Maximizing the utility of limited Earth observing satellite resources is a difficult ongoing problem. Dynamic Targeting is an approach to this challenge that intelligently plans and executes primary sensor observations based on information from a look-ahead sensor. However, current implementations have failed to account for realistic satellite operational constraints and have used static utility for repeat observations of the same target. To address these limitations, we implement a more general Dynamic Targeting framework that comprises a physics-based slew model, a dynamic model of observation utility, and an algorithm for gathering high-utility observations. To demonstrate this framework, we also supply complex dynamic utility models that are applicable to many missions and new algorithms for intelligently scheduling observations with slewing restrictions and changing utility, including a greedy algorithm and a depth-first search algorithm. To evaluate these algorithms, we test their performance across simulated runs through two datasets and compare to the performance of an algorithm representative of most scheduling algorithms aboard Earth science missions today as well as an intractable upper bound. We show that our algorithms have great potential to improve science return from Earth science missions.

Flooding has serious environmental and humanitarian effects. To track these effects, previous work has used remote sensing to achieve global monitoring at low to moderate resolutions or regional monitoring at high resolutions. We proposed that implementing a new sensorweb that had previously been prototyped only regionally in Thailand could combine moderate-resolution flood detection with targeted high-resolution observations to enable worldwide high-resolution flood monitoring. Furthermore, we aimed to integrate both commercial and government satellites into this sensorweb to improve upon previous efforts, which only used government satellites. To this end, we first gather data from several moderate-resolution sensors to identify large, flooded regions. We then task high-resolution sensors to observe these floods and analyze the resulting data, thus enabling global high-resolution flood monitoring. Overall, our approach improves worldwide analysis of floods, which can be further improved by our ongoing efforts to incorporate more flood detection sensors and generate more products.

We collect alerts of volcanic activity from space remote sensing, volcano observatories, and meteorological organizations [1], [2]. We use these alerts to autonomously generate <1 m/pixel observation requests for Planet SkySat [3] and 70m/pixel thermal infrared data from the ECOSTRESS instrument on the ISS. We also automatically download Dove imagery corresponding to alerts to supplement the SkySat and ECOSTRESS imagery. One new feature is that we can then automatically generate thermal products from the Planet image data using newly-developed band-comparison classifiers.This system has been running continuously for nearly 1.5 years without a dedicated operations team. It has generated over 650,000 alerts, tasked over 80 SkySat observations, and resulted in more than 6,000 scenes scheduled for ECOSTRESS.

Dynamic Targeting (DT) will enable future Earth Observing instruments to intelligently reconfigure and point instruments to dramatically enhance science return. In this work we present a realistic simulation study of DT for tracking of storm features. To this end we have developed several algorithms from Operations Research and Artificial Intelli-gence/heuristic search. We benchmark these algorithms and show that DT is a powerful tool with the potential to significantly improve science yield.

Deep space missions can benefit from onboard image analysis. We demonstrate deep learning inference to facilitate such analysis for future mission adoption. Traditional space flight hardware provides modest compute when compared to today's laptop and desktop computers. New generations of commercial off the shelf (COTS) processors designed for embedded applications, such as the Qualcomm Snapdragon and Movidius Myriad X, deliver significant compute in small Size Weight and Power (SWaP) packaging and offer direct hardware acceleration for deep neural networks. We deploy neural network models on these processors hosted by Hewlett Packard Enterprise's Spaceborne Computer-2 onboard the International Space Station (ISS). We benchmark a variety of algorithms trained on imagery from Earth or Mars, as well as some standard deep learning models for image classification.

The Smart Ice Cloud Sensing (SMICES) instrument aims at providing onboard smart autonomous observation of upper tropospheric water vapor and ice particle size distribution in clouds at various local times. SMICES is an active/passive combined sensor with sounding channels at 380 GHz, radiometric channels at 250, 310 and 670 GHz, and a radar instrument operating at 239 GHz. A low-noise, low-power radiometer command and data handling (C&DH) subsystem has been designed to acquire the 24 analog radiometer channels and 8 analog thermistor data. A radiometric power regulation system provides the required power supplies for the other radiometric subsystems of the SMICES instrument. An on-board FPGA provides command and control of other instrument subsystems, performs synchronous data acquisition. The radiometer electronics are designed to fit into less than 2U horizontal dimensions of a CubeSat instrument. An AI controller unit directly interfacing with radar and radiometer C&DH subsystems performs on-board artificial intelligence operations for full system autonomy. The AI unit will control the radar instrument depending on the system health conditions, including the battery level, and based on the observed scene through the radiometer instrument.

Future space missions will process and analyze imagery onboard placing greater demands on flight computing. Traditional flight hardware provides modest compute, even when compared to laptop and desktop computers. A new generation of commercial off the shelf (COTS) processors, such as Qualcomm Snapdragon, deliver significant compute in small Size Weight and Power (SWaP) and offer direct hardware acceleration in the form of Graphics Processing Units (GPU) and Digital Signal Processors (DSP). We benchmark a variety of instrument processing and analysis software (including machine learned classifiers) on a Qualcomm Snapdragon SoC currently hosted by HPE’ s Spaceborne Computer-2 (SBC-2) onboard the International Space Station.