非高斯噪声环境下的智能粒子滤波多用户检测方法  

Multi-user detection method based on intelligenceparticle filter under non-Gaussian noise

作　　者：彭涛[1,2] 李一兵[2] 高振国[3] 

机构地区：[1]海军驻哈尔滨地区舰船配套军代表室,黑龙江哈尔滨150001 [2]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001 [3]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001

出　　处：《应用科技》2011年第9期15-18,22,共5页Applied Science and Technology

基　　金：国家自然科学基金资助项目(60703090);船舶工业国防科技基础研究基金资助项目(10J3.1.6)

摘　　要：粒子滤波适用于任何非线性非高斯系统的状态估计问题,具有应用灵活、适用范围广等优点.建议分布的选择恰当与否直接决定着粒子滤波的估计精度和估计效率.针对这一难点提出了采用粒子群优化算法来确定粒子的建议分布.粒子群优化算法作为新的群智能算法同样适应于各类非线性非高斯系统,采用该算法确定粒子滤波的建议分布保证了粒子滤波广泛的适应性,同时提高了估计精度.最后在Alpha稳定分布噪声环境下对CDMA系统多用户检测进行了仿真,结果表明,采用智能算法来确定粒子的建议分布极大地提高了粒子滤波的估计精度.The particle filter is applicable to any kind of state estimation in nonlinear non-Gaussian system, and it has a wide range of application and flexibility. The choice of the proposed distribution directly determines the estimation accuracy and estimation efficiency. Aiming at the difficulty, a particle swarm optimization particle filter is proposed to determine the proposed distribution. The particle swarm optimization algorithm as a new intelligence al- gorithm also adapts all types of non-linear non-Gaussian systems, using the algorithm to determine the proposed dis- tribution of particle filter ensures a wide range of application of particle filter and improves the estimation accuracy. Finally, simulation of multi-user detection of CDMA system was realized based on particle filter in Alpha stable noise environment. The results showed that using intelligent algorithms to determine the proposed distribution of the particle filter greatly improved the estimation accuracy.

关 键 词：粒子滤波 多用户检测 群智能算法 粒子群优化 非高斯噪声 

分 类 号：TN92[电子电信—通信与信息系统]