The integration of the inertial navigation system and the global positioning system (INS/GPS) is a widely used procedure for position and attitude determination applications. The Kalman type filter (KTF) is the primary mechanism to perform the integration. In the KTF, the process noise is always assumed to be Gaussian distribution, which may be violated by the vehicle’s severe maneuver, resulting in a much degraded performance. In this paper, the Huber’s M-estimation methodology is investigated to suppress the process uncertainty, founded on the cascaded form of the M-estimation-based Kalman filter. An iterated algorithm is designed to construct the weighted matrix to rescale the prior state estimate covariance. The proposed process uncertainty robust algorithm is embedded into the newly derived modified unscented quaternion estimator to perform the standard inertial navigation equations-based INS/GPS integration. The car-mounted experiments are carried out to validate the proposed method against the process uncertainty.

Huber’s M-Estimation-Based Process Uncertainty Robust Filter for Integrated INS/GPS

Precise and robust localization in a large-scale outdoor environment is essential for an autonomous vehicle. In order to improve the performance of the fusion of GNSS (Global Navigation Satellite System)/IMU (Inertial Measurement Unit)/DMI (Distance-Measuring Instruments), a multi-constraint fault detection approach is proposed to smooth the vehicle locations in spite of GNSS jumps. Furthermore, the lateral localization error is compensated by the point cloud-based lateral localization method proposed in this paper. Experiment results have verified the algorithms proposed in this paper, which shows that the algorithms proposed in this paper are capable of providing precise and robust vehicle localization.

https://www.mdpi.com/1424-8220/17/9/2140/pdf

A Robust Vehicle Localization Approach Based on GNSS/IMU/DMI/LiDAR Sensor Fusion for Autonomous Vehicles

In this paper, a computational efficient attitude estimation method is proposed for the low-cost attitude and heading reference systems. In the proposed method, the velocity and position provided by the Global Positioning System and inertial sensors outputs are first used to construct the vector observations. With the constructed vector observations, an error equations based filtering model is established using the Euler angles as the attitude parameterization. If the attitude has been well initialized, the established model can reduce to a linear state-space model, which enables the application of standard Kalman filtering. For the established attitude estimation model, an indirect Kalman filter is detailedly designed. Car-mounted filed test results demonstrate that the proposed method possesses superiority over the existing methods with consideration of both accuracy and efficiency.

Indirect Kalman Filtering Based Attitude Estimation for Low-Cost Attitude and Heading Reference Systems

Due to the complementary nature of inertial navigation system (INS) and Wi-Fi positioning principles, an INS/Wi-Fi integrated system is expected to form a low-cost and continuous indoor navigation solution with better performance than using the standalone systems. In this paper, we explore the integration of Wi-Fi measurements with data from microelectromechanical systems (MEMS) based inertial measurement unit (IMU) for indoor vehicle navigation. Two enhancements, which employ adaptive Kalman filtering (AKF) and vehicle constraints, for supporting the integrated system are presented. One field experiment has been conducted for estimating the trajectory of a mobile robot vehicle. The numerical results show that the enhanced integrated system provides higher navigation accuracy, compared to using standalone Wi-Fi positioning and conventional INS/Wi-Fi integration.

https://www.zess.uni-siegen.de/cms/papers/Chai2012INSWiFibasedIndoor.pdf

INS/Wi-Fi based indoor navigation using adaptive Kalman filtering and vehicle constraints

