Robust adaptive UKF based on SVR for inertial based integrated navigation

Meng-de Zhang,Hai-fa Dai,Bai-qing Hu,Qi Chen
DOI: https://doi.org/10.1016/j.dt.2019.10.012
IF: 4.035
2020-08-01
Defence Technology
Abstract:<p>Aiming at the problem that the traditional Unscented Kalman Filtering (UKF) algorithm can't solve the problem that the measurement covariance matrix is unknown and the measured value contains outliers, this paper proposes a robust adaptive UKF algorithm based on Support Vector Regression (SVR). The algorithm combines the advantages of support vector regression with small samples, nonlinear learning ability and online estimation capability of adaptive algorithm based on innovation. Firstly, the SVR model is trained by using the innovation in the sliding window, and the new innovation is monitored. If the deviation between the estimated innovation and the measured innovation exceeds a given threshold, then measured innovation will be replaced by the predicted innovation, and then the processed innovation is used to calculate the measurement noise covariance matrix using the adaptive estimation algorithm. Simulation experiments and measured data experiments show that SVRUKF is significantly better than the traditional UKF, robust UKF and adaptive UKF algorithms for the case where the covariance matrix is unknown and the measured values have outliers.</p>
engineering, multidisciplinary
What problem does this paper attempt to address?
### Problems Addressed by the Paper This paper primarily addresses the issues in the traditional Unscented Kalman Filter (UKF) algorithm within the Strapdown Inertial Navigation System (SINS) when the measurement covariance matrix is unknown and the measurements contain outliers. Specifically: 1. **Unknown Measurement Covariance Matrix**: - In practical applications, due to complex and variable environmental conditions, it is challenging to accurately obtain the covariance matrix of measurement noise. 2. **Measurements Containing Outliers**: - In urban environments, satellite signals may be blocked or interfered with; in underwater environments, changes in water flow speed may lead to abnormal water speed measurements. To tackle these issues, the authors propose a robust adaptive UKF algorithm based on Support Vector Regression (SVR), termed SVRUKF. This algorithm combines the advantages of SVR in handling small samples, its nonlinear learning capability, and the online estimation capability of adaptive algorithms. ### Algorithm Features - **Automatic Outlier Detection**: By using the SVR model, the algorithm can automatically detect and replace the current innovation value, rather than simply reducing the weight of outliers. - **Enhanced Anti-Interference Capability**: The algorithm enhances anti-interference capability without reducing filtering accuracy. ### Experimental Validation - Through simulation experiments and shipborne experiments, it is demonstrated that the SVRUKF algorithm significantly outperforms traditional UKF, robust UKF, and adaptive UKF algorithms when the measurement covariance matrix is unknown and the measurements contain outliers. ### Summary This paper aims to improve the robustness and adaptability of the Strapdown Inertial Navigation System by introducing Support Vector Regression technology to the traditional UKF algorithm.