Status

[mgrouch]

• Switch to FIFO for correct IMU timing
• Kalman 3D Wave using adaptation parameters as control inputs. Estimate uncertainty of control inputs on-line
• Mag calibration crashes. Too much stack usage?
• Docs: Mention other pseudo measurements for drift correction on real IMU
• Docs: Mention GPS timing and uncertainty factors
• Docs: Mention adaptation using parameters as control inputs and RW
• Docs: Mention timing and FIFO use on real IMU importance
• Docs: Cross ref figures with text
• Kalman 3D Wave use of temperature correction
• Stricter acceleration biases calibration
• Docs: Mention alternative double integration method with pos, velocity pseudo measurements. Choice for noisy IMU
• Docs: Jacobian in doc for Ih two different formulas are in the document. One is too approximate
• Docs: explain why PM results worse
• Try to tune better with no noise and no mag first
• OU as regularization
• De-trend needs on real IMU. Adaptation of de-trending. Tests for de-trend. Discussion.
• Oscillatory models introduce "ringing" which looks like fake motion if spectral model is just slightly off. Integrator models have low frequency drifts (looks like fake swell)
• Adaptation is very important for different sea states