Reseach On Robust Optimization Algorithm Of Cooperative Relay Network

As the rapid development of mobile Internet, a mobile communication system in which users can high-speed access the network in real-time becomes a focus of research in recent years. Though cooperating with each other, cooperative relay network form a virtual multiple input multiple output(MIMO) systems, both can get the advantages of MIMO technology and avoid the problem that deploying multiple antennas on minidevice, thus is becoming one of the options for the future of wireless communications.When talk about cooperative relay network, this paper focuses on the design of weights matrix at relay nodes and equalization matrix at receiving nodes in an amplifyforward(AF) relay network. Currently most of the optimization algorithm only designed in consideration of Gaussian noise, but the non-Gaussian noise is widely present in fact, especially impulsive noise. Furthermore channel coefficients are estimated by the channel acquisition, which will exist channel estimation error and time delay caused by feedback. Above the existing problems can seriously decreased the performance of traditional optimization algorithms, so robust cooperative relay network algorithms become a research topic of this article. We focus on the robust design of the optimization algorithm under the environment of impulsive noise and robust designed algorithm in consideration of channel estimation error and time delay. Though reasonable setting, Gaussian mixture model can represent almost any distribution, so we can model the impulsive noise as a Gaussian mixture distribution. We find that the accumulation of multiple Gaussian mixture distribution is also fit for Gaussian mixture distribution, and obtain the probability density function. The performance of traditional beamforming algorithm based on second-order statistics dropped significantly under the impact of a bad environment noise, we use the maximum likelihood algorithm, which uses the probability density distribution of the noise to design the optimal value of relaying and equalization coefficients. Compared to the traditional Gaussian noise algorithms, our algorithm is more robust in the impulsive noise environment by simulation. Besides conventional relay optimization algorithms rarely consider the channel error or consider only one error, this article is considering channel estimation error and delay error at the same time. We build relationship channel between the estimated value and the true value by the appropriate mathematical models. On this basis, we propose the corresponding optimal algorithm to design the forward matrix and equalization matrix by minimizing conditional expectation of MSE. Comparisons with existing methods show that the proposed method provides better quality of communication when CSI estimation error and time delay are present.