Debugging Phase 3

Kalman filtering results can be examined for Argus using the routine showHourlyKalmanResults and specifying the station and the beginning and end times to examine. This routine calls the plotBathyKalmanStep to display each ongoing result in a series of sequential bathy estimates. For non-­‐Argus systems, you should call this function from within a wrapper m-­‐file that is consistent with your system.

For each bathy file within the sequence, this routine plots the prior running average bathymetry, the current estimate (from fCombined in this bathy record), the updated running average, as well as errors on the prior and current estimates. The last panel is the Kalman gain. Cycling through a series of results lets you examine how the Kalman filter ingests good and bad data and how the final averaged results evolve. It is instructive to watch the Kalman gain since these numbers should normally be small once the filter has settled down (ingested a few days worth of hourly data).