Minimum-output-energy method for blind equalization of OFDM and systems with sufficient or insufficient cyclic prefix

In this paper, a new blind equalization method is proposed for systems with cyclic prefix such as OFDM. With constrained minimumoutput-energy (MOE) optimization, a bank of equalizers can be derived without the help of FFT/IFFT. Algorithms are developed for blindly estimating the constraints and equalizers. When the cyclic prefix is sufficient, i.e., longer than the channel length, the complexity of the method is comparable to traditional FFT-based receivers. A special property is that the new method works similarly even if the cyclic prefix is insufficient, i.e., shorter than the channel length. Therefore, it provides an effective way to improve bandwidth efficiency with shorter cyclic prefix length. Simulations demonstrate the superior performance of the proposed method.