Research On Peak-to-Average Power Ratio Reducation In Multicarrier Communication Systems

Multicarrier communication technology is adaptive to high data rate transmission in the wireless environment thanks to its intrinsic robustness to multipath fading and high frequency efficiency. Recently, multicarrier technology, especially orthogonal frequency division multiplexing (OFDM), is being successfully applied in intelligent navigation, satellite communication, digital video broadcasting systems, wireless access networks and power line communication, thus it is considered as one of the key technologies in the next generation wireless communication systems; multiple input multiple output (MIMO) system can increase system capacity and improve spectrum efficiency effectively without increasing the bandwidth and transmit power. Consequently, the combined MIMO-OFDM system provides better performance. However, high peak-to-average power ratio (PAPR), which may cause signal distortion, spectrum spread and system performance degradation, is the main obstacle to the application of the multicarrier communication system. Therefore, PAPR reduction has been a popular topic in multicarrier communication system.In this thesis, we research on the property of multicarrier communication system and the cause of high PAPR, then propose applicable PAPR reduction algorithms for various multicarrier communication systems according to their own properties. Simulation results demonstrate that the proposed algorithms can improve the PAPR performance effectively.We analyze the statistical property of PAPR in multicarrier communication, and research on the PAPR performance of OFDM signals in discrete time domain and continuous time domain respectively, thus provide the selection criterion of over-sampling factor. Meanwhile, we analyze the corresponding relation of OFDM signals between the autocorrelation property and PAPR performance, which provide reliable theory basis for design of effective PAPR reduction algorithms.Aiming at the high computational complexity caused by full enumeration of all possible phase factors in partial transmit sequences (PTS) algorithm, we propose PTS algorithms based on nonlinear optimization. Since the search of optimum phase factors is formulated as a particular combinatorial optimization problem, efficient search known from the combinatorial optimization algorithms can be applied to PTS. Specifically, we apply the simulated annealing (SA) to avoid full enumeration of phase factors, thus yield good PAPR performance with largely reduced computational complexity. Moreover, we apply a modified discrete particle swarm optimization (DPSO) to search the optimum combination of phase factors. The hamming distance is adopted to adjust the velocity update formula of particles, which makes the phase factors always approach to the optimum ones quickly, thus achieve the signals with low PAPR. These algorithms are easy to be realized, and can obtain significant PAPR reduction performance with low computational complexity, hence have good serviceability.In fact, the iterative PTS (IPTS) algorithm can be regarded as a problem of traversing binary tree. From this view, we find that IPTS algorithm with linear approach is easy to be trapped in a local optimum solution, thereby, a nonlinear IPTS (N-IPTS) algorithm based on the thought of "probability acception" from Metropolis criterion is proposed. Since the N-IPTS algorithm breaks out of the linear search, it provides higher freedom for the search of optimum phase factor, and improves the PAPR performance significantly. Furthermore, because the IPTS algorithm is sensitive to initial phase factors, we repeat the iterative search process to eliminate the effect of initial phase factors. Simulation results show that N-IPTS algorithm can achieve good PAPR performance with low computational complexity, thus has high feasibility.Selected mapping (SLM) is an effective PAPR reduction algorithm, whose performance is related to the quantity of alternative signals, the more alternative signals exist, the better PAPR performance can be achieved, and the computational complexity increases accordingly, especially in MIMO-OFDM system. Hence, a SLM algorithm based on signal decomposition (D-SLM) is proposed to solve this problem. It is able to provide sufficiently larger number of alternative transmit signals, thus achieve better PAPR performance at the cost of slightly number of redundant bits (side information). Moreover, using the conjugate symmetry property of real sequence, the computational complexity of proposed algorithm remains unchanged. Furthermore, we propose a concurrent SLM based on signal decomposition (D-CSLM), which is able to utilize the additional degree of freedom offered by multiple transmit antennas, to yield excellent tradeoff between side information and PAPR performance.We also discuss the MIMO-OFDM system based on space time block code (STBC), and demonstrate that the conjugate symbols of two antennas have same PAPR property. The computational complexity cost of PAPR reduction algorithms with this property are reduced significantly, and the transmitted side information is also decreased, thus spectrum efficiency is improved. In addition, we propose a new signal selection criterion, which can reduce the possibility of encountering poor signals with high PAPR, thus achieve good PAPR performance.