基于一类SVM贝叶斯算法的DS-UWB系统多用户检测研究

在多径多址条件下，用户自身的前后符号之间会产生码间干扰，同时各用户之间会产生多址干扰，从而给用户信号的正确检测带来极大困难。提出了一种基于一类支持向量机贝叶斯分类器的DS-UWB系统多用户检测算法，证明了基于径向基核函数的一类SVM的分类函数归一化为密度函数，并将所得的概率密度函数用于构造贝叶斯分类器。仿真实验表明，在UWB信道环境下本算法的误码率性能明显优于最小均方误差（MMSE）检测和解相关检测等线性检测算法。相比传统SVM算法，本算法所需的核运算量和存储空间要小得多，有效地降低了运算负载，抑制了多址干扰。

DS-UWB multi-user detection based on one-class SVM Bayesian algorithm

A DS-UWB system multi-user detection scheme is proposed, which is based on one-class SVM Bayesian algorithm. It is proved that the solution of one-class SVM using the Gaussian kernel can be normalized as an estimate of probability density, and the probability density can be used to construct the two-class and multi-class Bayesian classifier. The simulation results show that in UWB channel, the BER performance of the proposed scheme was better than liner MUD algorithm, such as minimum mean square error (MMSE) detector and decorrelating detector. The size of kernel matrix of the new algorithm is greatly less than traditional multi class SVM, which leads to less training time and good performance of mitigating of MAI.