Spatio-Spectral Exploration Combining In Situ and Remote Measurements

Authors: David Thompson, David Wettergreen, Greydon Foil, Michael Furlong, Anatha Kiran

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We simulate exploration using alternative planning approaches, and show proof of concept results with the canonical spectroscopic map of a mining district in Cuprite, Nevada.
Researcher Affiliation Collaboration David R. Thompson Jet Propulsion Laboratory California Institute of Technology 4800 Oak Grove Dr. Pasadena, CA 91109 david.r.thompson@jpl.nasa.gov David Wettergreen and Greydon Foil The Robotics Institute Carnegie Mellon University 5000 Forbes Ave. Pittsburgh, PA 15213 {dsw,gfoil}@ri.cmu.edu Michael Furlong Intelligent Robotics Group NASA Ames Research Center Naval Air Station, Moffett Field Mountain View, CA 94035 furlong@gmail.com Anatha Ravi Kiran Jet Propulsion Laboratory California Institute of Technology 4800 Oak Grove Dr. Pasadena, CA 91109 ravi.kiran@jpl.nasa.gov
Pseudocode Yes Algorithm 1: Greedy spectrum selection
Open Source Code No The paper does not provide a specific link or explicit statement about the availability of open-source code for the described methodology.
Open Datasets Yes We used the Airborne Visible Near Infrared Spectrometer (AVIRIS) to simulate rover in situ spectra (Green et al. 1998). This instrument observes the spectral region from 0.38µm to 2.5µm at a spectral resolution of 0.01µm and high spatial resolution, making it comparable in spectral range and resolution to field instruments. Our case study centers on the mining district of Cuprite, Nevada. It contains a wide range of distinctive mineralogical signatures, and has been extensively studied through scientific expeditions and field campaigns (Kruse 2002). We also used orbital remote sensing data from the ASTER instrument (Fujisada 1995).
Dataset Splits No The paper describes a simulation setup and evaluation, but does not explicitly specify training, validation, or test dataset splits in terms of percentages or sample counts for model reproduction.
Hardware Specification No The paper only states that 'The entire planning process took just a few seconds to run on a modern laptop computer,' which does not provide specific hardware details (e.g., CPU, GPU model, memory) used for running experiments.
Software Dependencies No The paper does not provide specific software dependencies or library versions needed to replicate the experiment.
Experiment Setup Yes We simulated 256 random trials of hypothetical traverses. Each had a random start and finish location at the left and right of the image. We defined a path length budget of 125% of the Euclidean distance between start and end locations. This provided a challenging constraint with enough margin to visit a few targets of opportunity. ... Each algorithm selected waypoints from a grid of candidate locations spaced at 20 pixels. The ASTER spectra associated with these grid points became our target library for the unmxing objective.