If BG data is jumpy/noisy, AAPS may dose insulin incorrectly resulting in highs or lows. If you observe errors in your CGM data it is important to disable the loop until the problem is resolved. Depending on your CGM, such issues may be due to the CGM configuration in AAPS (as explained further below); or a CGM sensor site issue (which may require replacing the CGM sensor).
As of AAPS version 3.2, AAPS offers the option to smooth the data within AAPS rather than within the CGM app. There are four options available in Config Builder > Smoothing.
In general, this is the recommended option to start with, as it is most aggressive in resolving noise and rewrites the most recent value. However, see the table below for other specific recommendations.
This option works similar to back smoothing that was previously implemented on certain CGM platforms. It is more reactive to recent changes in BG value and therefore more prone to responding incorrectly to noisy CGM data.
Use this option only if your CGM data is being properly smoothed by your collector app before being transmitted to AAPS.
This option uses a statistical model to continuously estimate your true glucose value and its rate of change from the incoming (possibly noisy) readings, automatically adapting to how noisy your sensor currently is. It also detects and rejects sudden outlier readings. It is a more advanced alternative to Exponential and Average smoothing, useful if you want the smoothing to adjust itself to your sensor's changing noise level rather than using a fixed method.
(smoothing-xdrip-dexcom-g6)=
| Exponential | Average | None | |
|---|---|---|---|
| G5/G6/ONE | If noisy | Recommended | |
| G7/ONE+/Stelo | If noisy | If stable |
Libre sensors are noisy and can require smoothing. When using xDrip direct connection, or the patched app data source (receiving from another app, Juggluco included), smoothing is already done inside the app.
| Sensor / Data source | Juggluco | xDrip direct | xDrip bridge | xDrip patched app |
|---|---|---|---|---|
| Libre 1/14 days/Pro | N.A. | N.A. | Average | N.A. |
| Libre 2/2+ (EU) | Average | None | Average | None |
| Libre 2/2+/3/3+ | Average | N.A. | N.A. | None |
