Testing Near-Real-Time Remote Science Operations in the Field: NIRVSS in BASALT

Transect Days Area of Interest Maps Threshold Days Testing Near-Real-Time Remote Science Operations in the Field: NIRVSS in BASALT ! Amanda M. Cook, Anthony Colaprete, Ted Roush, Fritz Renema, Rich Bielawski, Erin Fritzler, Josh Benton, Josh Forgione, Bruce White, Stephen Battazzo NASA Ames Research Center ! A! B! C! D! E! 1! 2! 3! 4! Generally, how do remote scientists respond to realtime data? How are operational thresholds determined? What is the appropriate cadence for making nearreal-time operational decisions? Resource Prospector Mission The Resource Prospector (RP) mission is a lunar lander and prospecting rover with a one-meter excavating drill. The goal of the mission is to detect and measure concentrations of H2O and OH, and to make a quantitative assessment of whether these and other materials are extractable as a future lunar resource. Potential uses include fuel generation, water for human exploration, and rare earth elements and minerals for use in space and on Earth. The Near-Infrared Volatile Spectrometer System (NIRVSS) is one of the prospecting instruments included in the RP payload. NIRVSS includes dual near-infrared (NIR) spectrometers, a near-infrared tungsten light source, a visible-wavelength high-resolution CMOS camera accompanied by LEDs at 8 wavelengths, and four radiometers to measure surface temperatures. NIRVSS points downward from the underside of the rover, viewing the surface and collecting data in real time as the rover traverses, and during drilling operations. As the rover traverses, science data will be downlinked to Earth at a ~5 minute delay. This very short delay allows for near-real-time operator decisions to be possible. Decisions and traverse plans can be adjusted at a faster cadence (compared to, e.g., Mars rover operations), to achieve maximum science within a constrained mission duration. Time-constrained Operational Decisions The criticality of this decision-making process to the success of RP necessitates thoughtful testing and planning with instrument and science teams to maximize science output. To this end, the NIRVSS science and instrument team participated in a field testing campaign called BASALT in November 2016. The Biologic and Analog Science Associated with Lava Terrains (BASALT) Project is a three-year program with the goal of “enabling the human-robotic exploration of Mars” (Nawotniak et al. 2017). The NIRVSS Team participated in the BASALT deployment to Hawaii’s Mauna Ulu, to test operational scenarios in a higher-stakes environment where scientific discovery was possible. The goals of the test were to learn more about so-called “decisional” data (data that can generate changes in mission planning), to get a sense for the real-time value of each of NIRVSS’s data types, and to study the decision-making process with an eye toward improving data analysis tools to aid the science team in directing the mission. ! NIRVSS Instrument Detail NIR Spectrometers • Short-wave (1.5 – 2.3 μm) • Long-wave (2.2 – 3.6 μm) • Configurable Field of View Radiometers • 8, 10, 12, 14, 25 μm detectors Ames Imaging Module • Designed for Space – Thermal considerations, EMI/C, radiation tolerance, power, and bandwidth – Long-duration and beyond-LEO applications • CMOS detector for Vis-NIR imagery • InGaAs detector for IR imagery • LEDs: white, 410, 540, 640, 740, 905, 940, and 1050 nm BASALT Test Scenarios are meant to mimic Resource Prospector traverse types. What are the obstacles to fast-paced decisionmaking? Field Team Acts as Rover Portable Backpack Prototype: 40 lbs Area of Interest Marked by Orange Flags Ames Imaging Module 8 LED types for camera illumination Four radiometers Fiber-optics feed to spectrometers NIR tungsten light source for spectrometers For each Test Scenario, the remote science team was presented with the corresponding challenge below Transect Day Challenge: Select 2-3 target types for revisiting or sampling based on just spectra, or just images. Area of Interest Map Challenge: Using imagery and spectra, select 2-3 targets for revisiting or sampling. Identify unique zones in the Area of Interest. Threshold Day Challenge: Using data from all previous days, determine a band depth threshold at which more detailed measurements should be performed. Going-in Questions Methods 8 in 6.75 in 3.29 in Near-IR Spectrometers Early Outcomes: New Image Analysis Tool: Automated creation of composite images and visible spectra of user-selected targets Results & Lessons Learned • Our science team chose band depth thresholds that were too low. The field team ended up taking detailed measurements of more than half of the targets on the threshold day. Team needs practice with defining stricter thresholds that are still highlighting the most important target types. • The remote science team had fast tools to assess spectral data. It seemed very quick for them to differentiate target classes (e.g. unaltered, moderately altered, heavily oxidized) based on spectra alone. • The remote science team could not assess image data fast enough to make effective operational decisions. This revealed a strong need for better real-time image processing tools. See panel to right! • Area of interest mapping at 1-meter resolution provided excellent context of a 4 x 5 meter area. Using primarily spectra, the scientists were able to identify a region of heavy oxidation, that was centered around an active fumarole (“3d” in maps below). OH-bearing!minerals! Carbonates! Fe-oxides!(?)! High OH Map High Carbonates Map High Fe-Oxides Map 1. Field Team acts as Rover, moving instrument over targets at Mauna Ulu volcano in Hawai’i. Field team collected data in three test scenarios (see below) to mimic Resource Prospector traverse types. 2. Data was sent to the remote science team at Ames for time-constrained decisions. They were required to respond with a measurement plan within 1 hour.! Example of Active Fumaroles at Mauna Ulu