Blind detection designs with unique identification in two-way relay channels
In amplify-and-forward (AF) two-way relay networks, information is exchanged be-tween two source nodes via a relay node by AF relaying. This dissertation develops the blind detection designs for various two-way relay systems such as one antenna AF two-way relay network with direct link, one antenna AF two-way relay network without direct link, and multiple antenna AF two-way relay network without direct link. In the ?rst system, an e?ective signalling and transmitting scheme using four M-ary phase-shift keying (M-PSK) constellation sets is proposed to achieve a unique identi?cation of the transmitted symbols and the channel coe?cients in a noise-free transmission. Blind receivers with full diversity are derived for Rayleigh fading channels with Gaussian noise by the generalized likelihood ratio test (GLRT) and least square error (LSE) criterions. Constellation selections are investigated for the source nodes to transmit at a uniform bit rate in each time slot. Power allocation schemes are discussed to further improve the average error probability of the source nodes in the blind detections. In the second system, frequency selective fading channels are assumed between each node. The knowledge of channel state information (CSI) is assumed not available for all nodes. An e?cient transmission and signaling scheme is proposed. To deal with the high complexity in implementing the LSE receiver, a parallel iterative sphere decoding scheme is investigated. The low complexity of this scheme allows that the joint estimation and detection of the channel coe?cients and transmitted symbols can be e?ciently implemented. In the third system, multiple antennas are equipped on both source nodes. This part of work proposes an e?cient transmission and signaling scheme with time-reversal space-time-block-coded (TR-STBC) blind detection design for frequency-selective two-way relay channels. The similar sphere decoder is derived and simulated for this system to reduce the computational complexity.
Thesis (Ph.D.)-- Wichita State University, College of Engineering, Dept. of Electrical Engineering and Computer Science