Research On Channel Estimation Technique For3D Mimo Systems

With higher and higer demand for surfing, multimedia application and file transmission, the mobile communication system is promoted to apply more efficient techniques constantly. Through exploiting vertical direction spatially,3D MIMO (three-dimensional Multi-input and Mul-ti-output) technique can decrease inter-cell interference, improve throughput and spectrum efficiency further. Therefore,3D MIMO tech-nique is an effective way to satisfy the great demand for data rate. As an important part of3D MIMO systems, channel estimation provides chan-nel state information (CSI) to perform diversity combination, coherent detection and other transceiver processes. Channel estimation technique can be classified into two categories, namely blind channel estimation and non-blind channel estimation. Pilot aided channel estimation (PACE), belonging to the non-blind, is applied widely. This work focuses on the channel estimation technique for3D MIMO systems, and mainly involves following two parts:At the beginning, we study pilot pattern design for3D MIMO sys-tems. In3D MIMO systems, antennas’amount is large, so pilot overhead become high, resulting to the low data effectiveness. Therefore, aiming to reduce pilot overhead and meanwhile maintain estimation accuracy, we propose4D (four-dimensional) pilot pattern design. Through mathemati-cal processing, pilot spacing constraints in the transverse, vertical, time and frequency are provided. It is verified that this method can not only decrease pilot overhead substantially, but also improve channel estima-tion accuracy by exploiting spatial correlation of the channel. Futhermore, simulation results show that channel estimation in space domain cannot be decomposed out, and estimation in the transverse and vertical cannot be decomposed out of space domain. Otherwise, it will cause big perfor- mance degradation.Then, we study the channel estimation algorithm suitable for3D MIMO systems. By analyzing3D MIMO channel deeply, it can be found that this channel can be expressed by channel corresponding to antennas in one transverse and one vertical directions, which can be estimated by exsiting methods. According to this feature, we propose two channel es-timation algorithms with low pilot overhead. Numerical results show that these two alogithms have better MSE (Mean Square Error) performance than Winner filter when SNR (Signal-to-Noise Ratio) is large.The above are the mainly research and innovation points in this the-sis, which focuses on channel estimation technique for3D MIMO sys-tems. The4D pilot pattern design method proposed in this work can not only maintain channel estimation accuracy, but also need few pilots, re-sulting in the improvement of effective data transmission rate. On the other hand, channel estimation algorithms proposed in this work need lit-tle pilot and have a good performance by exploiting the feature of3D MIMO channel deeply. The decomposable characteristic of3D MIMO channel found in this work can be referred for the following study.3D MIMO technique can decrease inter-cell interference, improve throughput and spectrum efficiency further, and has been paid more and more atten-tion. The above achievements can provide a significant theoretical refer-ence for the research and application of3D MIMO technique.