The stochastic errors present in the inertial sensors are most commonly analysed using Allan variance technique. The traditional Allan variance technique requires large time for computing noise parameters wherein the cluster size is varied in an arithmetic sequence. In this paper, redundant data removal and K-means clustering based techniques have been adopted to reduce the size of dataset. The reduced dataset decreases the overall time taken to compute the Allan variance. The least square fitting has been applied on the log-log plot of Allan deviation for obtaining the bias instability and angular random walk of an inertial sensor. The proposed scheme of redundancy removal is found to have a promising performance at the cost of reduced accuracy. However, clustering based data reducing is not found to work suitable. The dataset for noise characterization has been obtained from an ADIS16448 sensor suite by logging the raw data offline.
[1]
Jintao Li,et al.
Not Fully Overlapping Allan Variance and Total Variance for Inertial Sensor Stochastic Error Analysis
,
2013,
IEEE Transactions on Instrumentation and Measurement.
[2]
Xiaoji Niu,et al.
Analysis and Modeling of Inertial Sensors Using Allan Variance
,
2008,
IEEE Transactions on Instrumentation and Measurement.
[3]
Jintao Li,et al.
Sliding Average Allan Variance for Inertial Sensor Stochastic Error Analysis
,
2013,
IEEE Transactions on Instrumentation and Measurement.
[4]
Jang Gyu Lee,et al.
Performance Improvement of GPS/INS Integrated System Using Allan Variance Analysis
,
2004
.
[5]
D. W. Allan,et al.
Statistics of atomic frequency standards
,
1966
.