Channel Estimation And Signal Detection For Massive MIMO Systems

In order to meet the exponentially growing demand in wireless services,future wireless communication systems require breakthroughs in communication technology.Massive multiple-input multiple-output(MI-MO)technology,which employs a large number of antennas at the base station,can significantly improve the spectral efficiency and energy efficiency,and has become a core technology of the new generation wireless communication systems.Acquisition of channel state information(CSI)is the basis for massive MIMO wire-less transmission,Orthogonal pilot signals are employed in conventional channel estimation methods,and the corresponding pilot overhead scales linearly with the number of antennas,which limits the performance gain brought by massive MIMO.Motivated by these reasons,we investigate channel estimation scheme with low pilot overhead.Then according to the result of channel estimation,the signal detection method with imperfect CSI is studied.Firstly,we investigate the traditional downlink channel estimation methods in massive MIMO systems,including least square(LS),linear minimized mean square error(LMMSE)and relaxed minimum mean square error(RMMSE).The derivations of these channel estimation methods are described firstly.Then we analyze the optimal pilot structure for each method,theoretical results show the influence of the number of antennas at the base station on the error performance of these methods.On the other hand,based on the result of channel estimation,we study two signal detection methods,including robust MMSE and generalized approximate message passing(GAMP)which considers the channel estimation error.These two methods are derived from MMSE and GAMP respectively by considering channel estimation error as a part of noise.Simulation results show that these two signal detection methods provide performance gains over MMSE and GAMP respectively when the CSI is imperfect.Secondly,we propose a channel estimation scheme based on compressed sensing for downlink massive MIMO systems.By exploiting the local sparsity characteristics ofthe outdoor wireless channels,a new sparse representation of the channel vector has been proposed by using the multiband modulated discrete prolate spheroidal sequences(DPSS)dictionary ?.A deterministic pilot structure that satisfies the ? restricted isometry property(?-RIP)is then proposed to guarantee the accurate recovery of signal that are nearly sparse in dictionary ?.To further reduce the pilot overhead,the ?-block-RIP is developed as a relaxed property to replace ?-RIP,a modified Block-Based CoSaMP(BBCoSaMP)algorithm based on ?-block-RIP is proposed to recover the channel vector.Numerical results show that the developed recovery algorithm combined with the proposed multiband modulated DPSS basis reduces the pilot overhead and provides significant gains over existing methods in terms of channel estimation error.Finally,we propose a robust approximate message passing(RAMP)detection algorithm with imperfect CSI based on minimizing Bethe free energy.For given pilot structure and channel estimation methods,the results of channel estimation should be denoted by a probability density function of CSI rather than its esti-mation values.Based on such observations,the MIMO detection in the presence of channel estimation error is formulated as a Bethe free energy minimization subject to appropriately imposed constraints and the given statistical model of CSI.The Lagrange multiplier theory is employed to identify the stationary points of the constrained Bethe free energy,which give us back the fixed-point equations.This results in an iterative algo-rithm for detection in massive MIMO systems.Numerical results show that this algorithm not only has a low computation complexity,but also improves the detection performance when the CSI is imperfect.