CR-PALP adjusts the phases of interferencesymbols and makes interference transform to constructive part which facilitatescommunication by taking advantage of the phase information of the constellation andthe channel state information. Then, the effective signal energy is enhanced withoutthe need to increase additional total transmitted power at the Cognitive Base Station(CBS). The complexity and simulation results show that CR-PALP offers less symbolerror rate and enhances information transmission rate with lower computationalcomplexity.3) This article proposes CR-MMSE-PLP and CR-MMSE-PALP which are bothbased on the Minimum Mean Square Error (MMSE) for the reason that CR-PLP andCR-PALP don’t take the noise in the system into consideration. The optimal criterionsof these two precoding schemes are that the mean square error of the received signaland the transmitted signal with the effect of interference classification andtransformation is minimum. CR-MMSE-PLP and CR-MMSE-PALP better balancethe noise and interference in the system and then have better performance, especiallyin the condition of low Signal to Noise Ratio (SNR).