Engineering topics
Publications and source records attributed to Castano, Rebecca.
This paper describes an integrated system for coordinating multiple rover behavior with the overall goal of collecting planetary surface data. The Multi-Rover Integrated Science Understanding System (MISUS) combines techniques from planning and scheduling with machine learning to perform autonomous scientific exploration with cooperating rovers.
The Autonomous Sciencecraft Experiment (ASE) will operate onboard the Earth Orbiter 1 mission 2004. The ASE software uses onboard continuous planning, robust task and goal-based execution, and onboard machine learning and patter recognition to radically increase science return by enabling intelligent downlink selection and autnomous retargeting. In this paper we will discuss how these AI technologies are synergistically integrated in multi-layer control architecture to enable a virtual spacecraft science agent.
Missions to Mars already under study call for very long autonomous traverses (on the order of hundreds of meters per sol) during which there will be limited contact with Earth. The onboard Autonomous Science Investigation System (OASIS) is a technology for increasing science return during rover traverses by prioritizing science opportunities. By prioritizing data for downlink onboard, it is expected that the set of images selected for downlink by OASIS will have a consistently higher scientific interest then any set of equal number of images of the same traverse obtained using random or periodic sampling. Thus, OASIS can be used to increase the science returned from a long traverse. OASIS prioritizes data and identifies science opportunities by extracting information from image data and then using this information to prioritize the data and identify potential new science opportunities. Currently, the information extracted from the images is the location of the rocks in the image and several properties of these rocks such as size, albedo, texture and shape, In this paper we give an overview of the role of rock detection in the OASIS system and describe the functionality of the rock detector. We then present results on representative images
Software has been developed to perform a number of functions essential to autonomous operation in the Autonomous Sciencecraft Experiment (ASE), which is scheduled to be demonstrated aboard a constellation of three spacecraft, denoted TechSat 21, to be launched by the Air Force into orbit around the Earth in January 2006. A prior version of this software was reported in Software for an Autonomous Constellation of Satellites (NPO-30355), NASA Tech Briefs, Vol. 26, No. 11 (November 2002), page 44. The software includes the following components: Algorithms to analyze image data, generate scientific data products, and detect conditions, features, and events of potential scientific interest; A program that uses component-based computational models of hardware to analyze anomalous situations and to generate novel command sequences, including (when possible) commands to repair components diagnosed as faulty; A robust-execution-management component that uses the Spacecraft Command Language (SCL) software to enable event-driven processing and low-level autonomy; and The Continuous Activity Scheduling, Planning, Execution, and Replanning (CASPER) program for replanning activities, including downlink sessions, on the basis of scientific observations performed during previous orbit cycles.
The Autonomous Sciencecraft Experiment (ASE) will operate onboard the Earth Orbiter 1 mission in 2003. The ASE software uses onboard continuous planning, robust task and goal-based execution, and onboard machine learning and pattern recognition to radically increase science return by enabling intelligent downlink selection and autonomous retargeting. In this paper we discuss how these AI technologies are synergistically integrated in multilayer control architecture to enable a virtual spacecraft science agent.
Explore the source record for details and available documents.
Currently Mars missions can collect more data than can be returned. Future rovers of increased mission lifetime will benefit from onboard autonomous data processing systems to guide the selection, measurement and return of scientifically important data. One approach is to train a neural net to recognize spectral reflectance characteristics of minerals of interest. We have developed a carbonate detector using a neural net algorithm trained on 10,000 synthetic Vis/NIR (350-2500 nm) spectra. The detector was able to correctly identify carbonates in the spectra of 30 carbonate and noncarbonate field samples with 100% success. However, Martian dust coatings strongly affect the spectral characteristics of surface rocks potentially masking the underlying substrate rock. In this experiment, we measure Vis/NIR spectra of calcite coated with different thicknesses of palagonite dust and evaluate the performance of the carbonate detector.
Breakthrough autonomy technologies enable a new range of spire missions that acquire vast amounts of data and return only the most scientifically important data to Earth. These missions would monitor science phenomena in great detail (either with frequent observations or at extremely high spatial resolution) and onboard analyze the data to detect specific science events of interest. These missions would monitor volcanic eruptions, formation and movement of aeolian features. and atmospheric phenomena. The autonomous spacecraft would respond to science events by planning its future operations to revisit or perform complementary observations. In this paradigm, the spacecraft represents the scientists agent enabling optimization of the downlink data volume resource. This paper describes preliminary efforts to define and design such missions.
In this paper, we will describe out methods for the prioritization of geologic data acquired by an in-situ rover. Our techniques are applicable to a wide range of data modalites, however out initial demonstration is focused on image analysis, as images consume a large volume of the downlink bandwidth for such missions.
Explore the source record for details and available documents.
In the quest to maximize the scientific return of future robotic missions, it is imperative that our rovers be capable of determining the importance of the science they collect so that they may prioritize the acquisition and relay of that data. As an important step in this process, we present an automated technique to allow a rover to classify the shape and other geologic characteristics of rocks from two-dimensional photographic images and three-dimensional stereographically produced data. Experiments were conducted in the Matlab environment using images returned by JPL's Mars Pathfinder mission. Our method begins by first segmenting the rocks from the background using a combination of image intensity and height data. Various metrics are then used to classify the region's sphericity, roundness, and other geometric properties. Preliminary experiments to determine the most useful metrics were conducted by characterizing the two-dimensional rock shape while the three-dimensional shape was later studied with metrics derived from these two-dimensional techniques. Seven measures were developed and implemented. The performance of each measure was characterized by analyzing imagesfrom the Pathfinder mission and ranking the rocks according to the measured properties. Combined, the measures would provide a tool by which an automated rover could discover a greater amount of information about the data it collects, leading to a more productive mission.
Many papers have been published concerning the analysis of visual texture and yet, very few application domains use texture for image classification. A possible reason for this low transfer of the technology is the lack of experience and testing in real-world imagery. In this paper, we assess the performance of texture-based classification methods on a number of real-world images relevant to autonomous navigation on cross-country terrain and to autonomous geology. Texture analysis will form part of the closed loop that allows a robotic system to navigate autonomously. We have implemented two different classifiers on features extracted by Gabor filter banks. The first classifier models feature distributions for each texture class using a mixture of Gaussians. Classification is performed using Maximum Likelihood. The second classifier represents local statistics using marginal histograms of the features over a region centered on the pixel to be classified. We measure system performance by comparison to ground truth image labels.
We provide preliminary evidence that existing algorithms for inferring small-scale gene regulation networks from gene expression data can be adapted to large-scale gene expression data coming from hybridization microarrays. The essential steps are (I) clustering many genes by their expression time-course data into a minimal set of clusters of co-expressed genes, (2) theoretically modeling the various conditions under which the time-courses are measured using a continuous-time analog recurrent neural network for the cluster mean time-courses, (3) fitting such a regulatory model to the cluster mean time courses by simulated annealing with weight decay, and (4) analysing several such fits for commonalities in the circuit parameter sets including the connection matrices. This procedure can be used to assess the adequacy of existing and future gene expression time-course data sets for determining transcriptional regulatory relationships such as coregulation.
We introduce a statistical data model and an associated optimization-based clustering algorithm which allows data vectors to belong to zero, one or several "parent" clusters. For each data vector the algorithm makes a discrete decision among these alternatives. Thus, a recursive version of this algorithm would place data clusters in a Directed Acyclic Graph rather than a tree. We test the algorithm with synthetic data generated according to the statistical data model. We also illustrate the algorithm using real data from large-scale gene expression assays.