基于群智能算法的盲源分离问题的研究

发布时间：2018-01-25 06:28

本文关键词： 盲源分离粒子群算法帝国竞争算法变异算子负熵　出处：《东北大学》2014年硕士论文　论文类型：学位论文

【摘要】：盲信号分离技术是一种仅利用观测信号恢复出源信号的方法,在语音处理、图像处理、通信和生物医学信号处理等各个领域均具有广阔的应用前景和发展潜力,对其研究已成为目前信号与信息处理、智能计算与信息处理等学科的研究热点。群智能优化算法作为一种模拟自然界生物体生存发展的行为方式进行目标优化的算法,是求解复杂优化问题的有效方法。因此,将群智能优化算法用于解决盲信号分离问题具有良好的发展前景。本文在对群智能优化算法和盲信号分离技术理论和算法研究的基础上,针对线性混合盲信号分离进行深入研究,所作的主要工作如下：(1)针对基本粒子群算法易于陷入局部最优解的问题,提出了一种基于变异算子的自适应粒子群优化算法。并采用信号的负熵的绝对值作为分离的目标函数,使用带变异算子的自适应粒子群优化算法对目标函数进行优化求解,从而实现从混合信号中将源信号分离出来。通过对源信号为亚高斯信号以及超高斯和亚高斯信号同时存在的情况进行分离的仿真实验,并通过相似系数和信噪比两种性能评价准则与基本粒子群算法进行比较,验证了算法的有效性。(2)为了改善基本帝国竞争算法易于陷入局部最优解的缺点,提出了一种基于混沌理论的帝国竞争算法。并采用信号的互信息作为目标函数,使用基于混沌理论的帝国竞争算法对目标函数进行优化求解,从而实现对混合信号的分离。通过仿真实验验证了算法的有效性,并通过相关系数和信噪比两种性能评价指标与基本帝国竞争算法进行比较,体现了改进算法在全局收敛性和稳定性的优势。
[Abstract]:Blind signal separation (BSS) is a method to recover source signals only by using observed signals. It has broad application prospects and development potential in speech processing, image processing, communication and biomedical signal processing. Its research has become the current signal and information processing. Intelligent computing and information processing and other disciplines of research hotspot. Swarm intelligence optimization algorithm as a simulation of natural organisms survival and development of the behavior of the optimization algorithm. Is an effective method for solving complex optimization problems. The application of swarm intelligence optimization algorithm to solve blind signal separation problem has a good prospect. In this paper, the theory and algorithm of swarm intelligence optimization algorithm and blind signal separation technology are studied. The main work of this paper is as follows: 1) aiming at the problem that elementary particle swarm optimization (PSO) is easy to fall into local optimal solution. An adaptive particle swarm optimization algorithm based on mutation operator is proposed, and the absolute value of the negative entropy of the signal is used as the separated objective function. An adaptive particle swarm optimization algorithm with mutation operator is used to solve the objective function. In order to achieve the separation of the source signal from the mixed signal, the simulation experiment is carried out to separate the source signal from the sub-#china_person0# signal and the superelev_person1# and ultra-high signal simultaneously. By comparing the performance evaluation criteria of similarity coefficient and signal-to-noise ratio with the basic particle swarm optimization algorithm, the effectiveness of the algorithm is verified. 2) in order to improve the basic imperial competition algorithm, it is easy to fall into the local optimal solution. An algorithm of imperial competition based on chaos theory is proposed, and the mutual information of signals is used as the objective function, and the algorithm based on chaos theory is used to optimize the solution of the objective function. The effectiveness of the algorithm is verified by simulation experiments, and the correlation coefficient and signal-to-noise ratio (SNR) are compared with the basic imperial competition algorithm. The advantages of the improved algorithm in global convergence and stability are demonstrated.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.7

【相似文献】