English

Abstract

In next-generation high throughput satellite (NGHTS), precoding (a preprocessing technique at the gateway (GW) can mitigate the interbeam interference with no additional resource at the user terminal. However, the calculation of precoding requires matrix inversion, which involves huge complexity in terms of hardware resources with real-time calculation for near error-free results. This paper proposes a novel approach for precoding matrix calculation using the Neumann series. In this way, the algorithm does not directly compute the matrix inversion. Hence the computational complexity is highly reduced. The proposed method uses the 4th-order iterative series with the initial guess as an inverse of the diagonal matrix of the input to construct the Neumann series. This leads to fast convergence of the matrix inversion process with fewer resources. The algorithm is claimed to be generic as it is seamlessly applied to any linear precoding scheme. The paper examines the resource complexity, convergence probability and Bit Error Rate (BER) performance for the proposed method with Zero Forcing (ZF) and Regularized Zero Forcing (RZF) based linear precoding scheme. Experiment results demonstrate that the proposed algorithm accomplishes superior performance with fewer resources compared to the Neumann series and Joint iterative Newton/Chebyshev-based Neumann series methods.