On Peak-to-Average-Power Ratio Reduction For OFDM Systems

Orthogonal frequency division multiplexing (OFDM) has been used widelydue to its inherent robustness against multipath fading and resistance to nar-rowband interference. However, one of the major drawbacks of OFDM signals isthe high peak to average power ratio (PAPR) of the transmitted signal. Manysolutions have been proposed to reduce PAPR in recent years, such as clipping,coding, selected mapping (SLM), partial transmit sequence (PTS), tone reserva-tion (TR) algorithm, tone injection (TI) algorithm etc. But each of these methodshas its own drawback. In the dissertation, we focus on some key technologies onreducing OFDM signal's PAPR.1. In order to overcome the PTS-based OFDM systems drawback, i.e. thecomputational complexity of PAPR reduction increases exponentially withthe number of sub-blocks, we modify the original arti?cial bee colony algo-rithm to solve the combinatorial optimization problems. Then we analysethe complexity of the algorithm. Simulation results verify the feasibil-ity and high e?ciency of the modi?ed arti?cial bee colony algorithm. Bycasting the originally combinatorial optimization problems for PTS-basedOFDM systems into an associate rare-event probability estimation, we usethe parametric minimum cross entropy algorithm to e?ciently solve theestimation. By simulation, the validity of the parametric minimum crossentropy algorithm is con?rmed.2. Two schemes of ?nding nearly optimal PRT set which are based on geneticalgorithm and particle swarm optimization are proposed to generate thepeak reduction tone set with lower secondary peak. Compared to the ex-isting algorithms, the complexities of the two schemes are very lower. Onthe basis of two schemes, we propose an adaptive clipping control algorithm to determine the optimal target clipping level, and analyse the complexityof the algorithm in detail. Simulation results demonstrate that our schemecan obtain better PAPR reduction regardless of the initial clipping level.3. A method based on hexagonal constellation is proposed to reduce PAPRand to eliminate increments in average transmitted power for OFDM sig-nals. The parametric minimum cross entropy algorithm is used to solve thesignal's optimal expression. In order to reduce the complexity, we proposean adaptive parametric minimum cross entropy algorithm to solve the sig-nal's optimal expression. Compared to the existing algorithms, simulationresults show that the two algorithms can obtain better PAPR reduction.4. We propose a new method to compute the continuous time PAPR forOFDM signals. Compared to the existing methods, the method overcomesthe drawback with larger computational load and improves computationale?ciency. Our method only needs to expand and simplify N ?1 Chebyshevpolynomials of the ?rst kind, and avoids the computation of the derivativeof Chebyshev polynomials of the ?rst kind. We also use the algorithm forpolynomial real roots isolation to e?ciently ?nd the real roots of polyno-mial.