Researches On Nonlinear Distortion Problem In The OFDM System

Orthogonal frequency division multiplexing (OFDM) technique is widely adopted in the high-speed broadband communication systems since it can turn the complex time-domain equalizer into the simple frequency domain equalizer (FDE). However, OFDM technique has the drawback of a high peak-to-average ratio (PAR) meanwhile needs the high-speed and high-precision analog-to-digital converter (ADC), which will lower the efficiency of the power amplifier (PA) and increase the power dissipation of the ADC. It is demonstrated that employing the soft limiter (SL) and the ADC with the low bitwidth can effectively solve this problem. But the SL distortion (i.e. limiter distortion) and the uniform symmetric quantizer (USQ) distortion (i.e. quantizer distortion) will be introduced at the transmitter and the receiver, respectively.The researches on the limiter distortion and quantizer distortion are performed and the state-of-the-art researches on these topics are reviewed in this dissertation. It is found that the current researches are only focused on the nonlinear distortion, not conducted from the viewpoint of the end-to-end link. Therefore, from the perspective of the end-to-end link of the OFDM system, the theoretical outcomes of the enhancement and optimization of the nonlinearly distorted OFDM system performance are obtained. They can be used as the guidance for the analysis and design of the OFDM system.Firstly, the signal-to-distortion and noise ratio (SDNR) quantifying the performance of the nonlinearly distorted OFDM system is defined and the approximate expression evaluating the bit-error-rate (BER) performance of the iterative decoding algorithm (IDA) is proposed thus circumventing the software simulation needed in the literature. Simulation results also verify that the approximate expression is reasonable. Since extra spectrum is required in the IDA to convey the clipping ratio (CR) parameter to receiver, therefore, an algorithm based on the BER of the pilot subcarrier is proposed. The algorithm considers the CR estimation problem as the rooting problem of a monotonic function, which can be solved with the bisection method. Simulation results validate the accuracy of the proposed algorithm.Secondly, the optimal quantization step-size of the USQ achieving the maximum channel capacity of the quantized OFDM system is acquired. The meaning of the optimality is presented. With the USQ possessing the optimal quantization step-size, a subcarrier reservation (SR) method is proposed to suppress the quantizer distortion. The relationship between the loss of the channel capacity and the BER gain is analyzed with the frequency domain model of the quantized signal. Simulation results show that the proposed algorithm is effective and the frequency domain model of the quantized signal is sufficiently accurate.Finally, the above results are extended to the quantized OFDM system with the clipping distortion at the transmitter. The channel capacity of the clipped and quantized OFDM system is calculated, which is the two-variable function with respect to the clipping ratio and the quantization step-size. Through selecting the PAPR gain and channel capacity as the optimization objectives, their Pareto fronts (PF) can be reached with the genetic multi-objective optimization algorithm.