Fig. 1. 3DM-GX4-25 X轴预处理后数据3DM-GX4-25 X-axis post-preprocessing data
Fig. 2. MTi-100 X轴预处理后数据MTi-100 X-axis post-preprocessing data
Fig. 3. Triaxial rate turntable and device packaging
Fig. 4. MTi-100 sensor gyro X-axis ARMA model parameters online update curve
Fig. 5. Comparison of output data before and after X-axis filter of MTi-100 gyro
Fig. 6. Comparison of output data before and after X-axis filter of 3DM-GX3-25 gyro
Fig. 7. Comparison of output data before and after X-axis filter of MTi-100 gyro
Fig. 8. Comparison of output data before and after X-axis filter of 3DM-GX3-25 gyro
Sensors | First set of data | Second set of data | Third set of data | Fourth set of data | Mti-100 | -0.341 | -0.344 | -0.346 | -0.360 | 3DM-GX4-25 | 0.054 | 0.060 | 0.063 | 0.067 |
|
Table 1. Output mean of two MEMS gyro
Model name | AR(p) | MA(q) | ARMA(p, q) | Self-correlation coefficient function | Trailing | q step cut-off | Trailing | Partial correlation coefficient function | p-step tailing | Trailing | Trailing |
|
Table 2. The traditional method of model identification
Parameters | Xk | Wk | A | B | ARMA (1, 1) | \begin{document}${\left[ {\begin{array}{*{20}{l}}
{{y_k}}&k
\end{array}} \right]^{\rm{T}}}$\end{document} | \begin{document}${\left[ {\begin{array}{*{20}{l}}
{{a_k}}&{{a_k} - 1}
\end{array}} \right]^{\rm{T}}}$\end{document} | \begin{document}$\left[ {\begin{array}{*{20}{l}}
{{\varphi _1}}&1\\
0&1
\end{array}} \right]$\end{document} | \begin{document}$\left[ {\begin{array}{*{20}{c}}
1&{{\theta _1}}\\
0&0
\end{array}} \right]$\end{document} | ARMA (2, 1) | \begin{document}${\left[ {\begin{array}{*{20}{l}}
{{y_k}}&{{y_{k - 1}}}&k
\end{array}} \right]^{\rm{T}}}$\end{document} | \begin{document}${\left[ {\begin{array}{*{20}{l}}
{{a_k}}&{{a_k} - 1}
\end{array}} \right]^{\rm{T}}}$\end{document} | \begin{document}$\left[ {\begin{array}{*{20}{c}}
{{\varphi _1}}&{{\varphi _2}}&1\\
1&0&0\\
0&0&1
\end{array}} \right]$\end{document} | \begin{document}$\;\left[ {\begin{array}{*{20}{c}}
1&{{\theta _1}}\\
0&0\\
0&0
\end{array}} \right]$\end{document} |
|
Table 3. The traditional method of model identification
Sensors | Methods | Mean | Mean variance | MTi-100 | Original output | -0.353 | 0.078 | KF | -0.353 | 0.037 | AKF | -0.353 | 0.019 | Method in this paper | -0.353 | 0.009 | 3DM-GX3-25 | Original output | 0.071 | 0.197 | KF | 0.071 | 0.195 | AKF | 0.071 | 0.099 | Method in this paper | 0.071 | 0.013 |
|
Table 4. Statistical characteristics of gyro output signal before and after filtering
Sensors | Methods | Mean | Mean variance | MTi-100 | Original output | 4.844 | 0.377 | KF | 4.844 | 0.333 | AKF | 4.844 | 0.315 | Method in this paper | 4.844 | 0.275 | 3DM-GX3-25 | Original output | 5.092 | 0.323 | KF | 5.092 | 0.321 | AKF | 5.092 | 0.317 | Method in this paper | 5.092 | 0.269 |
|
Table 5. Statistical characteristics of gyro output signal before and after filtering