Abstract
“Adaptive Calibration of CubeSat Radiometer Constellations” (ACCURACy) is a novel, constellation-level calibration framework developed to address the challenges in calibrating constellations of CubeSat radiometers. CubeSats are smaller than conventional monolithic systems, and so while cheaper and easier to develop and deploy, a lack of sufficient radiation shielding and thermal mass results in gain fluctuations due to the increased sensitivity to ambient conditions. Weight, cost, and power requirements also mean that CubeSats need to rely on vicarious calibration measurements, rather than external blackbody references, which cannot be collected frequently. ACCURACy is a multi-module framework which uses onboard telemetry data to cluster radiometers in similar states, creating cluster-level calibration data pools to share calibration data between radiometers in the same cluster. This is in contrast to existing constellation-level inter-calibration methods, in which a) one target sensor is selected within the constellation as an absolute calibration reference calibrated using blackbodies or RTMs, and then b) calibration measurements are collected from co-located constellation members in orbit when available and are used to eliminate biases between them in post-processing. This method is not suitable for constellations of identical CubeSat radiometers to obtain frequent-revisit, real-time, consistent observations with broad coverage.
ACCURACy has been developed and tested using a radiometer simulator, which was created in MATLAB to produce synthetic radiometer data for simulated constellations of CubeSats across a broad range of conditions. It has also been tested using various clustering algorithms and many different clustering parameters in order study the propagation of error in the system, and to reduce and quantify errors and uncertainties in calibrated products. This analysis establishes a relationship between uncertainties in calibration measurements and telemetry data in input data, cluster size and variance, and RMSE and uncertainty in calibrated products.
ACCURACy is also evaluated against existing constellationlevel inter-calibration methods using constellation simulations, comparing the RMSE and uncertainty in the calibrated antenna measurements for all radiometers in a constellation.
Adaptive Calibration of CubeSat Radiometer Constellations
“Adaptive Calibration of CubeSat Radiometer Constellations” (ACCURACy) is a novel, constellation-level calibration framework developed to address the challenges in calibrating constellations of CubeSat radiometers. CubeSats are smaller than conventional monolithic systems, and so while cheaper and easier to develop and deploy, a lack of sufficient radiation shielding and thermal mass results in gain fluctuations due to the increased sensitivity to ambient conditions. Weight, cost, and power requirements also mean that CubeSats need to rely on vicarious calibration measurements, rather than external blackbody references, which cannot be collected frequently. ACCURACy is a multi-module framework which uses onboard telemetry data to cluster radiometers in similar states, creating cluster-level calibration data pools to share calibration data between radiometers in the same cluster. This is in contrast to existing constellation-level inter-calibration methods, in which a) one target sensor is selected within the constellation as an absolute calibration reference calibrated using blackbodies or RTMs, and then b) calibration measurements are collected from co-located constellation members in orbit when available and are used to eliminate biases between them in post-processing. This method is not suitable for constellations of identical CubeSat radiometers to obtain frequent-revisit, real-time, consistent observations with broad coverage.
ACCURACy has been developed and tested using a radiometer simulator, which was created in MATLAB to produce synthetic radiometer data for simulated constellations of CubeSats across a broad range of conditions. It has also been tested using various clustering algorithms and many different clustering parameters in order study the propagation of error in the system, and to reduce and quantify errors and uncertainties in calibrated products. This analysis establishes a relationship between uncertainties in calibration measurements and telemetry data in input data, cluster size and variance, and RMSE and uncertainty in calibrated products.
ACCURACy is also evaluated against existing constellationlevel inter-calibration methods using constellation simulations, comparing the RMSE and uncertainty in the calibrated antenna measurements for all radiometers in a constellation.