基于免疫量子进化算法的惯性传感器信号重构  

Reconstruction of inerial sensor signal based on immune quantum evolutionary algorithm

作　　者：蒋行国[1,2] 罗珍珍 李海鸥[1,2] 欧少敏[3] 

机构地区：[1]桂林电子科技大学信息与通信学院,广西桂林541004 [2]桂林电子科技大学广西精密导航技术与应用重点实验室,广西桂林541004 [3]桂林电子科技大学信息科技学院,广西桂林541004

出　　处：《电子技术应用》2017年第10期132-136,共5页Application of Electronic Technique

基　　金：广西精密导航技术与应用重点实验室主任基金项目(DH201507)

摘　　要：针对惯性传感器信号的特点,提出一种基于免疫量子进化算法的正交匹配追踪重构方法。该方法以正交匹配追踪算法为核心,将免疫机制引入量子进化算法。首先,通过量子编码的叠加性构造抗体、免疫克隆操作实现种群扩张,以加速原子搜索进程,同时借助量子交叉操作避免算法陷入局部最优。然后,利用各次迭代选取的最佳匹配原子完成惯性传感器信号的重构,从而达到滤波的目的。仿真结果表明,在该算法下,静态信号的零漂值得到了改善,信噪比提高了10.48 dB,动态信号均方误差降低了28.551(″/s)。相同条件下,与现有重构算法相比,信号滤波效果提高的同时,重构时间均减少了4 s左右,最终实现了惯性传感器信号的实时性处理。For the characteristic in processing signal of inertial sensor, the immune quantum evolutionary algorithm combining with orthogonal matching pursuit algorithm was designed for inerial sensor signal. The orthogonal matching pursuit was the key of the method, and the immunologic mechanism was introduced into quantum evolutionary algorithm. Firstly, in order to accelerate atomic search processing, the superposition of quantum codes was used to construct immune body, and the immune clone operation was taking to expanding population. Quantum crossover operation avoided the algorithm falling into local optimum at the same time. Then inerial sensor signal was reconstructed by taking optimal matching atom at each iteration to reduce the noise effectively. The simulation results verify that static signal zero drift value was greatly improved, and SNR was increased by about 10.48 dB mostly, and the dynamic signal＇s mean square error was decreased 28.551 （″/s）. Compared with the traditional reconstruction algorithm under the same conditions, not only the noise was reduced effectively, but also time-consuming of the algorithm reduced by 4 s. Finally, the real-time processing of inertial sensor signal is achieved.

关 键 词：惯性传感器 正交匹配追踪 量子进化算法 实时处理 

分 类 号：TN911.7[电子电信—通信与信息系统] TP301.6[电子电信—信息与通信工程]