基于量子粒子群优化的数字冲击滤波器自动设计  

QPSO-based auto-design of digital impacting filters

作　　者：余静[1] 吴乐南[1] 靳一[1] 

机构地区：[1]东南大学信息科学与工程学院,南京210096

出　　处：《东南大学学报（自然科学版）》2012年第2期224-228,共5页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金资助项目(60872075);国家高技术研究发展计划(863计划)资助项目(2008AA01Z227)

摘　　要：为了提高数字冲击滤波器设计的效率,提出了一种基于QPSO算法的优化设计方法.首先根据EBPSK通信系统中数字冲击滤波器的特殊滤波机理,在考虑有限字长效应的情况下对数字冲击滤波器进行建模,然后设计出合理的优化变量、目标函数和约束条件,并利用QPSO算法的全局搜索能力和惩罚函数法的约束处理机制设计数字冲击滤波器.仿真结果表明:种群适应度值能够逐步达到收敛,从而保证了寻优结果的最优性;设计得到的滤波器幅频响应在中心频率处呈现出极窄的陷波-选频特性,并可将EBPSK调制信号的微弱相位跳变转化为明显的幅度冲击,有利于解调.与已有的手调滤波器相比,该设计方法快速可靠,可节约0.3～0.5 dB的信噪比增益.To improve the efficiency of designing digital impacting filters, a novel method based on the quantum-behaved particle swarm optimization （QPSO） algorithm is proposed. First, according to the special filtering mechanism of the digital impacting filter in the communication system of ex- tended binary phase shift keying （ EBPSK）, the model of the digital impacting filter is built consid-ering the effect of finite word length. Then the constraint conditions are designed. The global optimization variables, the objective function and the search ability of the QPSO algorithm and the con- straint mechanism of the penalty method are utilized to design the digital impacting filter. The simu- lation results show that the fitness value is gradually converged to make sure that the optimization re-suits are optimal. The amplitude response of the designed filter matches the digital impacting filter＇s characteristic of notch-select frequency in the carrier frequency, which can transform the phasehopping of the EBPSK signal into the amplitude impact so as to be good for demodulation. Com- pared with the existing impacting filters by manual design, the proposed design method is fast and re- liable. It can save the signal-to-noise（SNR） gain of 0. 3 to 0. 5 dB.

关 键 词：数字冲击滤波器 QPSO算法 EBPSK信号 有限字长效应 信噪比 

分 类 号：TN911[电子电信—通信与信息系统]