Advanced state estimation for navigation of automated vehicles

PS receiver has dominated the field of positioning and navigation for decades [1]. However, its performance depends on the signal environments. It provides a continuous navigation solution only when more than four satellites are in view. In order to solve this problem, other navigation systems, e.g., inertial navigation system, are often employed to integrate with GPS for having a robust and continuous navigation solution. Historically, due to the high cost in manufacturing the inertial sensors, the INS/GPS integration system are mostly employed in military and aerospace industry [2, 3]. Recently, the advent of micro-electromechanical systems (MEMS) technology drives the discrete, heavy and inflexible inertial sensor system to small, cost-effective, light-weight, portable and lowpower silicon-based inertial devices. Although the cheap MEMS-based inertial sensors do not exhibit high accurate navigation performance, they can meet the requirements of many land-based navigation applications when aided with GPS devices. An INS/GPS system combines the advantages of both sides and provides accurate and uninterrupted navigation results, working in all environments, and constituting a potential and powerful alternative to the GPS alone navigation devices. Nowadays, the INS and GPS integrated solutions are the back-bones of many modern navigation systems, which are employed in industrial and military applications. Substantial research effort has been devoted to extensive algorithmic developments and performance analysis. The objective is mainly at the promotion of system estimation accuracy with low-cost sensor systems, putting a focus of interest onto powerful sensor fusion algorithms. The so-called tightly-coupled integration is one of the approaches to fuse the INS and GPS measurements. However, when modeling the underlying problem, the system propagation and observation models are nonlinear. The most common application of the Kalman filter (KF) on nonlinear systems is the extended (or linearized) Kalman filter (EKF) [4-6], which is based on a first-order linearization of the nonlinear stochastic system models with the assumption of Gaussian distributed noises. Although the EKF maintains the elegant and computationally efficient update form of the KF, it suffers from a number of drawbacks. That is, the linearized transformations are reliable, only if the error propagation can be well approximated by a linear function, because the small error

Tightly-coupled INS/GPS using Quaternion-based Unscented Kalman filter

Applying Quaternion-based Unscented Particle Filter on INS/GPS with Field Experiments

In this paper, two quaternion-based nonlinear filtering methods are applied on the processing of measurements from the low-cost Micro-electromechanical Systems (MEMS) based Inertial Navigation system (INS) and Global Positioning System (GPS). One approach employs an Extended Kalman filter (EKF) propagating the quaternion vector using conventional vector addition operation. However, due to the fact that the unit sphere defined by the quaternion vector is not an Euclidean vector space, the vector addition and scaling should principally not be directly applied. Therefore, in the second approach, an Unscented Kalman filter (UKF) is used which propagates the quaternion vector based on the quaternion product chain rule, having a natural way of maintaining the normalization constraint. A field experiment based on the train ride is made for the comparison. The objective is to verify whether different handlings of nonlinearity in system models and different ways of propagating quaternion vector over time will practically yield differences in the estimation of attitude and sensor bias errors.

Quaternion-based Kalman filtering on INS/GPS

In this paper, we applied an eight-state Kalman filter on a software defined GPS receiver by replacing conventional tracking with the feedback solution from the navigation filter to control the local code and carrier generators. The code phase and carrier frequency outputs from the delay lock loop (DLL) and frequency lock loop (FLL) discriminators are employed as measurements to update the filter. A hardware-in-the-loop test is made using a GPS L1 signal simulator. Numerical results show that the integration of Kalman filter with tracking loops can improve the tracking robustness and bridges the short GPS outage periods (i.e., less than 4 satellites in view).

https://www.zess.uni-siegen.de/cms/papers/Yang2012GPSReceiverTracking.pdf

GPS receiver tracking loop design based on a Kalman filtering Approach

This paper presents a method on fine Doppler frequency acquisition in software GPS receiver, which is implemented in Matlab/Simulink. Compared with the existing FFT-based approach, it speeds up the handover from acquisition to tracking stages, and it requires much less computational burden. A field experiment is made to show the positioning performance of the software GPS receiver, which verifies the correctness of the applied acquisition and tracking techniques.

http://ieeexplore.ieee.org/iel5/6034696/6044246/06044301.pdf

Yuhong Yang

Papers

Tightly-coupled INS/GPS using Quaternion-based Unscented Kalman filter

Applying Quaternion-based Unscented Particle Filter on INS/GPS with Field Experiments

Quaternion-based Kalman filtering on INS/GPS

GPS receiver tracking loop design based on a Kalman filtering Approach

Applying a fine Doppler acquisition method on software GPS receiver with field experiment