基於星座圖整形之峰均值功率比降低方法

Abstract

在一具有多通道系統中，空時編碼經常被使用來達到最佳分集與多工取捨，然而，這些空時編碼對信號星座點的形變，往往導致輸出信號產生相對高的峰均功率比，使得訊號經過功率放大器時易產生非線性失真，而造成傳送端的放大器設計上的困難性增加。利用可逆整數對應的方式，則可對信號星座點加以整形，使得峰均功率比降低，並可維持分集與多工最佳取捨。然而，此對應技術運用於其他前置編碼多天線與正交分頻多工系統時，在數元錯誤率會有效能上的損失。因此，引入適合格柵點之解調之列舉球面演算法及更進ㄧ步的反覆解碼，使效能接近最佳理論值。

In a multichannel system, space-time codes are usually used to achieve the optimal diversity-multiplexing gain (D-MG) tradeoff. Nevertheless, these pace-time codes will make transformations on signal constellation, and often lead to relatively higher peak to average power ratio (PAPR) so that the signal will suffer nonlinear distortion through power amplifier. The design difficulty for transmit power amplifier will undoubtedly increase. By the use of integer reversible mapping method, the signal constellation can be shaped to lower down the PAPR without losing the optimal diversity-multiplexing gain (D-MG) tradeoff. However, in other precoding multi-input multi-output (MIMO) or orthogonal frequency division multiplexing (OFDM) system, this technique will lead to performance loss in bit error rate. Thus, we will introduce the list sphere decoding for closest lattice point search and a further iterative decoding to make the performance near to the theoretically optimal value.