| The development of data services and the popularity of mobile devices present further demand on spectral efficiency. As a promising candidate, large scale multiple-input multiple-output (also called massive MIMO) technology is proposed to enhance system capacity. However, extra antennas will burden the system power consuming and the reuse of resource in massive MIMO system will bring complex interference. Besides, the pilot contamination will finally limits the performance of massive MIMO. In order to solve these problems, power control schemes in massive MIMO systems is developed and user performance of massive MIMO systems is analyzed.Firstly, user performance is analyzed in massive MIMO system. At first, considering perfect channel state information, statistical signal to interference and noise ratio expression is derived with both maximal ratio combination receiver and zero forcing receiver. Then, considering imperfect channel state information, the achievable signal to interference and noise ratio is deduced under minimum mean square error channel estimation and zero forcing receiver. Using the statistical properties of high dimensions channel vector, we make a quantitative analysis of the achievable signal to interference and noise ratio and reveal the liner re-lationship between the pilot power and the interference and the noise. Meanwhile, an asymptotic analysis about the user achievable signal to interference and noise ratio according to pilot power is conducted and the upper bound is derived. Lastly, we proposed a pilot power allocation scheme for reference under condition that the pilot contamination is serious and per user energy efficiency is defined. Simulation results verify the accuracy of the deduction.Secondly, power control scheme based on standard interference function is proposed. By utilizing the properties of high dimension channel vector, Jensen's inequality and fractional programming, a convex energy efficiency optimization problem is derived. However, the coupling of power variables makes getting an explicit solution a challenging work. Alternatively, we get an implicit solution. By resorting to standard interference function, the optimal solution can be derived in few iterations based on the implicit solution. The simulation results prove that the proposed converges to the optimal point rapidly.Finally, power control schemes based on searching method are studied. On one hand, to simplify the en-ergy efficiency problem, some approximate transformations are adopted and a convex optimization problem is got. Comparing with the derivation method before, scheme based on particle swarm optimization algorithm is proposed in the following. The scheme relaxes demand on the convex property of utilizing function. Once there exists a global optimal point, it is guaranteed that the point can be got via particle swarm optimiza-tion searching method. On the other hand, considering condition there is no global optimal point, schemes mentioned before seem unreasonable. Hence, power control scheme based on multi objective optimization is studied. A model optimizing both the system throughout and power consuming is built at first. Chebyshev scalaxization is adopted in the next step to simplify the original problem. Then, the Pareto curve describing the tradeoff between system throughout and power consuming is derived via multi objective evolutionary algorithm. And optimal transmit power is finalized according to the certain system constrains. Simulation results confirm validity the of the two schemes... |
PDF Full Text Request