Variational Bayesian Based IMM Robust GPS Navigation Filter

This paper investigates the navigational performance of Global Positioning System (GPS) using variational Bayesian (VB) based robust filter with interacting multiple model (IMM) adaptation as navigation processor. The state estimation for GPS processing family Kalman (KF) may be degraded due to fact that in practical situations statistics measurement noise might change. In proposed algorithm, adaptivity is achieved by estimating time-varying covariance matrices on VB learning probabilistic approach, where each update step, both system and were recognized random variables estimated. iterated recursively at time approximate real joint posterior distribution learning. One two major classical adaptive (AKF) approaches have been tuning estimate (MMAE). IMM algorithm uses or more filters process parallel, corresponds a different dynamic model. Huber's M-estimation-based extended (HEKF) integrates merits Huber M-estimation methodology EKF. robustness enhanced modifying method filtering framework. referred interactive multi-model HEKF (IMM-VBHEKF), provides an effective way effectively handling errors outlying property non-Gaussian interference errors, such multipath effect. Illustrative examples are given demonstrate enhancement terms expense acceptable additional execution time.