Hopfield Neural Networks for Vector Precoding

We investigate the application of Hopfield neural networks (HNN) for vector precoding in wireless multiple-input multiple-output (MIMO) systems. We apply the HNN to vector precoding with N transmit and K receive antennas, and obtain simulation results for the average transmit energy optimization as a function of the system load α = K/N. We compare these results with lattice search based precoding performances, and show that the proposed method for nonlinear vector precoding with complexity O(K 3 ) achieves competitive performances in the range 0 <α ≤ 0.9 in comparison to lattice search based precoders. The proposed method is of a polynomial complexity and therefore, it is an attractive suboptimal approach for vector precoding. I. INTRODUCTION

[1]  Andrea J. Goldsmith,et al.  Duality, achievable rates, and sum-rate capacity of Gaussian MIMO broadcast channels , 2003, IEEE Trans. Inf. Theory.

[2]  Johannes B. Huber.  Iterated Soft-Decision Interfer-ence Cancellation for CDMA , 1998 .

[3]  B.M. Zaidel,et al.  On Spectral Efficiency of Vector Precoding for Gaussian MIMO Broadcast Channels , 2008, 2008 IEEE 10th International Symposium on Spread Spectrum Techniques and Applications.

[4]  Andrzej Cichocki,et al.  Neural networks for optimization and signal processing , 1993 .

[5]  Adriano Lorena Inácio de Oliveira,et al.  A Hybrid Hopfield Network-Simulated Annealing approach to Optimize Routing Processes in Telecommunications Networks , 2007, Seventh International Conference on Intelligent Systems Design and Applications (ISDA 2007).

[6]  A. Lee Swindlehurst,et al.  A vector-perturbation technique for near-capacity multiantenna multiuser communication-part I: channel inversion and regularization , 2005, IEEE Transactions on Communications.

[7]  Amir K. Khandani,et al.  Communication Over MIMO Broadcast Channels Using Lattice-Basis Reduction , 2006, IEEE Transactions on Information Theory.

[8]  Björn E. Ottersten,et al.  On the complexity of sphere decoding in digital communications , 2005, IEEE Transactions on Signal Processing.

[9]  Pekka Orponen,et al.  On the Computational Complexity of Binary and Analog Symmetric Hopfield Nets , 2000, Neural Computation.

[10]  Masato Okada,et al.  Approximate belief propagation, density evolution, and statistical neurodynamics for CDMA multiuser detection , 2005, IEEE Transactions on Information Theory.

[11]  László Lovász,et al.  Factoring polynomials with rational coefficients , 1982 .

[12]  Josef A. Nossek,et al.  Linear transmit processing in MIMO communications systems , 2005, IEEE Transactions on Signal Processing.

[13]  Ue-Pyng Wen,et al.  A review of Hopfield neural networks for solving mathematical programming problems , 2009, Eur. J. Oper. Res..

[14]  Robert F. H. Fischer,et al.  Precoding in multiantenna and multiuser communications , 2004, IEEE Transactions on Wireless Communications.

[15]  Babak Hassibi,et al.  On the sphere-decoding algorithm I. Expected complexity , 2005, IEEE Transactions on Signal Processing.

[16]  Ralf R. Müller,et al.  Vector Precoding in High Dimensions: A Replica Analysis , 2007, 2007 IEEE International Symposium on Information Theory.

[17]  Robert F. H. Fischer,et al.  Precoding and Signal Shaping for Digital Transmission , 2002 .

[18]  Max H. M. Costa,et al.  Writing on dirty paper , 1983, IEEE Trans. Inf. Theory.

[19]  Jirí Síma.  Energy-Based Computation with Symmetric Hopfield Nets , 2007 .