Hybrid Kalman Filter Design for Inertial Sensor Signal Processing
G Harish Babu1, N Venkatram2, N Giribabu3

1G Harish Babu, Electronics and Computer Engineering, Koneru Lakshmaiah University (KL U),GUNTUR, India.
2N Venkatram, Electronics and Computer Engineering, Koneru Lakshmaiah University (KL U), GUNTUR, India.
3N Giri babu, Electronics and Communications, RCI-DRDO, HYDERABAD, India
Manuscript received on April 02, 2015. | Revised Manuscript received on April 08, 2015. | Manuscript published on April 25, 2015. | PP:1-4 | Volume-3 Issue-6, April 2015. | Retrieval Number: F0946043615

Open Access | Ethics and Policies | Cite
© The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: Gyroscope is a rotation rate sensor. Olden days gyroscopes includes mechanical parts like motors, gimbals etc. which has suffered with noises and low accuracy. Developments in fiber optics has made the interferometric FOG practically realizable and reliable compared to mechanical gyroscopes. The fiber optic based gyro results in good stability, high disturbance rejection and ensure good tolerance to noises. Different noises like photon shot noise, quantization noise, filter noise, thermal noise and bias error many other noises degrades the gyro performance. Significant effect of these noise s result in random walk, bias unstability , power fluctuations etc .To overcome these noises denoising of gyro data play a crucial role in FOG. In this study kalman filter algorithm is used to denoise the FOG signal . The algorithm implemented here performs efficiently in both static and dynamic conditions. The existing KF algorithm is hybridized with adaptive moving average based dual gain kalman filter.In this study comparative study is made between different kf algorithms in both static and dynamic condition. In static condition three algorithms KF,ARMA KF & hybrid KF models are used for denoising the gyro data . Among all three algorithms hybrid model is provided to be more efficient. In case of dynamic condition kf algorithm is fails. ARMA model is used to identify the noise but fails in denoising the noise, where as hybrid kf model work efficiently this case. The denoising performance of hybrid model algorithm is validated on single axis FOG and three axis FOG with different input rotation rates
Keywords: Adaptive moving average, auto regressive moving average ,fiber optic gyroscope, kalman filter