Micro-electro-mechanical-systems (MEMS) inertial measurement unit (IMU) outputs are corrupted by significant sensor errors. The navigation errors of a MEMS-based inertial navigation system will therefore accumulate very quickly over time. This requires aiding from other sensors such as Global Navigation Satellite Systems (GNSS). However, it will still remain a significant challenge in the presence of GNSS outages, which are typically in urban canopies. This paper proposed a rotary inertial navigation system (INS) to mitigate navigation errors caused by MEMS inertial sensor errors when external aiding information is not available. A rotary INS is an inertial navigator in which the IMU is installed on a rotation platform. Application of proper rotation schemes can effectively cancel and reduce sensor errors. A rotary INS has the potential to significantly increase the time period that INS can bridge GNSS outages and make MEMS IMU possible to maintain longer autonomous navigation performance when there is no external aiding. In this research, several IMU rotation schemes (rotation about X-, Y- and Z-axes) are analyzed to mitigate the navigation errors caused by MEMS IMU sensor errors. As the IMU rotation induces additional sensor errors, a calibration process is proposed to remove the induced errors. Tests are further conducted with two MEMS IMUs installed on a tri-axial rotation table to verify the error mitigation by IMU rotations.

Huber M-estimation, as an estimation method based on mixed norm as cost function, provides an effective method for robust filtering to deal with measurement outliers. Based on statistical linear regression model approximating nonlinear measurement model, M-estimation algorithm is used to realize measurement update of states. However, when the measurement noise pollution rate is high, the filtering performance will be seriously degraded. In order to further improve the actual performance of data fusion under abnormal measurement noise, a vehicle cooperative positioning scheme based on adaptive M-estimation robust unscented Kalman filter (AMRUKF) was proposed. This algorithm combines Huber’s linear regression problem with covariance matching method, and calculates the adaptive matrix through the innovation covariance estimator based on fading memory index weighting to adaptively adjust the measurement noise covariance used in Huber M estimation method. The analytical and experimental results of tightly-coupled vehicle relative positioning show that the proposed AMRUKF method can effectively improve the robustness and accuracy of relative positioning and provide a practical control scheme for vehicle cooperative positioning.

It has been identified that the inertial system is not a completely observable system in the absence of maneuvers. Although the velocity errors and the accelerometer bias in the vertical direction can be solely observable, other error states, including the attitude errors, the accelerometer biases in the east and north directions, and the gyro biases, are just jointly observable states with velocity measurements, which degrades the estimation accuracy of these error states. This paper proposes an innovative method to improve the system observability for a Micro-Electro-Mechanical-System (MEMS)-based Inertial Navigation System (INS) in the absence of maneuvers by rotary motions of the Inertial Measurement Unit (IMU). Three IMU rotation schemes, namely IMU continuous rotation about the X, Y and Z axes are employed. The observability is analyzed for the rotating system with a control-theoretic approach, and tests are also conducted based on a turntable to verify the improvements on the system observability by IMU rotations. Both theoretical analysis and the results indicate that the system observability is improved by proposed IMU rotations, the roll and pitch errors, the accelerometer biases in the east and north directions, the gyro biases become observable states in the absence of vehicle maneuvers. Although the azimuth error is still unobservable, the enhanced estimability of the gyro bias in the vertical direction can effectively mitigate the azimuth error accumulation.

High accuracy positioning information plays an important role in the field of autonomous driving, where both the visual and inertial sensors have been widely noticed because the virtue that they do not rely on external information. However, the accumulation of sensor errors degrades system positioning accuracy in complex environments due to the fixed system model and variable measurement noise in traditional visual inertial odometry. A novel visual inertial odometry based on interactive multiple model and multi-state constrained Kalman filter is proposed. First, the trifocal tensor model constructed from three consecutive images is used as the measurement model of the system, and the corresponding position and orientation information is added to the filter state vector to form a multi-state constrained Kalman filter (MSCKF), which is then combined with the interactive multiple model (IMM) to form an IMM-MSCKF algorithm to interactively fuse the inputs and outputs of multiple sub-filters to improve the positioning accuracy of the visual inertial odometry (VIO). The proposed method is validated by selected urban environment and highway area data from publicly available datasets. The experimental results show that the proposed algorithm effectively improves the positioning accuracy of integrated navigation while reducing the positioning error of a single sensor compared to the conventional visual inertial odometry.

The real-time response and accuracy of the attitude (i.e., roll and pitch) estimation from low-cost inertial measurement unit (IMU) have become the key issues restricting related applications. This paper proposes a robust attitude estimation scheme which can perform well under dynamic conditions. When only accelerometers are used to calculate and correct the attitude, the external acceleration becomes the main source of attitude estimation errors. Moreover, the truncation error in the linearization process of the nonlinear system also affects the attitude estimation. As our first contribution, the external acceleration is modeled as a first-order Gauss Markov model, and its value is calculated under the indirect Kalman filter (IKF) framework. The measurement noise covariance matrix of the IKF is adaptively adjusted to enhance its robustness and reduce the negative impact caused by inaccurate modeling. In the second part of our work, the two-step cascade filter method is used for attitude estimation. The attitude obtained from the gravity field based on the gradient descent (GD) algorithm shows fast response capabilities, and hence, it is embedded as a measurement in the IKF by using the chain-derivation rule. The truncation error introduced into the linearization process of the nonlinear system is effectively avoided. Both simulation and experiments are carried out to verify the feasibility and accuracy of the proposed algorithm. The results show that the approach proposed in this paper can meet the accuracy requirements of consumer products.