Performance Of Inertial Navigation System Research Articles

Robust Interactive Multimodel INS/DVL Intergrated Navigation System With Adaptive Model Set

A Robust Interactive Multiple Model (RIMM) algorithm with the adaptive model set is proposed to improve the performance of Inertial Navigation System (INS) and Doppler Velocity Logs (DVL) integrated navigation system under complex measurement environment with undesirable heavy-tailed non-Gaussian noise. Specifically, an improved Huber kernel function is applied to the robust error state Kalman filter (ESKF) for its benefit of better resisting larger measurement outliers. In addition, a flexible adaptive model set update strategy is proposed where the model set is determined by the current and historical measurement information stored in the sliding window. And for model set update, the main model is no longer fixed, it will be determined based on the probability weights corresponding to each model. This innovative RIMM algorithm is compared with error state Kalman filter, Interactive Multiple Model (IMM) and Hybrid Interactive Multiple Model (HIMM) through simulations, AUV lake trial and semi-physical simulations. Experimental results show that our proposed algorithm has outstanding accuracy and robustness under heavy-tailed non-Gaussian noise environment.

An Integrated INS/Lidar Odometry/Polarized Camera Pose Estimation via Factor Graph Optimization for Sparse Environment

For the outdoor scenes with sparse features, the performance of inertial navigation system (INS)/Lidar odometry (LO)-based simultaneous localization and mapping (SLAM) pose estimation is drastically affected by the accumulated drift error and sparse features. A polarized camera (Polcam) that can obtain the absolute yaw angle from the polarized skylight is introduced to compensate for the accumulated drift error. A yaw measurement model based on the polarized camera is derived, and it will not suffer from the sparse and GNSS-denied environment. Moreover, it is integrated with INS/LO to construct a factor graph. The factor graph that consists of INS factor, polarized camera factor, lidar odometer factor, and motion model factor is optimized by nonlinear least squares. For the sparse environmental features, a robust and effective method named correntropy-based bidirectional generalized ICP (CoBigICP) has been adopted to register the sparse point cloud data. A loosely coupled INS/LO/Polcam integration based on factor graph optimization for sparse scenes is proposed, and it is tested and validated by an open dataset and experiment in the real-world environment, respectively. Different integration modes with different register methods are compared both in the simulation and experiment. The results show that the INS/LO integration with Polcam via factor graph optimization can realize better pose estimation accuracy than without in a sparse environment.

INS’s Performance Evaluation Methods Based on K Line Theory

By comparing the similarity between stock trend analysis and performance evaluation of inertial navigation system (INS), the feasibility of INS's performance evaluation and prediction of the K line theory is analyzed. The K line and drawing for INS performance evaluation are defined; parts of the K line indexes and their calculation formula are redefined and transplanted. These methods are tested and verified by using simulated and the INS actual data. Analytical results indicate that the K line theory can be evaluated to predict the INS performance and the method is precise and effective. Research results have an important guiding significance to develop INS performance evaluation standards and demonstrate that economic theories can be used in engineering application with a certain internal relations. It would be a promising work which should be further investigated.

An EMD threshold de-noising method for inertial sensors

Random errors of inertial sensors are key factors in influencing the performance of Inertial Navigation System (INS). Based on underlying white noise model, classical wavelet threshold de-noising method is incapable of eliminating colored noise. Since time-correlated colored noise is predominant, fractional Gaussian noise (fGn) is utilized to model sensor errors and the Hurst parameter of fGn is estimated by the periodogram method. Variances of the noise in Intrinsic Mode Functions (IMFs) decomposed by Empirical Mode Decomposition (EMD) are analyzed. The standard deviations of noise in the first tow IMFs are estimated by a robust estimator, and then the noise variances in other IMFs can be obtained after the variance relation among the IMFs decomposed from fGn is derived. Noise thresholds of IMFs are estimated through the obtained variances and an EMD threshold de-noising method using order-dependent thresholds is established. The method is firstly verified by a simulation example and then applied in INS and compared with wavelet de-noising method. Results show that wavelet threshold de-noising is poor at suppressing colored noise while EMD threshold de-noising is effective on reducing sensor errors due to its close association with proper noise model.

