INS Initialization & Self Calibration
Proper initialization of attitude states is in fact the toughest problem in any INS design task. Most people simply assume that the INS can be started with some arbitrary external attitude information. However such an over-approximation may cause serious complications.
The following scprits demonstrates a couple of ways to determine the initial orientation based on inertial sensors. Self-calibration examples further demonstrates the use of redundant sensors and ZUPTs to estimate sensor errors during stationary instants
| align_opt_v000.m | Algebraic and SVD based attitude solution from vectors |
| align_wander_v000.m | Platform leveling using a rotation vector |
| alingc_2s_v000.m | 2 step IMU Coarse Align Method (first platform, then heading) |
| alingc_hd_v000.m | Coarse Align: Heading |
| alingc_pl_v000.m | Coarse Align: Platform |
| alingc_std_v001.m | Standard Coarse Align Procedure |
| example_align.m | Usage Examples for coarse align methods |
| example_selfcalib.m | Usage example for self calibration |
| selfcalib1_v001.m | Stationarity based self calibration using Kalman filter |
| selfcalib2_v000.m | Multi source based self calibration using Kalman filter |
| sys_def.m | system definitions for above examples. |