Joint conjugate gradient and Jacobi iteration based low complexity precoding for massive MIMO systems
| Autor: | Xuming Lu, Liangliang Wang, Song Wei, Xiang Chen |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
Mathematical optimization
Computational complexity theory Preconditioner Iterative method Computer science 020208 electrical & electronic engineering Jacobi method 020206 networking & telecommunications 02 engineering and technology Precoding symbols.namesake Matrix (mathematics) Rate of convergence Conjugate gradient method 0202 electrical engineering electronic engineering information engineering symbols Series expansion Gradient descent Algorithm |
| Zdroj: | ICCC |
| DOI: | 10.1109/iccchina.2016.7636727 |
| Popis: | In massive multiple-input multiple-output (MIMO) downlink systems, the computational complexity of the matrix inversion for precoding is becoming a bottleneck for the system implementation, along with the increased antenna dimension at the base station and the number of served users. By traditional approximate matrix inversion algorithms, for example, series expansion, gradient or iteration methods, the proper system performance can be achieved by lower complexity for each recursive expansion or iteration, while keeping lower convergence rate. In this paper, a novel joint conjugate gradient (CG) and Jacobi (JC) iteration method is proposed to speed up the convergence rate as well as guaranteeing the bit error rate (BER) performance. In this proposal, the CG is firstly used twice to find a more accurate searching direction for the following low complexity JC iterations. Numerical simulations and complexity analysis show that, with lower computational complexity, our proposal can obviously outperform previous nonjoint iteration methods. Even compared with recently proposed joint steepest descent and Jacobi method in [7], our proposal can obtain the similar or even better BER performance with faster convergence rate, which indicates lower total hardware implementation complexity and latency. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|