Long-term stability of the SGA-WZ strapdown airborne gravimeter.
Bottom Line:
The test results reveal a quadratic drift in the strapdown airborne gravimeter data.After compensating for the drift, the drift effect improved from 70 mGal to 3.46 mGal with a standard deviation of 0.63 mGal.The quadratic curve better reflects the drift's real characteristics.
View Article:
PubMed Central - PubMed
ABSTRACT
Accelerometers are one of the most important sensors in a strapdown airborne gravimeter. The accelerometer's drift determines the long-term accuracy of the strapdown inertial navigation system (SINS), which is the primary and most critical component of the strapdown airborne gravimeter. A long-term stability test lasting 104 days was conducted to determine the characteristics of the strapdown airborne gravimeter's long-term drift. This stability test was based on the first set of strapdown airborne gravimeters built in China, the SGA-WZ. The test results reveal a quadratic drift in the strapdown airborne gravimeter data. A drift model was developed using the static data in the two end sections, and then this model was used to correct the test data. After compensating for the drift, the drift effect improved from 70 mGal to 3.46 mGal with a standard deviation of 0.63 mGal. The quadratic curve better reflects the drift's real characteristics. In comparison with other methodologies, modelling the drift as a quadratic curve was shown to be more appropriate. Furthermore, this method allows the drift to be adjusted throughout the course of the entire campaign. No MeSH data available. Related in: MedlinePlus |
Related In:
Results -
Collection
License getmorefigures.php?uid=PMC3472875&req=5
Mentions: As an example, consider the data mentioned above. Fitting the drift with a quadratic curve using the data for the first 10 days and last 10 days produces the result shown in Figure 9. The compensated result using quadratic curve is shown in Figure 10 after subtracting the starting value. Comparing Figure 10 with Figure 7 indicates the drift was compensated well. In Figure 10, the drift effect with regards to the starting time is only 3.46 mGal, which is much less than the pre-correction value of 70 mGal. For comparison, the results using a linear fit based on the two end-points and using a cubic fit of the data for the first 10 days and last 10 days are also shown in Figure 9. The compensated results by the linear and cubic models are shown in Figure 11 and Figure 12, respectively. The statistical results of the differences between the fitting curves and the drift are shown in Table 2. As Table 2 shows, when fitting the drift using static data in the two end sections, the quadratic fitting remains the best and the linear fitting is still poor while the cubic fitting worsened. Thus, modelling the drift as a quadratic curve is beneficial in practice. |
View Article: PubMed Central - PubMed
No MeSH data available.