針對多天線感知無線網路考慮量化通道資訊回授誤差下之波束型成技術

Abstract

本篇論文針對現有的感知無線電網路系統，提出了一個低回授冗餘的多天線波束型成演算法。由於感知無線網路系統中次級使用者需透過主要使用者完美回授量化後的通道狀態資訊來設計波束型成，然而，在現實世界裡，若要保證完美回授勢必會造成主要使用者負擔過多的訊號冗餘。因此，在本篇論文中，我們放寬了完美通道回授的假設，考慮回授通道誤差對於通道狀態資訊的影響，設計出一個最佳的波束型成，降低主要使用者的回授冗餘，並且能夠提供與傳統波束型成演算法有相近的效能。在設計出最佳的波束型成後，我們亦著手分析此波束型成的效能，推導出次級接收端平均訊雜比的數學明確表示式。在模擬結果中能發現即使回授通道環境變差，本篇論文的最佳波束型成效能還是能保持不錯的效能，且在運算上相較於傳統的波束型成演算法來的更為簡單。此外，模擬結果亦驗證推導出的理論值與現實的模擬值吻合。

Multiple antenna technique have been considered to further improve spectrum efficiency and link quality of cognitive radio (CR) systems. Employing beamforming at secondary transmitter is one such solution in cognitive radio systems. In order to design CR beamforming, channel state information (CSI) of the cross-tier communication link must be known at the secondary transmitter. In the thesis, we study the problem of the limited feedback of the channel state information (CSI) for the cognitive radio networks. Motivated by the fact that the assumption of perfect CSI feedback requires close coordination between primary and secondary networks such as employing powerful channel coding which may incur excessive signaling overhead, we consider that the primary system can not afford to pay too much overhead. More specifically, assuming that CSI feedback is no longer perfect, then we derive the a posteriori average interference power based on the imperfect feedback CSI. A closed-form formula for the optimal CR beamforming is derived, and the performance of average received signal to noise ratio (SNR) is analytically characterized. Besides, the energy efficient beamforming is also considered on our proposed feedback link. The performance of the proposed scheme can be illustrated throughout the computer simulation.