Not Fully Overlapping Allan Variance and Total Variance for Inertial Sensor Stochastic Error Analysis

Stochastic errors characterize the performance of inertial sensors and indicate the potential improvements of the device. Accurate modeling of these errors can enhance the performance of inertial navigation system. The simplest and generally adopted method to model stochastic errors of inertial sensor is the normal nonoverlapped Allan variance, but its estimation accuracy decreases in long cluster time. The fully overlapping Allan variance improves the estimation accuracy in long cluster time very much, while the traditional total variance based on maximal-overlapping cluster samples further increases the estimation accuracy in long cluster time greatly. However, with these two methods, better estimation is achieved at the expense of much longer computation time. Besides, the computation burden for large dataset and many variance data points is extremely large, while the inertial sensors need large dataset with many variance data points to fully characterize their stochastic errors. This is because the correlation time for some stochastic error is very long (e.g., the correlation time is 3 h for rate random walk), and the Allan variance curve of some stochastic error is very complicated (e.g., sinusoidal noise). Consequently, the fully overlapping Allan variance and the traditional total variance are not suitable for inertial sensor stochastic error modeling. This paper proposes a not fully overlapping Allan variance which has similar estimation accuracy to fully overlapping Allan variance, and whose calculation time is greatly reduced relative to fully overlapping Allan variance. Then, the not fully overlapping concept is integrated to that of total variance to further improve the estimation accuracy in long cluster time with respect to Allan variance, and the calculation time of total variance is greatly reduced as well. This method enables high accuracy and high computational efficiency of Allan variance analysis at the same time, especially for large dataset and many variance data points analysis. Finally, the proposed methods are applied to 12-hour static data of gyroscopes and accelerometers from a position and orientation system, and their advantages are demonstrated.

Embedded UAV model and LASER aiding techniques for inertial navigation systems

This work proposes the integration of two aiding information sources to enhance the performance of inertial navigation systems (INS), for precise maneuvering of uninhabited air vehicles (UAVs). A new methodology is derived to exploit vehicle dynamics (VD) information by embeding the VD equations directly in an extended Kalman filter architecture. In this technique, the INS estimates are propagated simultaneously using the VD and the INS differential equations, allowing for the estimation of the INS errors. The proposed technique reduces the computational load associated with classical VD aiding, while retaining the same accuracy enhancements. A LASER range finder sensor is also integrated for precise landing and takeoff maneuvers. Simulation results for the nonlinear dynamics of a Vario X-Treme model-scale helicopter are presented, illustrating the contributions of the proposed aiding techniques for precise UAV navigation.

Analytical Modeling of Sensor Quantization in Strapdown Inertial Navigation Error Equations

Although generally not considered a major contributor to system inaccuracy, inertial sensor quantization error, if not properly modeled, can lead to erroneously large estimates of its impact on inertial navigation system performance. Analytical methods are described for modeling inertial sensor quantization in strapdown inertial system error parameter propagation and measurement equations. Error propagation equations are derived in classical differential error state dynamic and discrete difference format. It is shown how the attitude, velocity, and sensor error parameters in these equations must be modified to enable proper sensor quantization error modeling as white noise and to account for differences in attitude, velocity, and position update frequencies. The discrete difference error equation form is used to develop values for attitude/velocity measurement noise covariances and for spectral densities of white quantization noise terms in the differential error propagation equations. A general discussion is included of how quantization white noise spectral densities should be computed for the differential error propagation equations for compatibility with two-speed attitude, velocity, and position updating algorithms. Validity limits for white noise modeling approximations and methods for reducing quantization noise are discussed. Numerical examples are provided.