FD-IVA Algorithm for Eliminating CCI in the Uplink of LEO Satellite MIMO Communication System with FFR

The use of full frequency reuse (FFR) for low-earth orbit (LEO) satellite multiple-input multiple-output (MIMO) communication systems can further improve spectral efficiency, but it also leads to serious co-channel interference (CCI). Compared with land systems, LEO satellites’ high dynamism and long transmission delay pose challenges in eliminating CCI due to the difficulty in obtaining accurate channel state information (CSI). This paper adopts a blind source separation (BSS) algorithm that does not require accurate CSI and avoids the extra occupation of channel resources to eliminate CCI. LEO satellite MIMO communication system with FFR has convolutional mixing characteristics. The inherent sorting ambiguity of the BSS algorithm leads to the difficulty of signal reconstruction after converting frequency-domain de-mixing signals to time-domain. This paper proposes a Fast Decoupled-Independent Vector Analysis (FD-IVA) algorithm. Firstly, FD-IVA algorithm uses IVA algorithm to solve the sorting ambiguity problem in the demixing process. Then, the iteration step size of each row of the separation matrix is optimized separately by decoupling to reduce the number of iterations of the de-mixing algorithm. Finally, the optimization of the initial matrix is achieved through joint approximative diagonalization of the eigenmatrix (JADE) algorithm to further improve the convergence speed of the de-mixing algorithm. Simulation analysis shows that the proposed FD-IVA algorithm has the best performance of CCI elimination, and the optimized initial separation matrix can obtain a better initial iteration point, significantly reducing the number of iterations.