On the state independency and log-linearity of error propagation for discrete group affine systems with application to attitude estimation

Purpose This paper aims to propose a general and rigorous study on the propagation property of invariant errors for the model conversion of state estimation problems with discrete group affine systems. Design/methodology/approach The evolution and operation properties of error propagation model of discrete group affine physical systems are investigated in detail. The general expressions of the propagation properties are proposed together with the rigorous proof and analysis which provide a deeper insight and are beneficial to the control and estimation of discrete group affine systems. Findings The investigation on the state independency and log-linearity of invariant errors for discrete group affine systems are presented in this work, and it is pivotal for the convergence and stability of estimation and control of physical systems in engineering practice. The general expressions of the propagation properties are proposed together with the rigorous proof and analysis. Practical implications An example application to the attitude dynamics of a rigid body together with the attitude estimation problem is used to illustrate the theoretical results. Originality/value The mathematical proof and analysis of the state independency and log-linearity property are the unique and original contributions of this work.

LiDAR DNN based self-attitude estimation with learning landscape regularities

This paper presents an EKF (extended Kalman filter) based self-attitude estimation method with a LiDAR DNN (deep neural network) learning landscape regularities. The proposed DNN infers the gravity direction from LiDAR data. The point cloud obtained with the LiDAR is transformed to a depth image to be input to the network. It is pre-trained with large synthetic datasets. They are collected in a flight simulator because various gravity vectors can be easily obtained, although this study focuses not only on UAVs. Fine-tuning with datasets collected with real sensors is done after the pre-training. Data augmentation is processed during the training in order to provide higher general versatility. The proposed method integrates angular rates from a gyroscope and the DNN outputs in an EKF. Static validations are performed to show the DNN can infer the gravity direction. Dynamic validations are performed to show the DNN can be used in real-time estimation. Some conventional methods are implemented for comparison.

Attitude Estimation Algorithm of Portable Mobile Robot Based on Complementary Filter

In robot inertial navigation systems, to deal with the problems of drift and noise in the gyroscope and accelerometer and the high computational cost when using extended Kalman filter (EKF) and particle filter (PF), a complementary filtering algorithm is utilized. By combining the Inertial Measurement Unit (IMU) multi-sensor signals, the attitude data are corrected, and the high-precision attitude angles are obtained. In this paper, the quaternion algorithm is used to describe the attitude motion, and the process of attitude estimation is analyzed in detail. Moreover, the models of the sensor and system are given. Ultimately, the attitude angles are estimated by using the quaternion extended Kalman filter, linear complementary filter, and Mahony complementary filter, respectively. The experimental results show that the Mahony complementary filtering algorithm has less computational cost than the extended Kalman filtering algorithm, while the attitude estimation accuracy of these two algorithms is similar, which reveals that Mahony complementary filtering is more suitable for low-cost embedded systems.

Multiplicative-modified-Rodrigues-parameters-based strong tracking unscented Kalman filter for satellite attitude estimation

In this paper, an improved strong tracking unscented Kalman filter (STUKF) based on multiplicative modified Rodrigues parameters (MRPs) is proposed for satellite attitude estimation. The multiplicative MRPs are superior to additive ones in terms of attitude representation, especially when attitude angles are large. By minimizing the loss function in Wahba’s problem, a novel method of weighted average for MRPs is derived to replace the simple procedure. The generation of Sigma points, update of state variables and calculation of covariance matrices are all different from the existing literature to maintain the multiplicative property of MRPs. Simulation results by raw telemetry data from the on-orbit CubeSat Enlai-1 demonstrate the excellent performance of the proposed filter under large attitude angles.