Adaptive Kalman filter based on multiple fading factors for fast in-motion initial alignment with rotation modulation technique

Abstract

In this paper, a novel fast fine alignment method is proposed on the basis of micro electro mechanical systems (MEMS) for unmanned aerial vehicle (UAV) under mooring conditions. Firstly, the model of single-axis rotation modulation with extended angular rate measurement is built to improve the speed of convergence, especially the azimuth misalignment angle and the constant error of gyroscope along the rotation axis. Both model mismatch and noise uncertainty of the extended measurement will arise for the difficulty to obtain the exact angular velocity of the carrier without auxiliary sensors. To address this problem, the adaptive Kalman filter based on multiple fading factors is proposed. Besides, the meta-heuristic beetle antennae search (BAS) algorithm is applied for the first time to optimize the fading factors in real time. Then, the effectiveness of the proposed method is verified by means of singular value decomposition and error covariance matrix analysis. Finally, the result of comparative simulations and experiments show that the observability degree and convergence speed of misalignment angles improve significantly compared to those conventional methods and the estimated values can satisfy the requirements of fine alignment. More importantly, the constant error of gyroscope along the rotation axis is estimated accurately and can be compensated in the subsequent navigation process.

Kalman Filter-based Signal Processing for Robot Target Tracking

“…The research object of moving target tracking is video sequence, or image sequence, which refers to the spatialtemporal changes of the moving target in the whole sled [1], such as the appearance and disappearance of the target, the position, size and shape of the target, etc. [2]. It is caused by the presence of illumination changes, background interference, shadows, camera jitter, and occlusion between moving signals.…”