Multiple-symbol Detection In UWB Systems

Ultra-wideband (UWB) is a communication technology without carrier,where the information is transmitted with very short duration pulses and is mainly used in short-range wireless communications. UWB communication technology has many advantages, including low-power consumption,low-complexity of the system implementation, as well as precise positioning.However, it also faces many challenges. UWB channel is a high-frequency multipath channel, where the number of multipath is up to a few hundreds, and the maximum delay is up to hundreds of nanoseconds. In order to avoid esti-mating the channel state information, we focus on the noncoherent UWB sys-tems in this paper. As a typical technology of noncoherent systems, currently,noncoherent detection plays an important role in the field of wireless commu-nications, and it has been widely concerned in the international academia and industry, where the most attractive scene is the UWB communication.In this thesis, multiple-symbol differential detection (MSDD) technology is investigated for the noncoherent systems. In order to avoid the inter-symbol interference, the adjacent pulse intervals are generally set to be large. Cor-respondingly, this arrangement limits the actual communication speed. MIMO(Multiple-input multiple-output, MIMO) technology increases transmission rates by multiplexing, and improves signal-noise-ratio (SNR) for detection through diversity. Thus, MIMO-UWB is likely to become one of the most attractive physical layer technology and is widely used in the future wireless communi-cation networks. As a typical noncoherent MIMO technology, MSDD is inves-tigated in the differential space-time block code (DSTBC) aided MIMO-UWB systems. Furthermore, due to the bandwidth of the UWB, the power of UWB is limited to be low to avoid mutual interference with the other systems. This correspondingly limits the distance of UWB communication. Therefore, we further investigate noncoherent detection in a two-way relay network.Compared with the ideal coherent reception, there always exists some per-formance loss for noncoherent reception both in point-to-point communication and a two-way relay network. Firstly, based on the theory of belief propagation(BP), a BP based noncoherent detection is proposed in this paper. Then, from the perspective of reducing the complexity, an efficient Viterbi algorithm (VA)based noncoherent detection algorithm is proposed. In addition, both semi-definite relaxation (SDR) aided algorithms are proposed for MSDD and multi-user detection. Furthermore, we have extended the noncoherent transmission scheme in the network coding aided relay system for single antenna and DSTBC aided MIMO-UWB systems, respectively. The proposed schemes with low-complexity relatively, therefore they can be implemented in noncoherent UWB systems. The major contributions and innovations of this thesis are summarized as follows.1) A BP based multi-symbol differential detection (BP-MSDD) is pro-posed for the DSTBC aided noncoherent MIMO-UWB systems. The proposed BP-MSDD outperforms the tradition GLRT-MSDD due to its forward and back-ward message passing mechanism. Furthermore, the joint iterative BP-MSDD and decoding algorithm takes advantage of the soft information between the detector and channel decoder and achieves performance gain.2) In order to reduce the complexity of GLRT-MSDD of DSTBC aided noncoherent MIMO-UWB system, VA-aided MSDD with hard decision (VA-HMSDD) algorithm is proposed by reformulating the metric function of GLRT-MSDD and searching the code with the minimum distance between the re-ceived sequence and the code sequence in the trellis. Furthermore, VA-aided MSDD with soft decision (VA-SMSDD) is proposed to reduce the complexity of the BP-MSDD. These two VA-MSDD algorithms have performance-and-complexity trade off. Thus, they are suitable for DSTBC aided MIMO-UWB system with low-complexity implementation.3) Semi-definite relaxation based multi-symbol differential detection (SDR-MSDD) and semi-definite relaxation based multi-user detection (SDR-MUD)are proposed, which reduce the complexity while maintaining detection per-formance. By slacking the conditions, non-convex optimization problems are transformed into convex optimization problems, i.e., the original nondeterministic-polynomial time hard (NP-hard) problems become the polynomial-time prob-lems. Therefore, the proposed SDR-MSDD and SDR-MUD algorithms have much lower complexity than the exhaustive search algorithms.4) The application of the proposed noncoherent detection to some more complex scenes is explored, especially in the UWB relay network. From the perspective of the existing detection theory, the receivers are designed suit-able for network-coding aided noncoherent UWB systems. In particular, in the time-division multiple access phase with three time-slots transmission, GLRT-MSDD and SDR-MSDD are proposed for the relay communication systems equipped with single antenna; and GLRT-MSDD and decision-feedback based MSDD are proposed for the DSTBC aided MIMO-UWB relay systems. In the broadcasting phase, a factor graph-based MSDD is proposed to improve the de-tection performance effectively. The proposed schemes avoid the estimation of the channel and realize the low-complexity transmission.