GPS/UWB/MEMS-IMU tightly coupled navigation with improved robust Kalman filter

详细信息    查看全文

• 作者：Zengke Li^a ; ^b ; ^{zengkeli@yeah.net" class="auth_mail" title="E-mail the corresponding author} ; Guobin Chang^a ; ^{guobinchang@hotmail.com" class="auth_mail" title="E-mail the corresponding author} ; Jingxiang Gao^a ; ^{jxgaocumt@yeah.net" class="auth_mail" title="E-mail the corresponding author} ; Jian Wang^a ; ^{wjiancumt@yeah.net" class="auth_mail" title="E-mail the corresponding author} ; Alberto Hernandez^a ; ^{albert_h_m@hotmail.com" class="auth_mail" title="E-mail the corresponding author}
• 关键词：Ultra-wideband ; Tightly coupled navigation ; Robust Kalman filter ; DGPS
• 刊名：Advances in Space Research
• 出版年：2016
• 出版时间：1 December 2016
• 年：2016
• 卷：58
• 期：11
• 页码：2424-2434
• 全文大小：1386 K

文摘

The integration of Global Positioning System (GPS) with Inertial Navigation System (INS) has been very intensively developed and widely applied in multiple areas. To further enhance the reliability and availability of GPS/INS integrated navigation in GPS challenging environment, range observation through ultra-wideband (UWB) is introduced in GPS/INS tightly coupled navigation. An improved robust Kalman filter is proposed and used to resist the influence of gross error from UWB observation in GPS/UWB/IMU tightly coupled navigation. The variance of the squared Mahalanobis distance in moving window is calculated, which brings as new judgement factor for gross errors in order to decrease the rate of false outlier identification. A simulation analysis shows that the improved robust Kalman filter is able to correctly identify gross errors and the rate of false judgment as zero. In order to validate the new robust filter, a real experiment is conducted. The results indicate that the improved robust Kalman filter used in GPS/UWB/INS tightly coupled navigation is able to remove the harmful effect of gross error in UWB observation. It clearly illustrates that the improved robust Kalman filter is very effective, and all the simulated small and large gross errors added to UWB distance observation are successfully identified.