NASA TROPICS Mission: microwave radiometer calibration using machine learning

The Time-Resolved Observations of Precipitation structure and storm Intensity with a Constellation of Smallsats (TROPICS) mission was selected by NASA as part of the Earth Venture-Instrument (EVI-3) program. Post-launch Cal/Val indicated a significant benefit to using a new calibration approach relative to the original, more traditional calibration scheme. The new approach uses early-orbit observations minus background (O-B) to optimize the calibration parameters by minimizing the difference between the O-B under clear-sky conditions. A multilayer feed-forward neural network was used in some channels to infer and correct complicated calibration artifacts. The calibration parameterization follows a generalized version of the periodic absolute calibration equation typically used for cross-track passive microwave radiometers. Calibration parameter regression and machine learning use various telemetry as predictors. Finally, the scan bias correction of the sidelobe contamination and a causal calibration sustainment system is applied. Absolute calibration performance is on par with present operational sounders.