基于自适应惯性权重粒子群优化的多跳频信号盲源分离

针对跳频通信中多跳频信号的盲源分离问题，提出了一种基于自适应惯性权重粒子群的盲源分离算法。该算法将分离信号的负熵作为目标函数，依据迭代前后每个粒子适应度值间差值自适应地调节惯性权重。把适应度值变差的粒子惯性权重设成零，以消除惯性分量不利影响，这样可以减少无效迭代次数，提高收敛速度。应用于盲源分离时，比经典算法分离效果好且克服了激活函数选取难题。实验结果表明该算法用于多跳频信号盲分离时性能稳定且收敛速度快，与经典算法比较优势明显，为智能算法在盲源分离方面的研究提供了一定的参考。

Blind source separation of multi-frequency-hopping signals based on adaptive inertia weight PSO algorithm

An adaptive inertia weight particle swarm optimization ( PSO ) based blind source separation method is proposed for multi-frequency-hopping signals in frequency hopping communication. This algo-rithm takes the negentropy of mixtures as an objective function to analyze the motion of each particle. After each iteration,the inertia weight of each particle is adjusted adaptively,according to the difference between the former fitness value and latter of each particle iteration. The particle’ s inertia weights are reset to zero, whose fitness value has become worse. The adverse effects of particle’ s inertia component on separation can be eliminated in the next iteration. It can reduce the number of invalid iterations,achieve blind source sep-aration and accelerate convergence speed. When applied to blind source separation, it is better than the classical algorithms and overcomes the select problem of the activation function. Simulation results show that the proposed method can make the performance stable and achieve rapid convergence in multi-frequency-hopping signal separation. Compared with the classical algorithms,this algorithm has advantage obviously. It has a certain reference value for the study of intelligent algorithm in blind source separation field.