一种基于Lagrange神经网络的多用户检测器

提出并讨论了一种基于Lagrange神经网络的多用户检测器,利用神经网络能有效地求解优化问题;推导了Lagrange神经网络多用户检测器(LNN-MUD)。理论分析和计算结果表明:在误比特性能和抗干扰性能上,该检测器均优于传统检测器和解相关检测器;在抗"远近"干扰能力方面,该检测器优于传统检测器而弱于解相关检测器,且易于实时应用和VLSI实现。

A Multiuser Detector Based on Lagrange Neural Network

According to the optimization theory and the neural network (NN) theory, a multiuser detector (MUD) is proposed, which takes the optimum MUD problem as combinatorial optimum problem. Using the neural network which has the ability of fast optimization computing, the Lagrange neural network (LNN) MUD is derived. Theoretical analysis and numerical results show that in aspect of bit-error rate and multiple access interfernce, the LNN MUD is better than the conventional and decorrelated MUD in aspect of "near-far" resistance, the LNN MUD is better than the conventional MUD and worse than the decorrelated detector and the LNN MUD can be easily implemented by VLSI technology.