文档介绍：第20卷第6期中国惯性技术学报 2012年12月 Journal of Chinese Inertial Technology Dec. 2012 收稿日期:2012-06-11;修回日期:2012-10-26 基金项目:航空科学基金(20100818018)作者简介:张华强(1982—),男,讲师,博士,从事组合导航和测控研究。E-mail:@文章编号:1005-6734(2012)06-0749-05 基于自适应概率神经网络的SINS/GPS组合导航系统故障诊断方法张华强1,2,赵剡2,连远峰2(,淄博 255049;2. 北京航空航天大学仪器科学与光电工程学院,北京 100191)摘要:为保障SINS/GPS组合导航系统稳定可靠地工作,提出采用自适应概率神经网络方法对组合导航系统进行故障诊断和隔离。采用状态χ2检测法对组合导航系统进行实时检测,提取状态χ2检测结果分量作为自适应概率神经网络的输入用于判断故障类型。采用高斯函数作为自适应概率神经网络的激励函数,为解决高斯核函数的宽度严重影响网络泛化能力的问题,使用交叉验证方法估算平滑因子,并采用粒子群优化算法优化平滑因子参数集。仿真结果表明,该算法能够准确地诊断出组合导航系统的故障并能确定故障发生的部件,保证了系统的安全可靠地运行。关键词:组合导航;自适应概率神经网络;粒子群优化算法;故障检测中图分类号: 文献标志码:A Adaptive probabilistic work fault-diagnosis algorithm for SINS/GPS integrated navigation system ZHANG Hua-qiang1,2, ZHAO Yan2, LIAN Yuan-feng2(1. Shandong University of Technology, Zibo 255049, Shandong, China; 2. Beijing University of Aeronautics and Astronautics, Beijing 100191, China) Abstract: In order to insure the SINS/GSP integrated navigation system work stably and reliably, a new method based on adaptive probabilistic work (APNN) was put forward to detect and isolate the fault of integrated navigation system. State chi-square test(SCST) was used to real-time monitor the navigation system. ponents of STSC were used as the input