Optimal Weighted Fusion Based on Recursive Least Squares for Dynamic North-Finding of MIMU on a Tilting Base

• Main Authors: Haifeng Xing, Zhiyong Chen, Xinxi Zhang, Haotian Yang, Meifeng Guo
• Format: Article
• Language: English
• Published: IEEE 2019-01-01
• Series: IEEE Access
• Subjects: MIMU; optimal weighted fusion; recursive least squares; fast Fourier transform; dynamic north-finding; tilting base
• Online Access: https://ieeexplore.ieee.org/document/8765725/

Accurate estimation of orientation is very important in the field of inertial navigation. With the development of the technology of micro electro mechanical systems (MEMS), suppressing the influence of a gyro drift by means of continuous rotation (“carouseling”) has become a research hotspot. However, most of the current studies are still experiments conducted on the horizontal plane but carrying out north-seeking research on non-horizontal planes is also much valuable because it is common in practical scenarios, such as in the mining and drilling environment. In this paper, a feasible dynamic north-finder method for a micro inertial measurement unit (MIMU) based on the MEMS technology on a tilting plane is proposed. The fast Fourier transform (FFT) algorithm is used to analyze the appropriate rotation rate, and the tilting angle and heading angle are calculated in real time by an optimal weighted fusion based on recursive least squares (OWFBRLS) algorithm. Higher-precision orientation results can be achieved through the introduction of the optimal weighted fusion (OWF) theory. This paper demonstrates the experimental method and data processing in detail. The experimental results indicate that the proposed algorithm can provide accurate inclinations and headings, and the standard deviation (SD) of heading angle can reach approximately 0.6° in 3 minutes, which is superior to the methods in the literature and still has high precision in the case of large tilting angles. These profound results prove the feasibility and effectiveness of the proposed method.