Research On Channel Estimation And Signal Detection Technologies Based On Multi-Way Arrays

Unlike low-rank two-way arrays decomposition, which has rotating ambiguity, low-rank multi-way arrays decomposition is unique when loading matrices satisfy some conditions. In wireless communication systems, dimensions of multi-way arrays model can be associated with several signal dimensions such as space, time and frequency. By fitting the multi-way arrays model, signal detection and channel estimation based on uniqueness property of decomposition can be derived without or with only a little specific information such as the channel state information (CSI) and coding matrices. Due to its ability to improve spectral efficiency and reliability of communication systems, the signal processing technique based on multi-way arrays has gained widely attention and studies.This thesis aims to study channel estimation and signal detection technologies based on multi-way arrays. By making a deep study on multi-way arrays models and fitting algorithms, and combining the low rank decomposition of multi-way arrays with the technologies such as multiple-input and multiple-output (MIMO) relay and space-time coding, some novel channel estimation and signal detection methods are proposed. The major work and contributions of this thesis are summarized as follows.(1) To improve the fitting speed of the parallel factor (PARAFAC) model, this thesis presents a low complexity fitting algorithm. In each iteration, the proposed algorithm sets up their own relaxation factors for two loading matrices which are required to be estimated, and gets the optimal couple of two relaxation factors by the joint optimization. Theoretical analysis and simulation results show that the proposed algorithm improves the fitting speed of the PARAFAC model without performance degradation compared with the existing bilinear alternating least-squares (BALS) algorithm.(2) This thesis presents a PARAFAC-based low complexity channel estimation method for a two-hop amplify-and-forward (AF) MIMO relay system. Based on the proposed estimation method, by using the PARAFAC model constructed, the proposed low complexity fitting algorithm can estimate full knowledge of all channel matrices involved in the communication efficiently. The linear minimum mean square error (LMMSE) approach is used to further improve the initial estimation of channel matrices. Theoretical analysis and simulation results show that the proposed PARAFAC-based channel estimation method yields smaller channel estimation error with lower complexity compared with existing channel estimation methods.(3) Compared with the channel estimation problem of two-hop MIMO relay systems, the channel estimation problem in multi-hop MIMO relay systems becomes more complicated. To the best of our knowledge, only a few related works have recently been done in this field. In this thesis, we propose a PARATUCK2-based channel estimation method for three-hop MIMO relay systems. The signal received by the destination can be constructed to a third-way arrays model, which satisfies a PARATUCK2 model. The identifiability and uniqueness issues are analyzed. By carrying out the PARATUCK2 model fitting, the proposed channel estimation method provides the destination with knowledge of channel matrix of each hop. Theoretical analysis shows that the proposed channel estimation method is also applicable for multi-hop MIMO relay systems with any number of hops. Compared with existing methods, the proposed method only requires the source to transmit the channel training signals. Moreover, the proposed method does not need relays to perform any task of channel estimation, and yields smaller estimation error. Numerical examples are shown to demonstrate the effectiveness of the PARATUCK2-based channel estimation method.(4) Conventional signal detection methods for MIMO relay systems need to know CSI, which is usually obtained by using channel training signals. However, the frequent use of the channel training signals will decrease the spectrum efficiency. In order to improve the spectrum efficiency of MIMO relay systems, this thesis presents a method for joint channel estimation and signal detection with the multiple Khatri-Rao space-time coding. Under the assumption that CSI is not available, the proposed method formulate the received signals as two PARAFAC models, and then a semi-blind receiver is proposed for joint channel and symbol estimation. Theory analysis and simulation results show that the proposed method can efficiently estimate all CSI in this communication system, and the presented semi-blind receiver operates close to the pilot-assisted zero forcing receiver.(5) Receivers based on the signal processing technique of multi-way arrays can be applied for channel and signal estimation with a blind or a semi-blind approach in point-to-point MIMO systems. Since existing receiver designs do not consider the cooperative link, they can not be applied to cooperative MIMO relay systems directly. In this thesis, we develop a receiver design scheme based on four-way arrays mode for two-hop cooperative MIMO relay systems. At the source node, the proposed transmit signal design combines the linear constellation precoding with time-domain spreading. We show that the received signals at the destination node can be formulated as a four-way arrays mode, and then a receiver is proposed for joint channel and signal estimation by fitting the formulated model. Theory analysis and simulation results show that the proposed receiver not only outperforms some existing tensor-based receivers, but also can operate under blind and semi-blind configurations.