计算机应用 ›› 2016, Vol. 36 ›› Issue (12): 3358-3362.DOI: 10.11772/j.issn.1001-9081.2016.12.3358

• 人工智能 • 上一篇    下一篇

基于智能粒子滤波的多传感器信息融合算法

陈伟强1, 陈军1, 张闯2, 宋立国2, 谭卓理3   

  1. 1. 中海油田服务股份有限公司 船舶事业部, 河北 三河 065201;
    2. 大连海事大学 航海学院, 辽宁 大连 116026;
    3. 大连船舶重工船业有限公司, 辽宁 大连 116052
  • 收稿日期:2016-05-20 修回日期:2016-06-20 出版日期:2016-12-10 发布日期:2016-12-08
  • 通讯作者: 张闯
  • 作者简介:陈伟强(1964-),男,广东龙川人,高级船长,主要研究方向:组合导航、智能控制;陈军(1970-),男,天津人,高级船长,主要研究方向:组合导航;张闯(1980-),男,辽宁昌图人,讲师,博士研究生,主要研究方向:组合导航;宋立国(1983-),男,山东德州人,讲师,博士研究生,主要研究方向:组合导航;谭卓理(1976-),男,山东德州人,工程师,主要研究方向:组合导航。
  • 基金资助:
    国家自然科学基金资助项目(61374114);中央高校基本科研业务费专项资金资助项目(3132016005)。

Multisensor information fusion algorithm based on intelligent particle filtering

CHEN Weiqiang1, CHEN Jun1, ZHANG Chuang2, SONG Liguo2, TAN Zhuoli3   

  1. 1. Division of Marine and Transportation Service, China Oilfield Services Company Limited, Sanhe Hebei 065201, China;
    2. Navigation College, Dalian Maritime University, Dalian Liaoning 116026, China;
    3. Dalian Shipbuilding Industry Company Limited, Dalian Liaoning 116052, China
  • Received:2016-05-20 Revised:2016-06-20 Online:2016-12-10 Published:2016-12-08
  • Supported by:
    This work is partially supported by the National Natural Science Foundation of China (61374114), the Fundamental Research Funds for the Central Universities (3132016005).

摘要: 针对粒子滤波中存在粒子质量低和粒子贫化的问题,提出了一种基于智能粒子滤波的多传感器信息融合算法。该算法分为两个模块,首先,将多传感器数据发送给相应的粒子滤波计算模块,以优化粒子分布为目的更新建议分布密度;然后,在智能粒子滤波模块中对多传感器数据构造完整的似然函数,引入设计的遗传因子将小权重粒子修正为大权重粒子,近似于真实的后验分布,重采样过程中保留了权值较大的粒子,又避免了粒子耗尽问题,进一步保持粒子的多样性,提高了滤波精度,最终得到最优的估计值。根据实船实验的数据进行了验证,将所提算法应用于GPS/SINS/LOG组合导航系统进行仿真计算验证了其有效性。实例仿真结果表明,所提算法能够得到精确的位置、速度和航向信息,而且也能有效改善滤波性能,提高了组合导航系统的解算精度,能够满足船舶高精度导航定位的要求。

关键词: 多传感器, 遗传因子, 粒子滤波, 信息融合

Abstract: In order to solve the low-quality and degeneration problem of particles in the process of particle filtering, a multisensor information fusion algorithm based on intelligent particle filtering was proposed. The process of the proposed algorithm was divided into two steps. Firstly, the multisensor data was sent to the appropriate particle filtering calculation module, and the proposal distribution density was updated for the purpose of optimizing the particle distribution. Then, the integrated likelihood function model was structured by using the multisensor data in intelligent particle filtering module, meanwhile, the small-weight particles were modified into large-weight ones according to the designed genetic operators. The posterior distribution was more sufficiently approximated, thus large-weight particles were reserved in the process of resampling, which avoided the problem of exhausting particles, further maintained the diversity of the particles and improved the filtering precision. Finally, the optimal accurate estimated value was obtained. The proposed algorithm was applied to the GPS/SINS/LOG integrated navigation system according to the prototype testing data, and its effectiveness was verified by the simulation calculation. The simulation results show that, the proposed algorithm can get accurate informations of location, speed and heading, and effectively improve the filtering performance, which can improve the calculating precision of the integrated navigation system and meet the requirement of high precision navigation and positioning of the ship.

Key words: multisensor, genetic operator, particle filtering, information fusion

中图分类号: